Research And Application Of Image Edge Detection Algorithm Based On Multi-scale And Multi-direction

Image edge detection is the basic technology of image processing and computer vision.The effect of edge detection has a great impact on image post-processing.With the development of remote sensing technology,more and more remote sensing data are used and applied to various industries.High-resolution remote sensing images contain a lot of noise information.In the process of edge detection of high-resolution remote sensing images,the way to remove noise and resist noise interference is particularly important.The main contents of this paper are as follows:1.The knowledge theory of image edge,image edge detection and edge detection algorithm in high resolution remote sensing image and the knowledge theory of image edge detection algorithm in denoising and anti-noise are studied.In this paper,the problems of typical algorithms in image processing are analyzed,and the possible improvement directions in denoising and anti-noise are summarized.Under the high resolution remote sensing image,the building presents the area structure,the edge characteristic is obvious,but at the same time,the high resolution remote sensing image contains more noise signals.The existing edge detection methods have some shortcomings in denoising and anti-noising.there is also room for optimization and improvement.In addition,the location of the edge is not accurate enough,the detected edge has an intermittent point,as well as missed detection and false detection.2.An improved optimization algorithm based on NSCT is proposed from two aspects of multi-direction de-noising fusion and multi-scale weighted fusion.The Canny operator,wavelet transform and Non-down-sampling profile transform(NSCT)are analyzed,and the principles,characteristics,advantages and disadvantages of these algorithms are introduced respectively.According to the advantages of translation invariance and multi-scale decomposition of NSCT,the optimal design of multi-direction de-noising and multi-scale weighted fusion for NSCT is carried out.Through a large number of simulation data experiments,it is verified that the optimized algorithm has certain advantages over Canny operator,wavelet transform and NSCT in dealing with 10% and 30% noise images,and the feasibility of this algorithm is verified.3.The optimized edge detection algorithm is applied to the online common basic service subsystem.Combined with the online common basic service subsystem,the building edge detection is carried out for high-score satellite remote sensing images and the detection results are analyzed.The main result of this paper is that on the basis of NSCT,the number and direction of decomposition are optimized and improved in de-noising and noise suppression,and applied to the online common basic service subsystem.Through the provision of online image analysis tools and image processing algorithm module to provide users with online image edge detection and other services.