Design and Implementation of the Algorithm for Blind Identification System Based on FPGA

Abstract

This paper discusses a blind sidewalk detection algorithm based on FPGA parallel processing technology and digital image processing technology. Depending on the colors of blind sidewalk and pavement, the algorithm includes two major modules which are processed in parallel. In one module, blind sidewalk is detected by the value of the hue when the colors of blind sidewalk and pavement are significantly different. In another module, blind sidewalk is detected by edge detection algorithms optimized by mathematical morphology, when the colors between blind and sidewalks are similar. The proposed design optimization based on morphological edge detection includes three parts: sobel operator edge detection, morphological dilation image processing and Hough transform for line detection. The simulation results show that the system for blind sidewalk detection has well real-time and accurate recognition accuracy.