The Edge Detection Method Research Of The Gray And Color Image

IN the digital age, Images, as a carrier of information, play an important role in ourdaily life.Images contain plenty of information and the edge which is thehigh-frequency part in the images is the basic feature in the image.The edge detectionalgorithm is the important basic of images analysis, pattern recognition and targetdetection.The ideal edge detection algorithm can balance the image noise and detail.However, there is a contradiction between the two, since both the noise and the detail isthe high frequency portion in the image.This paper studies the gray-scale images andcolor images edge detection method and improve the exsiting method.The maininnovation can be carried out as follows:(1) Traditional cellular neural network edge detection deduce the cellular neuralnetwork edge detection template though the ideal edge image and the orginalimage.Then we can calculate the template for imges using LMI.This method can’tobtain the unique template using this method.To solve this problem,I propose that usingGA algorithm can optimize the result. By choosing the appropriate adaptive function,the new algorithm can only get a unique set of templates. Finally, I use the template inthe binary and gray image. By comparing the simulation results, it is found that thisalgorithm can find more accurate and refined edge.(2) Traditional HSI space based noisy color image detection detect the edge usingmulti-scale space image.The edge we obtain using the exsiting algorithm is thick andthere is some pseudo-edge. In order to solve this problem, I propose that usingmulti-structure element to improve. By the simulation comparison, we can find colormorphological edge detection algorithm can detect more thin edge and false edges aregreatly reduced.Then comparging the result in different degree noise, this method Thismethod has good resistance to dry performance.