Approach to terrain pretreatment for the Yazidang reservoir based on image processing

Abstract

The terrain pretreatment based on image processing in the Yazidang reservoir is presented. The image processing is based on the image stitching and the image edge detection. The SURF algorithm is used to extract feature points of reservoir partial images. To match feature points and stitch reservoir partial images, the wide view reservoir image is stitching using KD-tree algorithm and linear gradient fusion algorithm. The improved mathematical morphology is adopted, and five different shapes and four different scales of structural elements are implemented to obtain precise bank edge of reservoir image. The *.dxf file is obtained by using CAD software to depict reservoir shore contours. It transforms *.dxf into *.kml to extract latitude and longitude information of reservoir bank edge. The satisfied mesh effect is generated by using the Delaunay triangulation algorithm. The result shows that accurate and detailed reservoir terrain can be obtained using the image processing method and the GE software, which provides a prerequisite for the numerical simulation of sedimentation in the Yazidang reservoir.