A distributed canny edge detector: Comparative approach

The aim of the system is to implement the new canny algorithm for parallel preprocessing block with no loss in edge detection results compared with the original frame-level Canny algorithm. The methodology employed here is distributed canny edge detection algorithm on the basis of threshold segmentation as the segmentation has good advantage in noise restraining with wide applicability and computes the thresholds based on the block statistics. The main objective of the proposed segmentation based system includes significant reduction in latency, better edge preservation. The simulation of proposed system with MATLAB version R2012a is carried out. The experimental result indicates the effect of mask size on the latency and effect of PSNR on quality of edges which quantify the overall system performance.