A Research On SAR Image Processing Method And Its Applications In Mining Remote Sensing Survey

In recent years, the technology of remote sensing develops rapidly, and it plays an increasingly important role in the national large-scale surveys. However, the capability of optical image acquisition restricts the work of the project process severely due to regional local climate, cloudy, more rain and snow areas. Therefore, SAR image processing technology makes up for the shortcomings in the current methodology. Achieving accurate efficient processing and establishing a basis of the SAR image in the application of remote sensing investigation become urgent need to address .According to summarizing previous results, I make a research on SAR image edge detection, geometric correction, noise filtering, multi-source data fusion systematically,choosing TerrySAR-X image of HongGe mine in Sichuan to analyze its prospect of application in mine Remote Sensing Investigation. I get some main achievements as follows:By comparing a variety of algorithm on SAR image edge detection, this paper improves Canny edge detection operator, and proposes SAR image feature extraction Objects Edge model on the basis of "Canny algorithm - morphological filtering - skeleton extraction" process, it balances the conflict between the SAR image noise and interference and the edge information effectively, and makes a foundation for selecting the control points in the geometric correction.A SAR image geometric correction program which based on the "Feature Objects Edge - RPC model Orthorectification - Perspective contraction of the control point encryption" is proposed in view of analyzing the existing SAR image geometric correction method and its shortcomings. It overcomes not only the difficulties of effectively selecting control points in SAR image,but also the seriousness of the perspective of corrected image contraction region" garland " and increase in the distortion of layerover area. The result demonstrates that this method can improve the overall accuracy substantially.A measure to decrease the Gamma MAP noise and suppress the brightness difference before and after of the slope of based on neighborhood model and iterative filtering is established. A multi-source data fusion means is investigated next. It achieves the enhancement of SAR images and improves the ability to identify SAR image’s objects.In support of the theories and methods, this paper uses TerrySAR-X image of Hongge of Sichuan to survey mining by remote sensing. The geometric positioning accuracy achieves to 3.68m,the object identification reaches 95.04% and the application effect is obvious.