A Research On Image Identification And Edge Detection

This thesis is to study the circle detection and edge detection which play an important role in the image area. Base on nowadays research, I get some achievement. This thesis consists of four chapters:Chapter 1 Give some introduces to the digital image, color space of digital image, and its research field including Image Restoration, Image Compression, Edge Detection, Image Segmentation, Image Enhancement, Image Registration, Image Trasform.Chapter 2 Detect circles from a digital image is very important in digital image process. Randomized Hough transform is a well-known method for circle detection. This chapter give An Efficient Improved Randomized Algorithm for Detecting Circles. Based on the Random Hough Transform, this Algorithm adopts a "window" method to improve the algorithm.Chapter 3 Give Another Efficient Improved Randomized Algorithm for Detecting Circles. This Algorithm applied the "window" method on Teh-Chuan Chen and Kuo-Liang Chungl's improved Random Hough Transform.Chapter 4 Using Hierarchic genetic algorithms to optimize edge operator to get the optimized operator. We design the operator using hierarchic genetic algorithms, then optimize the operator under the least square criterion by training the sample images. The optimized operator has enhanced the repellent ability for noise that can get the edge more exactly under various conditions.