Dsp Based The Tms320c6416 Adaptive Image Edge Detection

Edge detection is one of the most important steps in the computer vision system. To detect image edges is important significance to image analysis and recognition. The theory and the algorithm of the image edge detection are deeply studied in the paper. Combined with the texture recognition, edge detection algorithm for images is improved on DSP 6000 image process program.LOG operator is important in detecting edges in images with different scales. In the mean while, it has defects: value of σ in the operator can not be achieved and there are conflicts between the ability of eliminating noise and edges location, which means the Log operator is so sensitive to the variation of gray that it can detect edges precisely and cohere with real edges exactly with the lower value of σ. But there are lots of false edges because of residuary noise by Gauss Filter. With the higher value of σ, the operator can detect the edge of original images and noise is dramatically reduced while edges interfere with each other resulting in serious displacement.In order to avoid the defects of LOG operator, based on texture recognition, a method to get the optimum value of σ is proposed via co-occurrence matrix and LMS, a self-adaptable method. Detection of edges can be achieved automatically through Digital Image Processing system based on TMS320C6000 so that real-time of edge detection is reached. The results of the experiments show that Edges detected by the operator not only reduce most noise and little variation of gray, but also locate edges accurately to improve the function of LOG operator. The method can be realized easily and resolve problems concerning edges of images.