Design of Multi Direction Real-Time Online Edge Detection System Based on FPGA

According to the detection results of traditional single image or two direction edge is not ideal, the direction is not strong, the problem of low efficiency of detection, and proposes an improved edge detection algorithm based on FPGA multi direction, give full play to the FPGA internal resources the advantages of parallel processing of digital signal. The algorithm in the original 0 degrees, 90 degrees two direction on the basis of increase 22.5 degrees, 45 degrees, 67.5 degrees, 112.5 degrees, 135 degrees and 157.5 degrees in six directions of the Sobel template, can better detect multi direction edge. Considering the center pixel and its neighbor pixels in different distance from the impact algorithm are weighted according to the pixels of each pixel to pixel Euclidean distance to the center in the neighborhood, the distance of the center pixel closer, the greater weight, and by programmable adder and programmable parallel adder to improve the detection efficiency. The experimental results show that the algorithm can real-time online, efficient detection of complete image’s edge information and contour lines are clear, directional, and has a high significance.