Improved Edge Detection Algorithm of High-Resolution Remote Sensing Images based on Fast Guided Filter

In order to achieve fast and accurate extraction of edge and contour features of remote sensing images, a high-resolution remote sensing image edge detection algorithm based on fast guided filtering is proposed in this paper. The algorithm overcomes the disadvantage of losing a large number of edge information when the traditional Canny operator uses Gauss filter to smooth the image. We use fast guided filter instead of Gauss filter to reduce the loss of image edge information and keep the edge information of the image. Then the improved Sobel operator is used to accurately calculate the gradient and direction of the gradient. Finally, the edge of the remote sensing image is detected by the Canny operator. Compared with the previous edge detection methods, the edge detection contour information is more obvious and the pseudo edge is significantly reduced.