国产单极化SAR数据在锰渣尾矿库自动化识别中的应用

摘要

To explore the role of Chinese single polarization SAR data in auto-recognition of manganese slag pools and improve recognition accuracy for manganese slag pools coupled with optical data, the authors chose confusing goals with manganese slag pools in the optical image, and analyzed qualitatively their scattering properties and the difference in SAR and optical images. Based on the results of analysis, the authors carried out the research on auto-identification of manganese slag pools. The optical image was segmented using object -oriented approach. The authors selected the typical manganese slag pools and confusing objects in optical image, and calculated the texture characteristics of the samples. The experiment results show that manganese slag pools recognition rules in the SAR image could be established effectively through SAR texture features, and this provides the basis for the realization of the combination of radar data with optical SAR data for automatic identification of manganese slag pools on computer.%为了挖掘国产单极化SAR数据在自动化识别尾矿库中的作用，辅助光学数据以提高锰渣尾矿库的识别精度，针对锰渣尾矿库及其在光学遥感图像中容易混淆的目标，从定性的角度分析了二者的散射特性以及在SAR图像中的差异特点；并开展锰渣尾矿库自动化识别研究；利用面向对象的方法对光学遥感图像进行分割；选取典型锰渣尾矿库及其易混淆目标，进行纹理特性计算。实验结果表明：可以通过纹理特征有效地建立锰渣尾矿库在雷达图像中的识别规则，为实现光学遥感数据和雷达数据相结合的计算机自动识别锰渣尾矿库提供基础。