Target detection in mine based on fusion of color and edge information

In a coal mine, target and background have similar gray under low illumination, and sudden illumination changes can cause false targets. Due to the fact that the detection method based on Gaussian mixture model only consider the color information, problems of undetected objects and false detections appear. An improved target detection method is proposed, which integrates the color and edge information linearly. First, using Gaussian Mixture Model fitting the color information of background, extracting edge information of the image as supplement of color information, and classified edge as foreground and background edge, then the color background subtraction model and edge background subtraction model are established and normalized, finally integrate the two information by a linear method to detect the miners. The simulation results indicate that the proposed algorithm can effectively detect the miner of similar gray with the background and removes the false moving target caused by sudden illumination change, and thus improve the accuracy of detecting target.