With hardware and software algorithms development, necessity of methods with accuracy and velocity are more attention. For example edge detection is one of the most important operations in machine vision and the main purpose of edge detection is reducing the volume of data with preserving main structure and original form of images. For these purpose the accuracy of edge detection with retrieval edges by minimum position error rate and losing edges is one of more important approach in recent decades. It is clear those raw images because of no losing many features than JPEG images have better results. In this research consider a simulation neural network program and compare three famous edge detection “Sobel, Prewitt, Canny” with raw images and shows efficiency on results.

Keywords