This PDF file contains the front matter associated with SPIE Proceedings Volume 13170, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

With the purpose of better adapt to different types of complex terrain, this paper proposes an experimental scheme for ground image and LiDAR point cloud decision-level fusion block filtering and DEM generation. Firstly, the ground image and point cloud are matched through homologous feature points linear transformation to obtain more spectral texture information from the image. Then, based on decision-level fusion, the original point cloud is segmented into several independent blocks. The IPTD filtering algorithm is improved and optimized by using a search method according to the different multi-dimensional detail features of each block region. Finally, the filtered total ground points are interpolated to obtain a DEM model. The results show that compared with other overall filtering schemes, the DEM generated by this algorithm has the highest accuracy with an average absolute error and root mean square error of 0.100 meters and 0.157 meters respectively, greatly optimizing the generation effect of DEM.

For the impact of magnetic prospecting unmanned aerial vehicle route planning on the positioning accuracy of simulated unexploded ordnance, we conduct unmanned aeromagnetic prospecting flight tests and complete the localization of simulated unexploded ordnance using the normalized magnetic source intensity method, then analyze the localization results. The results indicate that as the survey line interval decreases to 0.5 m, the positioning error tends to stabilize. As the flight speed decreases from 5 m/s to 1 m/s, the positioning accuracy is significantly improved. As the flight altitude decreases from 3m to 1.5m, the positioning error decreases. By analyzing the impact of key parameters in magnetic prospecting unmanned aerial vehicle route planning on the positioning accuracy of simulated unexploded ordnance, we provide an empirical reference for the efficient and safe detection strategy of unexploded ordnance.

In order to make the positioning of civil airport PAPI system more convenient, accurate and flexible, based on the traditional positioning method of PAPI system, a positioning method of PAPI system based on Auto CAD drawing is proposed. The specific positioning process of traditional positioning method and drawing positioning method of PAPI system are summarized respectively. Through the comparative analysis of the two positioning methods, it is concluded that the drawing positioning method has the advantages of simple process, accurate positioning and flexible application compared with the traditional positioning method. It is suggested that drawing positioning method of PAPI system should be preferred in practical engineering design.

In order to enhance the predictive and forecasting capabilities of meteorological risk warning models for geologic hazards, this paper presents a new meteorological risk warning model for geologic hazards in Yichang City, Hubei Province, using heterogeneously integrated machine learning. Firstly, eight essential environmental factors were extracted from various data sources, including topographic, geological, and remote sensing imagery. Correlation tests were conducted alongside statistical analyses of landslide deformation and damage under diverse external hydrological conditions, such as rainfall intensity and duration. Consequently, nine rainfall warning indexes were derived, comprising metrics like previous period's effective rainfall, same-day inspired rainfall amount, and short-term rainfall intensity. These indexes, in conjunction with the fundamental environmental factors, formed the foundation of the landslide warning index system. Subsequently, multiple modeling techniques including logistic regression, CART decision tree, plain Bayes, a heterogeneous integration method as base models, were employed to establish the geologic hazards meteorological risk model. The results demonstrated that the integrated machine learning model achieved an accuracy rate of 0.835, surpassing the performance of the individual models. Then, the constructed model was further evaluated using historical rainfall and geologic hazards event data, which corroborated the alignment of the prediction outcomes with actual occurrences of geologic hazards. Collectively, this study's findings have important implications for the development of practical meteorological risk warning models for geologic hazards and provide a promising framework for improving regional disaster prevention, as well as emergency management planning.

With the resolution of spaceborne synthetic aperture radar reaching sub meter level, the mapping bandwidth is only 3km to 5 km. This reduces the effectiveness of aiming and capturing the observed target. Therefore, it is necessary to analyze the influencing factors for aiming accuracy of spaceborne SAR. This article conducts analysis from the aspects of ground telemetry/track and command system, ground operation control system, satellite platform and payload. An error model based on the influencing factors of entire satellite-ground link and on-orbit testing method suitable for any scenario are proposed. The results of theoretical prediction and real test show that 120m orbit height prediction error, 90m target elevation error, 5m slant measurement error, 0.03° satellite attitude pointing error and 0.01° systematic beam pointing error of SAR antenna will bring about 400m aiming accuracy.

The karst landforms in the southwest region of China make vegetation coverage susceptible to changes. Clearly understanding the spatiotemporal patterns of vegetation evolution is of great significance for ecological civilization construction and promoting green development in the region. The kernel normalized difference vegetation index (kNDVI) has overcome the "non-linearity" characteristic of NDVI, allowing for a more precise representation of vegetation growth status compared to NDVI. Hence, this paper utilizes Landsat data on the Google Earth Engine platform to retrieve kNDVI from 1990 to 2023. Subsequently, the Theil-Sen slope estimation method and Mann-Kendall test are applied to explore the spatiotemporal evolution patterns of vegetation, while the Hurst index is used to forecast the future trends of vegetation change. The results indicate that: (1) the overall vegetation in the study area has improved, with an increase slope of 0.0057. (2) the area with significant increases account for the majority, approximately 85.21%, suggesting an overall positive trend in the spatial distribution of vegetation in the study area. (3) the northeastern part of the study area and the central part of Guangdong are experiencing a slight but sustainable decrease, yet the future vegetation change is expected to be characterized mainly by significant and sustainable increases. The results of this study can provide scientific theoretical support for strengthening ecological environment protection and governance.

This paper uses Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR), taking the Anyuan District, Pingxiang City, Jiangxi Province as an example, with Sentinel-1A Synthetic Aperture Radar (SAR) data from January to December in 2022 to monitor the ground subsidence. The subsidence characteristics of different mines and settlement risk assessment are analyzed. The research shows that in 2022, the overall subsidence rate of Anyuan area is - 80mm/year to 59mm/year, and the subsidence rate on the southeast side is significantly higher than that on the southwest side. The currently closed Baiyuan mine area has no settlement trend. However, within the monitoring time, the Anyuan coal mine that is still in mining continues to subside. The method of subsidence risk assessment is given, which provides a theoretical reference for better understanding the law of surface subsidence.

The digital elevation model (DEM) is usually created based on a single vertical datum located at mean sea level. In most cases, however, the elevations of natural plains are trend downward (mostly toward the sea, quickly or slowly), which obscures the relationship between the natural descent and the subminiature terrain caused by tectonic. To remove this trend, relative elevation model (REM) was developed, which subtracted floodplain surface from DEM. The traditional method is limited to high resolution DEM, however, and is usually used in narrow band areas. Using medium-low resolution DEM, this paper presents a technique to create REM based on multi-rivers. This technique consists of three steps, namely finding rivers, calculating the trend and subtracting the trend from DEM. This technique was applied to the study area of Linyi district, Shandong, China. There is a great agreement between REMs and the location of faults, which means this technique is helpful to the study of subminiature terrain and active tectonics.

Aiming at the problem of intelligent generation of semantic information of vector planar water bodies, a graph neural network method that integrates morphological features, terrain characteristics, and spatial relationship features is studied. This method can generate more detailed type information under the broad category of planar water bodies, including Three categories: lakes, reservoirs, and ponds. Experimental results show that this method can effectively distinguish types of planar water bodies and has higher accuracy than existing methods, providing a new idea for the intelligent generation of semantic information about water bodies.

The data of polar-orbiting meteorological satellites and geostationary meteorological satellites from 2021 to 2022 were used, combined with land use types and administrative boundary vector data, and contextual inversion algorithms to adjust the key thresholds for fire point identification and monitor straw hot spots. UAVs and others tested the positioning accuracy, hot spot fire authenticity and hot spot classification results of satellite hot spot monitoring. The results showed that: satellite fire monitoring and hot spot classification were basically correct, and the positioning deviation was within 300m. The inversion method of satellite hot spot and the performance results are reliable. The high-incidence areas of straw hot spots in Hunan are located in the Dongting Lake Basin, Hengyang-Shaoyang-Loudi area, northern Yongzhou and eastern Chenzhou. The straw burning phenomenon was more serious in 2022 than that in 2021, and the high-density distribution range of straw hot spots was significantly expanded compared with 2021. There are two concentrated periods for straw burning: January to April and October to December. The number of straw hot spots in January, October and December accounts for 66.1% of the whole year. The daily distribution of straw hot spots is single peak. The number of straw hot spots began to increase rapidly at 11a.m and reached the peak at 14p.m. After that, the number of hot spots decreased rapidly, and the phenomenon of straw burning at night was significantly reduced.

Microwave remote sensing inversion of soil moisture (SM) products has gained increasing prominence in agricultural, meteorological, hydrological, and other research fields, and analyzing the regional applicability of microwave remote sensing SM data is a crucial prerequisite for their effective utilization. In order to assess the accuracy and applicability of SMAP, ASCAT, and AMSR2 remote sensing SM products in the North China Plain, we employed MERRA2 simulated SM as reference data, the triple collocation (TC) method was utilized to compare and evaluate the daily SM data obtained from remote sensing products in the region. The results of the TC analysis indicate that SMAP SM exhibited superior performance in the North China Plain. The majority of errors had a standard deviation of less than 0.023 m³ / m³, and the cumulative frequency of errors less than 0.027 m³ / m³ approached 1, surpassing ASCAT and AMSR2 overall. Additionally, SMAP exhibited the lowest deviation and root mean square error. The median correlation coefficient was highest for SMAP, followed by AMSR2 and ASCAT. Similarly, the median deviation and root mean square error were lowest for SMAP. The verification results of the evaluation index align with the TC findings. The error standard deviation of SMAP demonstrates the best performance across various land use types. The land use types with the best quality of SMAP, ASCAT and AMSR2 are cultivated land and grassland (low vegetation coverage area), among which ASCAT is better than AMSR2 in cultivated land and grassland. The error of urban construction land is the largest. This study aims to provide a scientific reference for enhancing the inversion algorithm and facilitating the extensive application of data products, and lay the foundation for future research such as im-proving product resolution and drought monitoring.

As a form of geographical combination of production layout, economic circle is a new spatial form of urbanization, a brand new governance scale to implement the national strategy of regional coordinated development, and an important bearing space to promote regional coordinated development. Therefore, it is crucial to study how land use is changing dynamically and its EcosystemServiceValue (ESV) response from the economic circle scale for regional land use planning and guaranteeing regional ecological security. Taking the Jiaodong Economic Circle as an example, based on the land use remote sensing monitoring data of five periods in Jiaodong Economic Circle from 2000 to 2020, an estimation model was established to assess ESV using the equivalent factor coefficient correction method, and sensitivity analysis, spatial autocorrelation were employed to explore the characteristics of spatial and temporal variations of ESV in Jiaodong Economic Circle. The results show that: from 2000 to 2020, the land use type of Jiaodong Economic Circle is dominated by cropland, accounting for more than 62% of the total area, and the expansion of construction land area is obvious, and the main sources of diversion are cropland and grassland; the ESV of Jiaodong Economic Circle as a whole shows a decreasing and then an increasing trend, with a total of 39.58 billion yuan during the 20-year period; the ESV of cultivated land, forest land and grassland continued to decrease, and the ESV of water area increased significantly; the ESV of Jiaodong Economic Circle " south is low while the north is high, high all around, low in the middle" was the spatial pattern that was visible, and there was a significant spatial clustering effect.

A water balance equation is used to calculate the water conservation capacity of Hehuang Valley, and ArcGIS analysis is carried out. The functional divisions of regional water conservation are divided comprehensively with the distribution of natural reserves (national nature reserves, national parks, and water sources). The results show that the areas of Lenglongling and Daban mountains in the northern Qilian Mountains are the most critical areas for water conservation, accounting for 25% of the total area, and the areas of general and medium important are mainly distributed in the middle and small rolling mountains, accounting for 66.3% of the total area; the areas of unimportant water conservation are concentrated in Huangshui and Yellow River Valley, where human activities are the most concentrated, accounting for 8.7% of the area.

Bathymetry data is an essential element in marine survey and mapping, and plays important roles in researching the Earth and guiding the underwater works. Satellite altimetry is one of the most effective ways to obtain global ocean gravity anomalies and vertical gravity gradients. In this study, the transfer function between gravity anomalies and gravity gradients and seabed topography is derived in spectral domain. Based on the error propagation analysis between altimetry data and gravity filed products, a bathymetry prediction method using different data fusion according to different water depths is proposed. A simulation is carried out in the Mariana Trench Seabed Terrain to verify the proposed method. The results show that vertical gravity gradients data does perform better in shallow water areas than gravity anomaly data does. And when the vertical gravity gradient data is used in the area with water depth shallower than 1000m, and the gravity anomaly data is used in the area with water depth deeper than 1000m, the predicted bathymetry has higher accuracy than that derived from gravity anomalies alone or vertical gravity gradients alone, which improves the RMS 11.35% compared to just using anomalies alone.

With the development of remote sensing technology, remote sensing images have gradually entered the mass view, bringing great convenience to public life. However, if the public maps labeling information is not properly handled, it is very easy to endanger national security. Based on the research and analysis of the characteristics of Internet remote sensing image maps, combined with the national requirements for the use of open Internet remote sensing image maps, this paper proposes the classification of sensitive information points of Internet remote sensing image maps and the detection method of sensitive information of Internet remote sensing image maps based on sensitive information points. The actual test results show that the proposed method can comprehensively, truly and objectively reflect the security status of Internet remote sensing image maps, and can be applied to the security assessment and supervision management of surveying and mapping geographic information data.

In view of the insensitivity of the available radar vegetation indices to dryland ecosystems, it is the purpose of this paper to develop a more pertinent new radar vegetation index, i.e., dryland radar vegetation index (DRVI), for land characterization using Sentinel 1 data based on the scattering entropy, scattering angle and cross polarization. After a comparison with the available radar indices and sensitivity analysis versus (vs) the optical remote sensing indices, the new index, DRVI, was found to have a stronger correlation with the optical vegetation indices such as the normalized difference vegetation index (NDVI), generalized difference vegetation index (GDVI), and mining and restoration assessment indices (MRAIs) and a higher sensitivity to dryland biomes than other radar vegetation indices. It is hence concluded that this DRVI has high potential of application in characterizing the dryland ecosystems despite it is derived from the Mu Us Sandy Land, China.

Land use change detection by remote sensing (RS) and simulation-based prediction may provide support to rational spatial planning of territory, which is of great significance to ensure sustainable land management. For this purpose, this paper employed multitemporal Landsat images dated 2010, 2015 and 2020 and county-level socioeconomic data for achieving such an integrated study taking Hefei city as an example. After atmospheric correction using the COST (Cosine Theta) model, the maximum likelihood approach and post-classification differencing were applied respectively to classify the images and to reveal the land use changes. The Logistic-Multi-Criteria Evaluation-Celluar Automata-Markov (LMCM) hybrid model was harnessed to simulate the land use pattern of 2025 under four different scenarios namely business as usual (BAU), ecological protection (EP), economic development (ED) and sustainable development (SD). The results show that (1) during the period 2010-2020, the built-up areas increased by 804.69 km² and the croplands decreased by 568.14 km²; (2) There seems to be a big difference between the predicted croplands and built-up areas in the four scenarios in 2025. Under the EP scenario, the built-up areas may likely gain an increase of 244.27 km², indicating that the environment shall be protected but the urbanization will be also limited. Under the ED scenario, the built-up areas may be extended by 829.54 km² and the croplands and forests are likely to decline significantly in 2025, meaning that overemphasizing urban development would not only threaten food security, but also bring harm to the local ecosystem. Under the SD scenario, the increase of built-up areas and the reduction of croplands would both come between the other scenarios, and the land use structure seems more stable and rational. Clearly, the SD scenario is the potential pathway for sustainable land use management in the long run. These results may provide the local governments with reference for their decision-making to conduct sustainable land use and urban planning.

Due to the unclear distribution of near-surface atmospheric carbon dioxide (CO₂) in small areas, by choosing Yulin City and Xi'an City as the research areas and applying Unmanned Aerial Vehicle (UAV) to carry out atmospheric CO₂ concentration monitoring at multiple carbon monitoring points, this study explores the impact of ground object type, height and time on atmospheric CO₂ concentration. Regarding the temporal variation, the atmospheric CO₂ concentration of the vegetation-covered area in the afternoon is lower than that in the morning in summer, and only the evergreen vegetation-covered area in winter has similar characteristics as that in summer. The near-surface atmospheric CO₂ concentration in the vegetated area changes greatly in summer and winter. For the characteristics of vertical distribution of CO₂, the CO₂ concentration gradually decreases with increasing altitude. In summer, the CO₂ concentration in areas affected by human activities decreases rapidly from 0 to 50 meters and decreases slowly from 50 to 150 meters. The CO₂ concentration in vegetation-covered areas decreases slightly from 0 to 150 meters. In winter, the areas affected by human activities and the vegetation-covered areas both decrease rapidly from 0–20 meters and relatively slowly from 20 meters to 100 meters. The experimental results indicate that using UAVs can accurately monitor the atmospheric CO₂ concentration near the surface of typical land features in small areas. UAVs can form air-space-ground carbon monitoring networks with satellites and ground observation points, which is of great significance for quantitative analysis of carbon reduction paths to achieve peak carbon dioxide emission and carbon neutralization.

First, the acoustic signal detection environment is preprocessed, and the target for detecting the depth of cracks in road bridge pavement is set. Then, based on this, an interleaved low-rank group convolutional hybrid depth detection structure is established, and the ILGCHDN bidirectional detection model is designed. Finally, the transient acoustic wave method is used to achieve crack depth detection in road bridge pavement. The test results show that compared to the traditional vertical crack depth detection group and the traditional mixed morphology segmentation detection group, the crack depth error mean obtained by the designed acoustic signal crack detection group is relatively small, indicating that the testing method is more reliable and the final results are more accurate, thus having practical application value.

Susceptibility assessment is an effective tool for landslide disaster prevention and control. Machine learning algorithms can effectively reduce the subjectivity of determining the weights of factors in disaster risk assessment, which has become a hot spot in current research. We selected maximum monthly rainfall during the year, slope, FVC, distance from rivers, distance from faults and vegetation type as the influencing factors for landslide susceptibility assessment, and used BP neural network and random forest algorithms to evaluate the susceptibility of landslide hazards in the Qingyijiang river basin, and compared the evaluation results. The results show that the fitting results of the random forest algorithm are better. According to the fitting results of the random forest algorithm, the landslide hazard in the Qingyijiang river basin is closely related to the distance from the fault and the distance from the river, and about half of the area (48.75%) in the basin is at risk of small landslides, and most of them are distributed near the river.

The national mine remote sensing monitoring is an important basic national condition investigation work. Using remote sensing data in July 2022, through remote sensing interpretation and field verification, the remote sensing monitoring of open-pit mines in China was completed. The open-pit mining land in China is 993.4 thousand hm², accounting for 1.04‰ of the national land area, which is mainly distributed in Xinjiang Uygur Autonomous region, Inner Mongolia Autonomous region, Qinghai Province, Shanxi Province, Yunnan Province. Among them, the ecological rehabilitation rate of land damaged by mining is 9.57%. The area of open-pit mining land of chemical raw materials non-metallic, building materials and other non-metallic minerals and energy minerals is more larger than other minerals. The basic farmland is occupied by open-pit mining land , which is 11686.64 hm², accounting for 1.18% of the total area; the national open-pit mining land located within the ecological red lines is 7410.62 hm², accounting for 8.39 % of the total area. According to the above results, this paper preliminarily analyzes the current situation of mining damaged land distribution, ecological rehabilitation and existing problems in open-pit mines in China, and puts forward corresponding countermeasures and suggestions.

The inherent uncertainties of input parameters in satellite may affect the inversion accuracy of the column-averaged dryair mole fractions of carbon dioxide (XCO₂). Regarding the design specifications of the next-generation carbon satellite (Tansat-2), a sensitivity analysis is conducted on seven input parameters that may affect the XCO₂ inversion. The paper employs the SCIATRAN model to retrieve XCO₂ from short-wave infrared spectra with central the wavelengths at 0.76 μm and 1.61 μm. The results indicate that, considering the uncertainties in measuring these parameters, the parameters influencing XCO₂ inversion are ranked in the following order: the pressure, oxygen, temperature, water vapor, surface elevation, surface albedo, and ozone. To achieve the XCO₂ inversion accuracy within the range of 0.3%-0.5%, it is necessary to ensure the temperature uncertainty of approximately 0.4 K-0.8 K, the pressure uncertainty of approximately 0.2%-0.4%, the water vapor uncertainty of approximately 7%-14%, the oxygen uncertainty of approximately 0.25%- 0.5%, along with the surface albedo uncertainty of approximately 0.2-0.4, and the surface elevation uncertainty of approximately 22.5 m-45 m.

Point cloud filtering is a very important link in Lidar data processing, that is, separating ground points and non-ground points in the point cloud, which lays a foundation for subsequent data processing. The filtering accuracy will directly affect the accuracy of the final DEM. Based on the "rotation invariance" of point cloud, this paper firstly uses Euler transformation, axis angle transformation and homogeneous transformation to flatten the target point cloud. Then the mainstream filtering methods such as cloth simulation filtering algorithm, statistical filtering method, bilateral filtering and slope filtering are used to filter the point clouds before and after flattening, and the filtering results are evaluated by ISRIS index. Finally, the corresponding DEM model is constructed through the filtering results, and the overall error of DEM constructed by each filtering algorithm is quantitatively analyzed through the evaluation indexes of DEM average error, median error and maximum error, so as to study the influence of point cloud flattening on point cloud filtering.

High temperature stability is an important parameter in thermal design of space cameras, because high resolution space cameras need highly stable thermal system. However, severe orbital thermal environment and complex satellite attitudes bring big challenges for thermal control of space camera. Traditional thermal control method can not meet high stability. Therefore, a novel thermal control method of high temperature stability is proposed to deal with the problem. Indirect radiation thermal control and differential temperature control method is applied for the structure of the camera, which significantly decreases temperature fluctuation and disturbance of external thermal environment. The in-orbit temperature data indicate that the temperature fluctuation of structures of the camera is better than ±0.2 °C and the mirrors are better than ±0.1 °C. The orbital temperature performance demonstrates that the thermal control method of high temperature stability for the high resolution space camera is reasonable and feasible and the thermal system satisfies the high stability demand. The novel method solves the demand of high thermal stability under harsh environment, and plays an important role in enhancing space camera resolution.

Color distortion of remote sensing image is a common problem, and the color distortion of remote sensing image is not conducive to its subsequent interpretation. The existing color correction methods have problems such as poor correction effect and distortion of texture details. To solve these problems, this paper proposes an improved CAP-VSTNet style migration network. The proposed structure-color loss function can make the texture and color of the result close to the target image. The AdaIN adaptive normalization layer is introduced to improve the correction effect without increasing the calculation amount. Experiments on public datasets show that the proposed method outperforms other comparison methods in mean square error, peak signal-to-noise ratio, structural similarity and IL-NIQE(a completely blind image quality assessment method).

Due to the complex background, small proportion of target pixels, and large scale difference in remote sensing images, this paper designs a Attention Fusion and Context Enhancement Network (AFCE-Net) to improve the detection accuracy of YOLOv5 for large and small targets in remote sensing images. Firstly, the C3 module is replaced with Transformer Encoder module in large target detection to capture the target and surrounding environment information by improving global attention, adds SimAM module to reassign channel and spatial weights and Context Enhancement module to suppress background noise and expand the perceptual field. Secondly, Another Context Enhancement module is designed to obtain feature information focusing on small targets in the feature fusion stage of small target detection and then the attention fusion module filters the network channel weights. Finally, the localization loss is replaced by SIoU to improve the target localization accuracy. In this paper, the experiments results on DOTA and DIOR datasets are 5.81% and 3.03% higher than the benchmark algorithm, respectively, and the FPS reaches 135.14 frame/s, which provides a reference for the research of remote sensing image object detection.

This paper uses surface temperature observation data from July to August 2022 in Sichuan to assess the applicability of the National Meteorological Information Center's 0.01° resolution hourly temperature real-time product in extreme high-temperature weather in Sichuan for 2022. The evaluation covers hourly temperature, daily extreme values and temperature ranges, and high-temperature days. The following conclusions are drawn: (1) The product effectively characterizes hourly temperature, with the absolute value of the mean error ≤ 0.15°C and a correlation coefficient ≥ 0.98. Performance in the Sichuan Basin surpasses that in complex terrain areas such as the Western Sichuan Plateau, Liangshan, and Panzhihua (called Panxi). In conditions of high-temperature weather (temperature ≥ 35 °C), accuracy peaks at 35 °C ~ 36 °C, with underestimation becoming more pronounced as temperature increases.; (2) The products exhibit underestimation in the daily maximum temperature and daily temperature range, but the absolute mean error value in each region is ≤ 1 °C. The daily range of products is most accurate between 400 ~ 2000 m above sea level. The daily minimum temperature is slightly overestimated, with stations featuring large errors predominantly in the southeast of the basin; (3) The products underestimate the number of high-temperature days, with the absolute value of the mean error in each region being ≤ 3 days. Stations with substantial errors are primarily distributed in the western and northern parts of the basin and the northern part of Liangshan. Areas with more high-temperature days exhibit smaller errors.

In the past few years, transformers have attracted the attention of many scholars due to their excellent performance in fields such as natural language and machine vision. However, the transformer's self-attention mechanism makes it difficult to extract long-distance contextual information and is difficult to apply to semantic segmentation tasks with high-resolution remote sensing images. Therefore, we propose a bionic eye movement transformer model with visual perception capabilities for cross-domain semantic segmentation of high-resolution remote sensing images. Specifically, we designed three different bionic eye movement attention models to enhance the transformer's ability to extract long-distance contextual features. Experiments conducted on two international standard remote sensing image data sets show that our proposed bionic eye movement transformer has excellent cross-domain semantic segmentation capabilities.

Aiming at the problem that it is difficult to obtain the elevation of the target in the point cloud data of inclined photogrammetry, an automatic extraction method of the elevation of the center point of the target based on the K-Means clustering algorithm and the minimum root-mean-square error is proposed. Firstly, the point cloud of the target area is determined based on the known plane coordinates and divided into elevation intervals, and the elevation intervals are rejected according to different thresholds to achieve the denoising effect; secondly, the ground point cloud and target point cloud are distinguished based on the K-Means clustering algorithm; and finally, the elevation of target center point is determined by the minimum root mean square error from the target point cloud to the center point elevation. The point cloud data generated by UAV photogrammetry is used to carry out several experiments on the method of this paper, and the experimental results show that this algorithm can realize the rapid and automatic extraction of the elevation of the center point of the target, and the difference between the results and the manual extraction is better than 2cm, and it has a certain role in improving the recognition accuracy and work efficiency.

In order to explore the spatial and temporal changes of land use in Yangzonghai basin of Yunnan Province, this paper used the data of 2009 and 2019 to analyze the spatial and temporal changes of land use in the study area, which based on the dynamic variation degree of land use, transfer matrix, Sankey diagram, standard deviation ellipse and center of gravity change, to achieve the goal of promoting the effective protection, rational utilization and sustainable development of the ecological environment of the lake basin. The results showed that: 1) During the period from 2009 to 2019, garden plot, forest land, construction land, water and the irrigation works land increased by 646.89 hm², 2661.17 hm², 492.72 hm² and 98.30 hm² respectively, while cultivated land, grass land and other land areas decreased by 1631.27 hm², 1677.81 hm² and 590.00 hm² respectively. 2) From the land use transfer matrix and sankey diagram, it could be seen that all kinds of land types increased or decreased during 2009 to 2019, in which cultivated land mainly flowed to forest land, construction land and garden plot, forest land mainly flowed to cultivated land, grass land and construction land, and grass land mainly flowed to forest land and cultivated land. 3) From the changes in standard deviation ellipse and center of gravity during 2009 to 2019, the dynamic changes of garden plot, grass land and other land were more intense, followed by construction land and forest land, and the dynamic changes of cultivated land, water and the irrigation works land were relatively small. Meanwhile, the center of gravity of other land and grass land shifted greatly, followed by construction land, forest land and garden plot, and the center of gravity of cultivated land and water and the irrigation works land changed little. In order to provide data support and decision-making basis for land use and ecological security system construction, the changes of land use quantity, transfer matrix and center of gravity were studied in Yangzonghai basin from 2009 to 2019. It aims to effectively promote the effective protection, rational utilization and sustainable development of ecological environment in lake basin.

This study investigated the spatial changes in reed distribution in Baiyangdian, using high-resolution remote sensing imagery. It revealed a substantial alteration in the distribution of reeds within Baiyangdian from 2016 to 2022. The reeds have converted with five types of land cover, including water, cropland, orchards, woodlands and building land. The reed area increased by 14.3% to 103.7 km² in 2022 compared to 2016. The expansion primarily occurred in the shallow waters of Zaozhadian in the western part of Baiyangdian. The expansion of reed area is attributed to the continuous water replenishment in order to preserve the ecology and the restoration of cultivated land to wetland ecosystem. After the removal of the dyke, the 26.55 km² of reed swamp on the dyke became open water. It increases the hydraulic connectivity of Baiyangdian lake and accelerates the restoration of ecological function of Baiyangdian Lake.

In the context of global climate change and accelerated urbanisation, the urban heat island effect is becoming increasingly prominent, with significant impacts on the human habitat and ecosystems. In order to study the urban thermal-environmental relationship, the study designed a research method combining remote sensing technology. In the process, remote sensing satellites were selected as the data source for the study of the thermal-environmental relationship between urban green belts and blue-green spaces, and the processing methods of different categories of data were designed, after which the research index system containing the morphological factors of green belts and the morphological factors of blue-green spaces was constructed, and the computation methods of each factor were designed, so as to carry out the study of the thermal-environmental relationship through the index system. The experimental results show that the research method measured that between 2016 and 2017, the largest change in the surface temperature contribution rate of Shanghai is the green belt area factor, and the surface temperature contribution rate increased from 24.25% to 35.23%. It shows that the research method can effectively carry out the study of the correlation between urban green belts and the thermal environment of blue-green space.

Targeting the deficiencies of the full prediction mode and simplified prediction mode provided by the CCSDS 123.0-B-2 compression standard of the International Committee for Space Data Systems, this paper proposed a novel hybrid prediction mode combining the full prediction mode and the simplified prediction mode, and optimizes the local sum and local difference calculation parts of the predictor. The proposed hybrid prediction mode reduces the prediction complexity of edge pixels while ensuring the compression effect, which is favorable for deployment on FPGA hardware systems. The improved compression algorithm was tested and analyzed using a 25-channel hyperspectral image captured by a self-made snapshot mosaic hyperspectral imager developed by our research team. The experimental results show that the improved algorithm can realize the lossless compression of snapshot mosaic hyperspectral images (SSM HSI) with the compression effect of 2.67 ~ 4.32 bits/sample and the compression ratio of 1.86 ~ 3.01. The hybrid prediction mode can be applied in the fields of real-time imaging of unmanned aircraft-carried hyperspectral and wireless transmission of large-capacity hyperspectral images.

The development of cities is influenced by geographical and geological conditions. Taking Jiangxi in China for example, this paper compared the geographical conditions of the urban areas between 1984 and 2020, and predicted the areas suitable for future urbanization in the study area through Random Forest (RF) modeling. By identifying the changes of the average elevation, and slope of cities, we understand that cities tend to expand toward the plains but in the valleys and basins, cities are also forced to develop toward high elevation and high slope belts.

The built-up areas of 194 cities and towns (> 10 000 in population) in 2020 were taken as positive samples, and the buffer ranges of 400 non-urban random points generated in areas where slope > 10° and distance to river > 10 km as negative samples. Positive and negative samples were divided into a training set and a validation set in a ratio of 7:3 by random selection, and 300 classification trees were set within the RF model to classify the whole study area. The results were verified through the validation set and an overall accuracy (OA) of 94.88% and a Kappa Coefficient (KC) of 89.03% were obtained. This means that the prediction of the optimal urbanization candidate areas seems reliable. These areas are distributed in basins and valleys around the Poyang Lake, along the rivers Yangtze, Ganjiang, Xinjiang, Yuanhe, Jinjiang and Fuhe, and those along the deep and large fault zones are the most suitable ones for urban development.

Semantic segmentation is an emerging topic in the processing of high-resolution Synthetic Aperture Radar (SAR) images of oil spill. Due to the complexity of information in SAR images of oil spill, the efficacy of most semantic segmentation methods for images of oil spill hinges on the quantity of labels; however, annotating images demands considerable time and manual effort. To address this issue, we propose a semi-supervised semantic segmentation approach based on pseudolabeling. This method amalgamates multiple functions to optimize the model. Specifically, we introduce DeepLab v3 as the backbone network, combined with a dual-student framework, to enhance the model's resilience to image noise through consistency and stability constraints. Moreover, we employ per-pixel contrastive learning to explore the global semantic relationships between sample pixels. By using a truncated Gaussian function to weight pseudo-labels, we reduce the entropy in pseudo-labels. Experimental results on public datasets demonstrate that our method can fully comprehend the latent information in unlabeled images and reduce the number of labeled images required while ensuring accuracy.

The relationships between urbanization and landscape metrics have been extensively documented. Few studies, however, have examined whether such relationships are affected by scale effects. This study aims to (1) study the variations in the relationships among different selected transects; and (2) study the impacts of scale effects on landscape diversity and other landscape metrics in an entire area. To reach such goals, land uses and land covers in the Shunyi District of Beijing were classified based on Landsat images, then the landscape metrics were calculated both at the regional scale and at four transect scales using Fragstats and patch analysis. The results show that transect selection and spatial scales did have some impacts on the relationships between the landscape metrics and urbanization. The relationships between urbanization and the landscape metrics differed across different scales, as well as different transects. Hence, the threshold of scale should be carefully considered since it is site-dependent. A suitable scale found in a particular area may not be suitable for another area. These results suggest that the proper selection of scales and transects is crucial to ensure the quality of ecological urbanization assessments.

Urban heat island effect has become increasingly severe, causing a series of negative impacts on people's physical and mental health. Urban parks have a significant cold island effect, which can effectively alleviate the thermal environment. From the perspectives of the physiological and psychological needs of urban residents, this study establishes four indicator criteria: surface temperature (LST), heat island patch area (S), population density (PD), and accessibility (AI), with aiming for constructing a quantitative evaluation system for cold island demand; Indicators and weights are calculated through research methods such as surface temperature retrieval, mean standard deviation method, kernel density analysis, improved two-step moving search method, subjective and objective weighting method. This study selects 59 urban parks within the main urban area of Wuhan and uses the above evaluation system to calculate the cold island demand index, in order to generate a map of Wuhan park’s cold island demand distribution. The results show the following: (1) the total cold island demand urgent area in Wuhan’s main urban area is 459.4 km2, among which the most urgently needed area (demand level III) is 78.2 km2, accounting for 11.1% of the research area. This demand area has obvious spatial characteristics, particularly scattered multicore and spatial distribution along the main road. (2) The cold island demand in Wuhan, particularly in areas classified under demand level III, can be attributed to three main reasons: lack of walking accessibility, high population density, and lack of ecological coverage. Finally, on the basis of three categories, the most typical and urgent points are selected and then optimization measures are proposed based on the actual nature of land planning and ecological conditions, which can contribute to the alleviation of urban thermal environment in Wuhan.

The Dynamic Resolution Detection System of aerial camera is used to simulate the motion state of ground object relative to aircraft during flight in laboratory. The virtual instrument technology and motion control technology of dynamic resolution detection system based on LABVIEW are studied, and the motion curve of dynamic object simulation is designed according to the image motion velocity equation of aerial camera, the traditional PID control algorithm is improved by using friction fluctuation compensation algorithm, and the uniform motion curve of high precision linear and motor control is completed.

Wheat is one of the most important cereal crops in the world. Efficiently and accurately extracting spatial distribution information of winter wheat from high-resolution remote sensing images has become crucial research and application fields in agricultural monitoring. However, the complex texture features and spatial heterogeneity in high-resolution imagery often require deep learning models to be trained with a substantial number of samples to achieve the desired outcomes. In response to the challenges posed by insufficient sample volumes, which impair the accuracy of recognition in the field of extracting spatial distribution information of winter wheat, this paper introduces a new methodology that synergizes deep learning neural networks with machine learning technologies to address these issues. The new process first applied a Grey-Level Co-occurrence Matrix (GLCM) to enhance the image texture. It then utilized a random forest (RF) model to generate a confidence map of winter wheat distribution. Next, the method recommended a supervised correction to this map. It integrated the corrected map into an improved Segformer model as a priori knowledge to increase the weight of the high-confidence regions to achieve high extraction accuracy and robustness. This study selected Shouguang City, Shandong Province, as the case study and the SuperView-1 satellite data as the input image. This study compared the proposed GLCM-RF-Segformer model with ResUnet, DeepLabV3Plus, Hrnet, FCN8S, and RF to validate the experiment's effectiveness. The experimental results have shown that the new model has achieved average metrics of 91.387% for Precision, 94.467% for Recall, 94.180% for Overall Accuracy (OA), 87.053% for mean Intersection over Union (mIoU), 85.087% for Kappa coefficient, and 92.630% for F1-Score. These accuracy statistics surpass those of the five other comparative models, highlighting the model's high accuracy and robustness in extracting the spatial distribution of winter wheat with limited samples.

The Yellow Sea green tide has occurred every year on a large scale since 2007, causing harm to marine ecology, fisheries, tourism, and so on. Affected areas usually formulate green tide containment and removal programs based on satellite images green tide monitoring information. However, due to the influence of clouds and fog and the width of satellite images, it is difficult for one satellite image to monitor the green tide comprehensively, so it is necessary to fuse green tide monitoring information from multi-source satellite images to obtain comprehensive green tide information. However, we face the problems of different scope, time, and accuracy when fusing green tide multi-source monitoring information. Therefore, this paper developed a fusion module based on resampling technology, area refinement methods, green tide drift/tracking technique, and spatial overlay analysis and integration technique. A case study found that the fusion module generated a spatially and temporally uniform green tide fusion product with comprehensive coverage, and the product greatly reduced the number of green tide points, which effectively reduced the amount of drift prediction calculations, improving the prediction efficiency and thematic mapping fluency. It can be concluded that, comparing the single satellite monitoring information, the green tide multi-source fusion product can better provide technical support for Green Tide disaster prevention and mitigation emergency decision-making.

With the purpose of better adapt to different types of complex terrain, this paper proposes an experimental scheme for ground image and LiDAR point cloud decision-level fusion block filtering and DEM generation. Firstly, the ground image and point cloud are matched through homologous feature points linear transformation to obtain more spectral texture information from the image. Then, based on decision-level fusion, the original point cloud is segmented into several independent blocks. The IPTD filtering algorithm is improved and optimized by using a search method according to the different multi-dimensional detail features of each block region. Finally, the filtered total ground points are interpolated to obtain a DEM model. The results show that compared with other overall filtering schemes, the DEM generated by this algorithm has the highest accuracy with an average absolute error and root mean square error of 0.100 meters and 0.157 meters respectively, greatly optimizing the generation effect of DEM.

Soil moisture is an essential soil parameter affecting energy transfer and surface heat exchange. Accurately acquiring surface soil moisture is important for guiding agricultural production, understanding the global water cycle, and studying climate change. This paper conducted an inversion study of soil moisture in the Shandian River Basin based on the change detection method (CD) using Sentinel-1A and MODIS time series data. It was found that the traditional CD has a mismatch problem between the amount of change in soil moisture and the amount of change in radar backscattering coefficient, which reduces the accuracy of soil moisture inversion. This paper proposes an improved CD that corrects for soil moisture variations and radar backscattering coefficient variations using the normalized difference vegetation index (NDVI) piecewise function. Compared with the traditional CD and the Alpha model, the improved CD proposed in this paper has better applicability in the vegetation cover area and improves the accuracy of the soil moisture inversion, with a Pearson correlation coefficient of 0.813, a coefficient of determination of 0.645, and a root mean square error (RMSE) of 0.046 cm³ /cm³ . The improved CD does not require complex parameters and has the potential for large-scale application. Using the improved CD to invert soil moisture, the inversion of soil moisture value can capture the changing trend of measured soil moisture value in a time series.

Based on sensor VIIRS, this article establishes a nonlinear split-window model suitable for calculating vegetation reflectivity of the mid-infrared channel M12 of VIIRS. The model is based on the split-window algorithm and the idea of using two mid-infrared channels to invert the reflectivity of one channel. Analysis shows that the uncertainty caused by various error sources to the model is only 0.82%, so it is believed that this model is suitable for calculating the vegetation reflectivity of mid-infrared channel M12 of VIIRS. This article uses the split-window model to calculate the mid-infrared reflectivity of multiple vegetation research sites in China and its surrounding areas of VIIRS image. The reflectivity of dense vegetation are relatively stable and within a small range, while the reflectivity of low-sparse vegetation are large and have significant differences. These values are consistent with actual data, indicating that the mid-infrared reflectivity obtained by this research model has potential in evaluating large-scale forest volume.

The vector information of mountainous roads is great significance, including regional economics, emergency response, disaster management, among others. Nevertheless, there are irregular deformations in existing vector data, coupled with incomplete extraction results for mountainous roadways. These issues pose challenges in establishing a matching model between road extraction results and existing vector data, significantly reducing the efficiency of automated updates for the current vector data. Targeting roads as a distinct artificial feature and leveraging their topological connectivity, we propose a vector data updating method based on matching point pair grouping. Firstly, we employ the SDUNet model to extract mountainous road and utilize the Superglue model to match road binary images with vector binary images, resulting in sets of initial matching point pairs. Subsequently, we propose a road line extraction model that groups matching point pairs, eliminates misaligned pairs based on road topological connectivity, and establishes a mapping relationship between road lines and vector lines. Ultimately, through road line optimization and vector line optimization methods, we accomplish the automatic updating of vector data for mountainous roadways. Through a comprehensive analysis of over 2600 vector data, our method demonstrates substantial advantages in terms of both F1 and IoU compared to alternative methods.

The Yellow River Delta wetlands are rich in vegetation species and the ecosystem is sensitive and fragile, so it is of great practical significance to accurately extract the information of the vegetation in the Yellow River Delta wetlands and analyze its spatial and temporal changes. Based on Sentinel-2A multispectral images and ZY1-02D hyperspectral images, this study addresses the problem of high spectral resolution and low spatial resolution of hyperspectral images, and adopts the Gram-Schmidt fusion method to fuse the two kinds of images, and uses the Random Forest Classification method to complete the classification of the typical vegetation of the Yellow River Delta region, and compares the classification accuracies of the three kinds of images. The study shows that: (1) Gram-Schmidt is more effective for image fusion, and the fusion of fused hyperspectral images based on the NIR band of Sentinel-2A images has the best fusion effect, which maximally retains the original spectral reflectance values, and at the same time improves the spatial resolution. (2) The classification accuracy of the fused hyperspectral images is high, with an overall classification accuracy of 85.62%, which is higher than that of single hyperspectral images and multispectral images.

Soil moisture is a crucial parameter for agricultural production, water resources management, global climate, et al. Vegetation and roughness, as two key influencing factors, are the focus of research. Synthetic aperture radar is an important tool of obtaining moisture parameter. This article conducts research on the following aspects. First, the overall backscattering coefficient is obtained based on Sentinel-1 radar data, and the backscattering coefficient of vegetation is separated using the water cloud model based on three Landsat-8 vegetation indices (i.e., NDVI, NDWI, and MSAVI). Then, the Dobson model combined with the advanced integral equation model (AIEM) was used to establish a backscattering coefficient table lacking surface roughness, and the minimum cost function was used to obtain the optimal roughness parameters. Finally, the least squares method is used to determine the coefficients of the inverted soil moisture empirical equation. The experimental results show that in the plateau area, the model inversion results are in good agreement with the ground measurement results. Among them, the inversion results using the normalized water index (NDWI2) input water cloud model combined with the optimal roughness are the best. The combined coefficient reaches 0.8402, and the root mean square error is 0.02732cm³/cm³.

The Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2) is the most advanced laser altimetry satellite at present, emitting a laser wavelength of 532 nm, which not only realizes the observation of land areas, but also provides a new solution for the shallow sea areas around the islands and reefs that are difficult to measure. However, due to the large number of noisy photons contained in the satellite raw data, it brings difficulties to the subsequent data processing work. In order to solve this problem, this paper uses two photon point cloud denoising algorithms proposed independently: one is the joint denoising algorithm of DBSCAN and OPTICS; the other is the denoising algorithm based on BP neural network, and the validity of the two methods in the shallow sea area around the islands is verified by comparing them with the experiments in a certain island area in the South China Sea.

In order to research the selection of leg laying zone under the offshore operation of deep-sea exploration platform, considering the need for real-time collection of information such as terrain elevation and small scale terrain roughness, the submarine terrain feature extraction technology based on gradient and variance of Digital Elevation Model (DEM) is used to fuzzy process and reconstruct the seabed terrain, and generate the point cloud map of the seabed terrain. The results show that the point cloud map information can reflect the roughness of the terrain in this zone, and also reflect the slope information in this zone to a certain extent. This method can provide a rough assessment of the topographic characteristics of the leg laying zone and determine the candidate leg laying zone for the deep-sea exploration platform.

On the task of remote sensing image target detection, aerial images have detection difficulties in terms of small data volume, small imaging, and few pixel features, while traditional target detection algorithms have poor real-time and accuracy, and high false and missed detection rates for small targets. Targeting the aforementioned issues, a remote sensing target detection algorithm based on the improved YOLOV7 is proposed. First, on the basis of the traditional YOLOV7 algorithm, the BiFormer attention mechanism is added to the YOLOV7 backbone network, and BiFormer proposes a new dynamic sparse attention method through two-layer routing to achieve more flexible computational allocation with content-awareness, which improves the algorithm's feature extraction capability; second, Wise IoU is adopted as the bounding frame regression loss function instead of the original CIoU, which improves the convergence speed of the algorithm. The experimental results on the DOTA dataset show that the accuracy of the improved YOLOV7 algorithm reaches 76.8%, and the mAP is improved by 1.7%. To a certain extent, the detection accuracy of YOLOV7 algorithm on remote sensing images is improved.

This paper aims to improve the efficiency of remote sensing image processing. The reflectance retrieval module developed based on IDL language uses the images obtained by the new DMCIII digital aerial survey instrument, introduces the parameters of the DMCIII camera and various modules of the whole process of multi-thread processing, and analyzes the accuracy by combining the field data collected. It proves the operability and practicability of multi-threaded processing in the whole process.

Traditional measurement methods are faced with problems such as large labor consumption and difficult implementation in the process of surveying and mapping in mountainous areas. To solve these difficulties, this paper studies the feasibility of using UAV photography technology to carry out the completion measurement of mountain land increase and decrease project under the application background of the acceptance measurement of rural demolition land reclamation into cultivated land in a mountain land increase and decrease project. The implementation results show that the UAV photogrammetry is convenient and efficient in the project of land increase and decrease connection in mountainous areas. The measurement accuracy meets the requirements of the specification and achieves the effect of efficiency and quality guarantee.

In order to achieve stronger penetration and higher imaging resolution of future spaceborne radars, a radar system with lower frequency band and wider bandwidth must be designed. However, such designs will lead to an increase in the impact of the ionosphere, and traditional ionospheric research methods may no longer be applicable, and new analysis methods are urgently needed. This study conducts a detailed quantitative analysis of the performance of future spaceborne P-band ultra-wideband radar under the influence of the ionosphere. In the study, an in-depth evaluation of the expansion orders of the ionospheric assessment model, including Taylor expansion and Legendre orthogonal polynomial expansion, was carried out. Analysis shows that in the application of ultra-wideband radar, whether Taylor expansion or Legendre orthogonal polynomial expansion is used, the existing third-order coefficient expansion method cannot meet the accuracy requirements of ionospheric impact assessment. Correspondingly, the Taylor expansion needs to be extended to fifth-order terms, and the Legendre orthogonal polynomial expansion needs to be extended to fourth-order terms. In addition, through the fitting analysis of the point target diffusion model under the influence of the ionosphere, the quantitative relationship between the radar resolution of the P-band ultra-wideband radar and various ionospheric parameters was clarified. The results of this study can provide relevant reference basis for future high-performance radar design.

Aiming at the problem that ship targets in synthetic aperture radar (SAR) images in challenging environments are easily affected by background clutter, resulting in false positives and missed detections, in SAR image ship detection.a SAR ship detection method in challenging environments based on improved VarifocalNet is proposed. Firstly, ConvNeXt is incorporated as the core network to enrich the receptive field of the model to improve the feature extraction ability of the network of multiple sized ships in complex backgrounds. Secondly, replacing the nearest neighbor interpolation upsampling module with the flow alignment module can compensate for the challenge of feature alignment within the feature pyramid and bolstering multi-scale feature fusion proficiency. Finally, introducing MPDIoU loss function is designed to expedite model convergence and refine target localization precision. The empirical outcomes demonstrate that the detection accuracy of the refined model on the SSDD and SAR-Ship-Dataset datasets has been enhanced by 4.3% and 5.9%, respectively, in comparison to the original VarifocalNet model. This substantiates that the algorithm can perform high-quality ship detection on SAR complex background images.

The spatial-temporal distribution and heterogeneity of carbon emissions are important issues in ecological environmental protection and climate change monitoring. Based on DMSP/OLS night light remote sensing data and socioeconomic data, this paper uses machine learning methods to study the carbon dioxide emissions of 281 prefecture-level cities in China from 2005 to 2020, and reveals the characteristics of spatial-temporal distribution, growth trend and agglomeration of carbon emissions. The results show that in recent years, Chinese carbon emissions show an increasing trend, and regional carbon emissions have obvious characteristics of spatial agglomeration and unbalanced distribution. From the spatial dimension, the pattern of carbon emissions gradually evolves to "high in the east and low in the west" and "high in the north and low in the south". This study can provide scientific reference for carbon emission estimation and prediction, energy conservation and ecological environmental protection.

Modern manufacturing industry has a strong measurement demand of specular surfaces. Fringe reflection method has the advantages of simple structure, low cost, high accuracy, and is becoming more and more popular for 3D measurement of these surfaces. Most of previous methods achieve the fringe reflection measurement using a single camera, and suffer the problem of gradient-height ambiguity. Here, a binocular fringe reflectometry method is proposed with the surface normal is obtained through normal vector consistency of the measured surface. To enhance the computation efficiency, a polar constraint and bisection method are also developed to get the matching points on the two camera images. After that the normal vector and the gradient distribution can be calculated. The 3D measurement result can be obtained according to integral reconstruction algorithm subsequently. To verify the effectiveness of the method, measurements were performed on a spherical mirror. Compared with the result from interferometer, the deviation of peak to valley value is 314.3 nm, and the root mean square value is 52.8 nm, which indicates high accuracy of the method.

Antennal control system software in telemetry ground station adopts C# language to realize switching of various tracking modes and monitoring of working state. The software running on the controller and jade on the data of multiple interfaces on the base of the original data storage, analysis, real-time calculating target azimuth and pitch information output to the servo control system, puts forward all kinds of control functions and strategies of telemetry antenna, complete antenna precision of target tracking, real-time display the running state of the system software at the same time, and timely response to abnormal situation. The practical application results show that the software has high reliability and expansibility, and can fully meet the needs of real-time control system, and has good practicability and development prospects.

Optical satellite imaging is easy to be affected by weather, and imaging in the area with high cloud cover will result in poor quality or even complete unavailability of satellite imaging products. Addressing the lack of effective methods for evaluating data acquisition in large-area, multi-satellite joint observations, this paper utilizes the VIRR global cloud daily product data from the FY-3A/3C meteorological satellites spanning 2010 to 2020. This data is analyzed to extract cloud amount information, subsequently used to obtain the regional clear sky index sequence. The SARIMA time series prediction model is then employed to forecast the clear sky index of the observation area effectively. This forecast is combined with satellite transit information to establish a data acquisition validity prediction model. Utilizing the predicted monthly clear sky index for Beijing in 2021, alongside the satellite transit information for each period, the model predicts data validity. The average discrepancy between the predicted and actual values is 0.36. This data validity prediction model can identify optimal observation periods for multi-satellite joint coverage of large-area targets and provide effective data support for coverage prediction within the observation period.

The GF-5 satellite has obtained nationwide hyperspectral images, which will play a dominant role in its widespread geological application. The GF-5 hyperspectral data is featured by integrated images and spectra, and can simultaneously obtain both the location and types of surface altered minerals. This work identified altered minerals and conducted geological application based on GF-5 hyperspectral data in the Yingshan area of Aksu City, Xinjiang. Through radiometric calibration and atmospheric correction of the hyperspectral data, the image reference endmember spectra were acquired. An improved spectral angle mapping method was employed to map altered minerals, and three types of altered minerals limonite, calcite, and dolomite were extracted. Based on previous geological data, this work analyzed the distribution of altered minerals, and determined the relationship between altered minerals and formation lithology. As a result, inaccurate geological boundaries were discovered in previous geological maps. It is inferred that the GF-5 hyperspectral data can effectively extract altered minerals, and is of great significance in stratigraphic division, lithology boundary delineation, and non-metallic mineral resource prediction.

This study underscores the significance of accurately identifying minute objects within satellite imagery. We propose an innovative GAN framework tailored for enhancing the detection of minute objects in such imagery. Our approach primarily emphasizes differentiating these objects from their surroundings by leveraging the core mechanisms of GANs. Moreover, we address the varied spectral characteristics of the background by integrating a similarity constraint into our GANs architecture. This strategy effectively distinguishes between the target object regions and non-object regions, significantly reducing the occurrence of false positives. Additionally, we employ a smoothness regularization technique to preserve the structural integrity of the detected objects. Through extensive experimentation on publicly available remote sensing datasets—spanning mineral, fabric, and vehicle detection—we compare our method quantitatively against conventional small object detection techniques and state-of-the-art GAN-based models. The comparative results consistently showcase the superior performance and robustness of our GAN-based approach in detecting minute objects within remote sensing imagery.

With the rapid development of facility agriculture in China, the distribution of plastic greenhouses is becoming more and more extensive. Plastic greenhouses along transmission channels are easy to float, which is one of the main external breakage hazards of transmission lines, and it is of great significance to carry out remote sensing identification of plastic greenhouses along transmission channels for the management of external breakage hazards. This study proposes an improved Segformer model based on the Segformer decoder. In it, high-level semantic information is fused by the feature fusion module to realize the deep mining of implicit features of multi-scale feature maps. Then, the accurate reduced representation of image spatial details is realized by the inverse convolutional upsampling module. Experiments are carried out in a study area along a transmission corridor in Lianyungang City, Jiangsu Province, China, where the overall accuracy and the intersection and concurrency ratio of the improved Segformer are 0.876 and 0.849, respectively. The ablation experiments demonstrate that both designed modules play a positive role and the maximum accuracy can be achieved by using them jointly. The study analyzes the identification results and the distance to the transmission lines to predict them into high, medium and low risk classes. The study shows that based on high-resolution satellite remote sensing images, it is feasible and effective to use the model to carry out the identification of plastic sheds in the areas along transmission corridors, which can provide a decision-making basis for the comprehensive management of external breakage hazards.

WorldView-3 is one of the highest spatial resolution multispectral commercial satellites currently available on the market, with great potential for applications in land analysis, geological mineral surveys, and disaster assessments. In order to fully utilize the high spatial resolution of WorldView-3 data, fusion processing was performed using Principal Component Analysis (PCA), Gram-Schmidt (G-S), Histogram Contrast Stretching (HCS), and Nearest-Neighbor Diffusion (NNDiffuse) methods to combine the multispectral images with the panchromatic image. Qualitative and quantitative evaluations were conducted on the fused images, and the results showed that in the visible-near infrared (VNIR) wavelength range, the HCS method had the best overall fusion effect, with the closest mean value to the original multispectral image, highest standard deviation, information entropy, and average gradient value. The NNDiffuse method had the highest correlation coefficient and spectral fidelity between the fused image and the original multispectral image. In the shortwave infrared wavelength (SWIR) bands, the HCS fused image had the best spectral and color fidelity, while the NNDiffuse and G-S methods produced fused images with richer information and hierarchy, higher clarity, and more distinct spatial structure and texture features.