Research And Realization On The Method Of Counting Drillpipe’ Number Based On Machine Vision

Gas explosion,outburst of gas accident is one of the largest category in the process of the coal mine production and a hidden safety trouble [1].Therefore, gas extraction technology is the most effective method of comprehensive control of gas.At present, for the gas extraction of drill hole depth measuring the basic use of MWD method;For the supervision and measurement after the drill, still with the method of artificial sampling detection, there is great potential safety hazard.Machine vision technology is a multidisciplinary integration of technology, the main purpose is to obtain the target image signal, and through image processing module to a series of image processing, able to detect the characteristic information of the target and then coordinate with decision-making module and control module to make certain.Study a drill pipe counting system based on machine vision technology, firstly the system do some processing with the video image intelligently,so as to realize the automatic counting of drill pipe, the subject for strengthening the safety management in coal mine, it is a great significance.In the first place,the system set a identification device in the drill,and get the coal mine video for drilling by cameras, and transmission image through Ethernet ring network to industrial computer image processing system, thus image preprocessing, feature extraction,object detection, target tracking, finally realizes the automatic counting of drill pipe.Aimed at bit target detection problem, this paper proposes a method based on improved Vibe target detection, interframe difference method was applied to initialize the background model, the phase detection and updating background using the attenuation of the update methods to remove ghost area, and extract the Vibe to detect the edges of the target with the original goal or operation,finally got ideal moving targets.Aimed at tracking problem for drill pipe, traditional Camshift algorithm of target tracking is vulnerable to shelter and workers head light interference, and counting in theinitial need to manually choose identity goal, therefore, this paper proposes a bit marker target tracking algorithm based on improved Camshift algorithm: adopt the method of template matching positioning tracking target position in the first frame, instead of manual frame chosen way;the motion characteristics and the integration of color features, color interference to solve similar problems; kalman filter is used to solve the problem of part of the cover.The system hardware mainly includes the coal mine explosion-proof camera, optical transceiver, video server, industrial control, lighting and so on.Application Visua1 Studio2012 + OpenCV function library, developed friendly interface of the image processing software system, realized the automatic counting of drill pipe.According to the field test results show that the images do this design institute of coal mine underground pipe automatic counting system of counting validity can reach more than80%, basically meet the actual needs.