Research On Algorithm Of Digital Image Edge Detection And Algorithm Of Distortion Correction

The edge of digital image plays a crucial role in the digital image analysis and processing. But the difficulties of the edge detection of digital image is the contradiction of anti-noise and edge location precision. This paper were studied all kinds of classic edge detection operators respectively, and put forward a new edge detection algorithm which considers both multi-scale morphology and Canny operator.The concrete work of the thesis are as follows:Firstly, this paper introduces the basic principle of digital image edge detection,and introduces the principle of several classic algorithms and the specific implementation methods in detail. I have simulated several classic algorithms,through a comprehensive comparison, the result showed that Canny operator and morphological methods performance better than other classic algorithms in the actual effective.Then, the paper introduces a method of multi-scale morphological edge detection and image fusion based on wavelet transform. On this basis, I propose a new edge detection algorithm which considers both multi-scale morphology and Canny operator.This method do edge detection by using Canny operator or multi-scale morphology assay respectively to original image, then we can get a Canny operator edge detection image and a multi-scale morphology edge detection image. To obtain the final edge image, the two edge images should be fused by the wavelet transform.Finally, the new algorithm is simulated to achieve, and the simulation results is used to compare with the test results of Canny operator and some other traditional morphological edge detection methods. Comparative results show that the new algorithm is more effective than the traditional Canny operator and morphological edge detection.Aiming at the serious barrel distortion problem of wide-angle lens digital image,two Improved Distortion Correction Algorithms are put forward in this paper. The distortion Correction Algorithms in literature are divided into two steps: the first step is changing color image to grayscale, the second step is to correct image distortion,but these methods can’t restore original color information of image. The nonlinear model method and equivalent spherical method in this paper use original data of image. The numerical simulation examples and experiment results are given to illustrate the superiority of the obtained results to those in literature.