Miner face detection is based on improved AdaBoost algorithm

摘要

This article connects with Coal mine video monitoring image be impacted for special environment, which be vulnerable to mineral dust in coal mines, light, as well as miner's safety helmet for the realization of face detection in real-time and accuracy, I will study on face identification and analysis on the characters of behavior in the follow-up work for getting a good foundation, which will be in intelligent Coal mine video monitoring. This article simulates rectangle Haar-like character and Extended Haar-like character of the AdaBoost algorithm about face detection in real-time and accuracy, is based on OpenCV, also describes briefly the rectangular Haar-like characteristic model and about computational algorithm and faster algorithm of the characteristic value, analysis detailedly extended Haar-like character model and the characteristic value of computational algorithm-integral image. Experimental resulted show that extended Haar-like characteristic model can be implemented more quickly and more accurately in the miners' face detection, as well as real-time.