The Coal Mine Video Monitor The Study Of Abnormal Behavior Recognition Algorithm

In recent years, intelligent video surveillance and identification is anew hot spot in the forefront of the topic and research direction of thefield of computer vision. With the development of network technology,monitoring technology is to unmanned, intelligent, digital directionprogress continuously, and thus the field of computer experts timely putforward a new generation of video monitoring concept. For coal mineunderground, it because of frequent accidents, easy to cause a largenumber of casualties, the intelligent monitoring system can timely findout problems before the accident, timely warning can reduce casualties,the analysis results with using of video image data directly related to theeffect of early-warning is good or not, and whether can accurate warning.In this paper, through a new generation of intelligent monitoringtechnology used in coal mine underground abnormal conditionrecognition, achieve the purpose of early warning. For video image of theactual situation, the different processing method, due to insufficient lightwell, take video image enhancement processing, due to the specialcondition of the underground, and take in the maximum variance betweenmethods for the selection of threshold value, then the hierarchical imageprocesses, to achieve the desired effect.Combined with early treatment, and obtained the coal mineunderground video data, know gas bursting and rib spalling and the characteristics of abnormal situation, because video material can’t includeall situation, simulated the abnormal situation, and combining with thethree frame differential method for simulation of the video imagerecognition, whether gas bursting or rib spalling and anomaly appear tobe able to achieve high recognition rate and is expected to reach the goal.Simulation is completed in the analysis of the image processingtechnology and architecture, for actual monitoring conditions andrequirements, the design of the underground detection and with therecognition algorithm using MATALB development platform for imageextraction and pre-processing, segmentation and recognition of themoving target. The experimental results show that the algorithm is able tohandle the general static background conditions underground disasteridentification.