Measurement of edge detection algorithms in clean and noisy environment

The subject of identification edge in images has a wide application in various fields for that it's considered one of the important topics in a digital image processing. There are many algorithms to detect the edge in images, but the performance of these algorithms depends on the type of image, the environment of the image and the threshold value of the edge algorithm. The objective of this paper is to evaluate five algorithms of edge detection which are Roberts, Sobel, Prewitt, LOG, and Canny in multi environments clean and noisy by using several types of original images (binary image, graphic image, high frequency image, low frequency image, median frequency image, and texture image) and then determine the best algorithm. In noisy environment the following noises was used Gaussian, salt and pepper and speckle. It's known that each edge detection algorithm has a threshold value, if the current pixel value is less than the defined threshold in strength, it will be considered an edge pixel. The change rate of the threshold value in all environments is also explained through this study.