Remote Sensing Image Edge Detection Algorithm Based On Region And Contourlet Transform

High spatial resolution remote sensing is used in Each field of the society more and more, so Extraction and recognition of image has become a hotspot of Image processing technology. Edge detection technology is the key link after the image preprocessing and before the image analysis and understanding. Its accuracy can directly affect the image of the subsequent operation. Study on the edge detection is difficult in a question that how to find a balance between the noise and the positioning accuracy, and make the results of edge detection have a high noise-signal ratio. Therefore, researchers have proposed many algorithms of edge detection, but these algorithms have their own deficiencies in practical applications, There is not a uniform and good one in image edge detection algorithm. This paper based on two considerations as the anti-noise and precise positioning of the remote sensing image edge detection, combine various technique such as contourlet transform, region segmentation, image enhancement, edge detection, and propose new edge detection algorithms. The concrete work of the thesis are as follows:(1)A new edge detection algorithm has been proposed combined with contourlet transform and region growing method. The algorithm first uses contourlet filter to Remove the noise of the remote sensing image, after Enhancement and uses region growing method to Combine the homogeneous region segmentation image, the final uses LOG operator pick up the image edge. The algorithm can greatly restrain noise and can provide better edge information.(2) A new edge detection algorithm has been proposed combined with contourlet transform and threshold segmentation. The algorithm uses contourlet filter to Remove the noise of the remote sensing image first, and then enhanced image, next uses Otsu method in threshold segmentation and the traditional LOG operator for adaptive threshold segmentation to extract image edge. This algorithm can reduce noise and reduce the false edge image, improve the positioning accuracy.(3) For the two algorithms mentioned above, this paper Based on the MATLAB platform compiler, and combined with the experimental and comparative analysis of the relevant, Verify the effectiveness of the proposed algorithms.