作者: Xiaodong Xian , Qiqi Jiang , Xudong Liu
DOI: 10.1109/ICAMECHS.2017.8316498
关键词:
摘要: In a coal mine, target and background have similar gray under low illumination, sudden illumination changes can cause false targets. Due to the fact that detection method based on Gaussian mixture model only consider color information, problems of undetected objects detections appear. An improved is proposed, which integrates edge information linearly. First, using Mixture Model fitting background, extracting image as supplement classified foreground edge, then subtraction are established normalized, finally integrate two by linear detect miners. The simulation results indicate proposed algorithm effectively miner with removes moving caused change, thus improve accuracy detecting target.