Automatic reading recognition of pointer barometer based on machine vision

To develop a calibration system for a pointer barometer based on machine vision, an automatic reading recognition scheme of the barometer based on color images is proposed. To locate the dial while extracting the pointer, an optimized three-point-based circle determination method is proposed to calibrate the dial center and segment the dial. At the same time, an improved angle method is proposed to realize the automatic reading recognition of pointer instruments. First, the pointer area is detected and extracted by the three frame difference method, and the pointer centroids are extracted by least circumscribed rectangle fitting, and the optimal three points are selected to determine the center of the circle and locate the dial. Then, the two-color image matrix elements corresponding to different pointers are compared point by point to obtain the maximum value, and the dial without the pointers is also obtained. Through cropping, color channel component extraction, and binarization processing on the dial without pointers, a noninterference dial scale image is refitted, and each scale is fitted with a minimum circumscribed rectangle to calibrate the centroid of the scale line. Finally, using the centroids of the calibrated scale lines, the centroid of the fitted pointer, and the circle center obtained by three-point-based circle determination, the improved angle method proposed is used to realize the automatic reading recognition of the pointer instrument. Experimental results show that this scheme can effectively improve the accuracy of machine vision-based automatic reading recognition for pointer instruments. (c) 2022 SPIE and IS&T