Study Of Methane Alarming Device Automatic Verification Method Based On Machine Vision

The most important problem of mine safety is gas leaking, in order tomaking mine safe, the staff should go down to the bottom of the mine and test theconcentration of gas. Therefore, the mathane alarm instrument is an essentialinstrument of mine to detect methane gas, the alarm instrument must be acurate,neither false to alarm nor leakage to alarm. This requeires regular examination onthe methane alarm device. Currently, the method of verification on methaneinstrument is manual testing. For large mines, the workload is very large, thereare hundreds of this device shloud be manual tested. In order to meeting the needof industrial production, it is essential to design an automatic verification method.In the modern industrial production and testing field, Digital displayinstrument has been widely used by its unique advantage. But most digitalinstrument is not given the data communication interface to computer, whichbrings a lot of inconvenience to instrument automation calibration. In traditionaltechnology, if want to verification for the coal mine methane test instrument,need a special worker to fill into the standard gas for detected instrument, then,entry the display data of detected instrument into computer, in order to analyzingdata, processing, workload is very big. This artificial verification methodvulnerable to the influence of the artificial factor, it was difficult to ensure theefficiency and accuracy of calibration. For some Pointers instruments there iscertain research for automatic verification method both at home and abroad, butto digital instrument is not much.This paper is based on the realization of automatic verification of methanealarming device, and we resolve the core prolem of the test data automaticverification in the digital instrument display window which is shown up in the realization of the verification of the methane alarming device through combiningthe machine vision technology and image processing technology, which realizesthe automatic verification of the mechane alarming device. Through tocontinuous image acquisition for the digital instrument data display window, andthen do the following treatment: color image with gray, binary, charactersegment and character enhancement, template matching and so on a series ofdigital image processing technology, to realize the automatic verificationfunction of machine vision instrument, not only improve the efficiency andaccuracy of the recognition but also reduce a lot of manpower material resourcesconsumption. This study in image recognition link collection a group of highdefinition image as a template and gathered a large number of instrument imagesin order to match testing with template, the results showing that the recognitionrate is100%, meet the requirements of the application in engineering. At thesame time it shows the feasibility of machine vision instead of the human eyereading instrument character to realize the automatic verification function.