Mining heavy metal ore deposits may lead to an increase in heavy metal element content in surrounding soils, which could pose irreversible harm to the ecological environment and human health. Therefore, analyzing and classifying soils from different mining areas is of great significance and can provide reference for soil management and environmental pollution control. Laser-induced breakdown spectroscopy (LIBS) has gradually become a research hotspot in soil detection due to its fast and pre-treatment-free characteristics. However, traditional LIBS technology has problems such as low sensitivity, high noise, and poor repeatability, which affect its accuracy. Therefore, this paper proposes a soil classification method based on Principal Component Analysis (PCA) of LIBS technology coupled with K-Nearest Neighbor algorithm (KNN). This method first conducts data standardization and PCA pre-processing to eliminate redundant information and improve signal-to-noise ratio. Then, autonomous sampling technology is used to design the KNN machine learning algorithm structure to generate continuous analytical networks for training and testing sets. Finally, the results show that the soil classification accuracy of the PCA-KNN machine learning model can reach 97.531%, proving that the combination of LIBS technology and PCA-KNN can achieve rapid and accurate classification of soils from different mining areas. Therefore, this method has the significance of providing new ideas and methods for soil classification in different regions.
Currently, China is still a major consumer of coal resources. Coal can be used in various fields such as industry and civil use, and can be used for power generation, heating, and building materials. There are many types of coal, each with its unique composition and properties. It has specific requirements for its use in various fields, which make the use of coal more reasonable and important for the sustainable development of the environment and resources. Therefore, the classification research of coal is of great significance. Due to the same component influence among various coals, there are certain challenges for coal classification. Therefore, a laser induced breakdown spectroscopy (LIBS) based on principal component analysis (PCA) combined with convolutional neural network (CNN) method was proposed to classify and recognize coal samples from six different regions. Through laser ablation of coal samples and collection of corresponding data, the data are dimensionalized and standardized, and then the spectral data are classified and trained through PCA-CNN optimization model. The final results indicate that the coal classification accuracy of the PCA-CNN deep learning network model can reach 98.15%. From this result class, it can be seen that laser induced breakdown spectroscopy technology combined with PCA-CNN can achieve rapid and accurate classification of coal samples from different regions, and provide a new coal quality detection data analysis and processing scheme.
The development of heavy metal mining area will pollute the surrounding soil and do great harm to the ecological environment and human health. Soil classification in different mining areas is of great significance to soil management and environmental pollution control. Soil is easily affected by matrix effect due to its complex physical properties and composition. Thus, accurately classifying soils is a challenge. Laser-induced breakdown spectroscopy has developed rapidly in the past two decades. It has been widely used in the detection of various physical samples due to its characteristics of fast analysis speed and no need for sample pretreatment. However, traditional LIBS technology has disadvantages such as low sensitivity, obvious noise and poor repeatability, which affect the accuracy of quantitative analysis. In this paper, a soil classification method based on principal component analysis (PCA) based laser-induced breakdown spectroscopy (LIBS) and random forest (RF) algorithm was proposed, and the standard soil samples from six different mining areas were accurately identified and classified. The final prediction results based on this combination show that the accuracy of soil classification by PCA-RF machine learning model can reach 97.86%. From the aspect of classification accuracy, it can be found that laser-induced breakdown spectroscopy combined with PCA-RF can achieve rapid and accurate classification of soil in different mining areas, which also provides a new method for soil classification in heavy metal mining areas.
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.