Research on Instrument Recognition Technology of Inspection Robot in Coal Mine
In order to realize the unmanned and intelligent inspection robot in coal mine and ensure that the inspection robot can read the display data of various instruments,a multi-type instrument visual recognition system based on deep learning is designed in this paper.The system uses the YOLOv3 network model to distinguish and locate the instrument type,and uses the improved radial gray method combined with Hough transform and Canny edge detection algorithm to read the data of the pointer instrument.The EasyOCR is used to get the readings of the digital meter.The test results show that the system has good accuracy in instrument identification,and the reading error of the three types of instruments is less than 3.8%.This provides a practical and effective solution for the vision system of inspection robot in coal mine.
NETL
NSTL
万方数据
