Pointer instrument detection and reading identification method based on mobile robot

Abstract

The invention relates to a pointer instrument detection and reading identification method based on a mobile robot. The method comprises the following steps: obtaining a deep neural network detection model M for a pointer instrument; the mobile robot moves to a designated place to obtain an original environment image containing the instrument equipment at present in a mode that the mobile robot carries a camera; S serves as system input and is transmitted to a deep neural network model M, whether an instrument exists in S or not is detected, the position of the instrument is framed out, an image in a frame is intercepted and subjected to height setting processing, the length-width ratio is not changed, and the processing result is represented by J; contrast enhancement is performed on the image, and a processing result is represented by E; local adaptive threshold segmentation is performed on the E to obtain a reverse binary image B; pointer extraction processing is performed based on aprobability circle; a central straight line L of the pointer part is extracted as pointing information of the instrument pointer; and a coordinate system is established based on a probability circlecenter projection algorithm and reading by an angle method.

Description

technical field [0001] The invention relates to a mobile robot-based method for automatically detecting and identifying readings of pointer meters. The algorithm realizes accurate readings of pointer meters by combining deep learning with traditional image processing. Background technique [0002] Due to its simple structure, convenient maintenance, anti-electromagnetic explosion-proof and other advantages, pointer meters are widely used in oil and gas pipelines, power metering and other industries. At present, meter reading and inspection work are mostly completed by manpower, with a low degree of automation. With the increase in the number of instruments and the requirements of the factory for the development of information automation, manual inspection and meter reading can no longer meet the needs of the factory. Automatically reading the readings of the pointer dial has become a problem that needs to be solved urgently. However, the method of monitoring by placing came...

AI Technical Summary

Problems solved by technology

However, compared with the YOLO V3 model used in the present invention (the basic network is Darknet-53), the SSD model used in this article (the basic network is VGG16) has a poorer detection effect on small targets, a higher missed detection rate, and a lower overall accuracy. Low, does not meet the requirements for high detection accuracy in industrial environments

In addition, the ellipse detection algorithm and MSRCR (Multi-Scale Retinex with Color Restoration) enhancement algorithm used by it have poor actual effect in the case of uneven illumination of the image.

[0007] To sum up, the current pointer meter reading algorithm has high requirements on image quality, and cannot completely solve the interference of factors such as low definition, complex lighting conditions, and angle inclinations, and has poor robustness.

Images