Shadows in polarized images often interfere with the acquisition and analysis of the polarization state of light. By removing the shadows, these interferences can be eliminated, and the polarization information can be extracted more accurately for subsequent processing. To solve the problem of insufficient illumination and information recovery in the shadow region of polarization images, we propose an attention condition generative adversarial network (GAN) for shadow removal in polarized imaging. The method uses conditional GAN s as its basic framework and introduces attention modules into the generator, which enhances the network’s ability to localize and recognize shadows. At the same time, the polarization shadow removal images from different directions are fused to maximize the generation of polarization shadow-free images. The acquisition of shadow images in four different scenarios uses a polarization camera. The proposed method is compared with the recent method. Experimental results show that the polarization image after shadow removal using our method is closer to the real ground image, and the evaluation index is better than other methods.
Due to the limitation of the detector, the spatial resolution of the polarization image obtained by space-modulated full-polarization computed imaging is low. In general, there are a lot of valuable high-frequency components in low resolution space, but it is difficult to reconstruct high frequency information by deep learning method. Through the research of the previous super-resolution reconstruction method based on deep learning, it is found that it is difficult to obtain better lifting effect only by stacking residual blocks to build a deeper network. A polarization imaging bi-attention recursive residual super-resolution reconstruction method is proposed. In the network structure, bi-attention recursive residual group is used as the deep feature extraction module, and the hybrid bi-convolutional attention module contained in this module is used to adaptively learn image features of different channels and different transformation Spaces in the deep network. In order to provide the concentrated learning efficiency of high-frequency information, jump connections are introduced into the network structure, and the shallow feature extraction and reconstruction of images are completed by a convolution layer and a sub-pixel convolution algorithm respectively. The experimental verification is carried out under the actual imaging system and simulation data and compared with other methods in terms of visual effect and quantitative results. Experimental results verify that the proposed method can effectively reconstruct the contour structure and texture details of the detection target in subjective image quality and is superior to the comparison method in objective evaluation index.
With the development of unmanned aerial vehicle (UAV) in aeronautical monitoring field, the performance requirements are continuously improved, each application scene also puts forward higher and higher requirements for target detection accuracy and speed. The traditional target imaging method is difficult to meet the image quality requirements, and the artificial target recognition method can’t cope with the rapid changes in the detection environment. Combined with the development of deep learning and polarization hyperspectral imaging technology, a ground target detection method based on Faster R-CNN was proposed. We proposed region proposal network (RPN) module for model training. In the target detection phase, the proposed feature map is obtained by pooling operation of interest regions. Finally, we used the proposed feature map to complete the target category classification. Three scale models were used in the experiment, and through polarization hyperspectral camera, the image data of target in different scene conditions was acquired in indoor and outdoor simulation environment for training and validation of models. The experimental results showed that the proposed method could achieve ideal detection accuracy and speed when the ground target was effectively detected.
In order to enhance the restoration quality of Wiener filter, and widen the range of its application, an improvement is made on its basic model, then discuss how adaptive Wiener filter works on motion images, which is based on detecting blur’s direction and depth, and on recursive iterations. As for the process of motion-blurred image of the fast-moving object, experiment indicates an ideal effect can be achieved by this method.
Polarization imaging is another photoelectric imaging detection technology. It has obvious technical advantages in revealing camouflage, penetrating haze, and getting target details. It can gain multiple polarization features images and achieve target detection and recognition through specific polarization information analysis methods such as synthesis and fusion. Because there is a mis-match problem between the polarization features images, polarization image registration performs first. However, existing methods such as mutual information registration and related registration methods are hard to solve the problem of mis-match because of distortion of the polarization imaging lens. In this paper, we present a matching optimization SIFT polarization image registration algorithm found on the standard SIFT registration algorithm. In the sub-matching description, a reversed matching is added, that is, matching in both directions performs to form a symmetrical matching. In the matching set of positive and negative directions, matched feature points pairs satisfying both sets extract. The pair of matching points are only when the pair of feature points are the best matching points. This increases the matching accuracy of feature points and reduces the mismatching rate of descriptions. At the same time, numbers of feature points add in the algorithm using the gray leveling method. Registration experimental results show the registration accuracy of this method is better than the mutual information registration method.
In the target detection process of polarization optics imaging, due to the turbulent effect of the target signal transmitted in the atmosphere and the photoelectric conversion of optical imaging sensors and other factors, Salt-and-Pepper noise which affects the detection accuracy. According to the statistical characteristics of the Salt-and-Pepper noise probability density, a new structure preserved polarization image Salt-and-Pepper noise removal method is proposed. With the new signal sparse representation theory and image inpainting method, only the noise regions is restored by the noise point detecting. In the inpainting process, the structural similarity is considered which can improve the structural information retention ability of polarization image. Numerical simulation results demonstrate the validity of the proposed method both subjectively and objectively.
KEYWORDS: Associative arrays, Principal component analysis, Image denoising, Denoising, Global system for mobile communications, Visualization, Image processing, Inverse problems, Image compression, Image segmentation
To get better denoising results, the prior knowledge of nature images should be taken into account to regularize the ill-posed inverse problem. In this paper, we propose an image denoising algorithm via non-local similar neighbor embedding in sparse domain. Firstly, a local statistical feature, namely histograms of oriented gradients of image patches is used to perform the clustering, and then the whole training data set is partitioned into a set of subsets which have similar local geometric structures and the centroid of each subset is also obtained. Secondly, we apply the principal component analysis (PCA) to learn the compact sub-dictionary for each cluster. Next, through sparse coding over the sub-dictionary and neighborhood selecting, the image patch to be synthesized can be approximated by its top k neighbors. The extensive experimental results validate the effective of the proposed method both in PSNR and visual perception.
KEYWORDS: Polarization, Associative arrays, Image compression, Principal component analysis, Super resolution, Visualization, Lawrencium, Cameras, Chemical species, Imaging systems
The different polarization phase angle (orientation) low-resolution images of the same scene have much redundant and complementary information which can be used to construct a high-resolution image. In this paper, we propose a super-resolution (SR) algorithm via sparse and redundant representation with considering the non-local self-similarity in different polarization orientation images. As the redundant over-complete dictionary has many irrelevant atoms which not only reduce the computational efficiency in sparse coding but also reduce the representation accuracy, we learn a local dictionary by applying the principal component analysis (PCA) technique. For an image patch to be coded, the best fitted sub-dictionary is adaptively selected by an adaptive sparse domain selection strategy. To improve the stability and accuracy of sparse coding, the centralized sparse coding algorithm is used. The extensive experimental results demonstrated that the proposed method can effectively reconstruct the polarization image with edge structure preserved and detailed information obtained in terms of PSNR, SSIM and visual perception.
Polarization imaging provides abundant information of object, i.e. surface roughness, texture, physical and chemical characters. Independently, intensity and polarimetric features give incomplete representations of an object of interest. These representations are complementary, and it is expected that the combination of complementary information will reduce false alarms, improve confidence in target identification, and improve the quality of the scene description. Polarization parameter images include the degree of polarization, the angle of polarization, azimuth angle etc. There are not only strong correlations between polarization parameter images, but also different characters, which gives image fusion challenges, namely, how to find the optimal polarization parameter image to take part in image fusion with intensity image. This paper presents a polarization image fusion method based on choquet fuzzy integral. Using this algorithm the best polarization parameter image and intensity image are fused, and the fusion result is evaluated. The experiments show that this method could automatically select the best polarization parameter images from multi-polarization parameters image, the resulting images can yield more detail and higher contrast, and can reduce the noise effectively. It is conducive to the subsequent target detection.
In this paper we analyse the polarization imaging theory and the commonly process of the polarization imaging detection. Based on this, we summarize our many years’ research work especially in the mechanism, technology and system of the polarization imaging detection technology. Combined with the up-to-date development at home and abroad, this paper discusses many theory and technological problems of polarization imaging detection in detail from the view of the object polarization characteristics, key problem and key technology of polarization imaging detection, polarization imaging detection system and application, etc. The theory and technological problems include object all direction polarization characteristic retrieving, the optical electronic machinery integration designing of the polarization imaging detection system, the high precision polarization information analysis and the polarization image fast processing. Moreover, we point out the possible application direction of the polarization imaging detection technology both in martial and civilian fields. We also summarize the possible future development trend of the polarization imaging detection technology in the field of high spectrum polarization imaging. This paper can provide evident reference and guidance to promote the research and development of the polarization imaging detection technology.
In this paper we propose a polarization image fast fusion approach based on online dictionary learning for sparse non-negative matrix factorization, aiming at improving the efficiency of fusion methods for polarization image based on non-negative matrix factorization. Firstly, all of the polarization parameter images are taken as source data sets for sparse non-negative matrix factorization using online dictionary learning algorithm, so as to extract three feature basis images. Then, after histogram matching, the three feature basis images are mapped into three color channels of IHS color space. Finally, the fused image is achieved via the transform from IHS to RGB color model. Experimental results show that, the proposed method not only has better capacity of color representation capability and effectively pop out detailed information of objects but enhances the running efficiency evidently as well.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.