Research and development of icon recognition system based on machine vision

Abstract

Machine vision represented by two-dimensional code and barcode is widely used, but both are not friendly enough for human eye recognition and cannot be used in human-machine collaborative work. Therefore, this paper proposes a machine vision-based icon recognition system that can quickly identify human-friendly icons. The system first converts the real-time image captured by the camera into binary image. Secondly, divide the binary image into 9*9 cells, get the binary information of each cell center point, match with the template information, and finally judge and draw the recognition conclusion. Experiments show that the method is fast and accurate in identifying icons, and the effect is good.