A detection and recognition system of pointer meters in substations based on computer vision

Automatic reading of pointer meters is of great significance to meter verification and industrial inspection robots. Many automatic reading algorithms of pointer meters based on machine vision have been proposed. However, the pointer shadow problem has not been effectively solved, which hinders the automatic reading of inspection robots in complex lighting environments. Also, pointer line recognition is easily influenced by the images around the pointer. With the aim of addressing these problems, this paper proposes a robust and high-precision pointer meter automatic reading algorithm. In meter center recognition using a round-degree method, this paper uses the eight chain code shape to represent the pixel image shape to prevent abnormal calculation of the round-degree. A novel binarization threshold determination method based on the symmetry degree is proposed to segment the pointer region, which effectively solves the pointer shadow problem. In addition, for pointer recognition, this paper proposes an improved random sample consensus (RANSAC) algorithm to overcome background image interference. The experimental results prove that the algorithm effectively removes pointer shadows and that the reading precision is high.

https://iopscience.iop.org/article/10.1088/1361-6501/aaed0a/pdf

A robust and high-precision automatic reading algorithm of pointer meters based on machine vision

Take advantage of the good matching performance of scale-invariant feature transform (SIFT) features between multiple images about a textured object, our automatic meter reading method calculates a homography matrix H between the template image and the test image, uses random sample consensus (RANSAC) to filter out false matches, ensuring an accurate H, and finally maps the measuring curve from the template image to the test image. On the path of this measuring curve, the relationship between the pointer and the scale is calculated based on the gray scale change in the image. The pointer is detected and the reading can then be obtained using a particular formula. Our method provides rotation invariance, scale invariance and higher reading accuracy. Experimental results show the effectiveness of our method in the substation environment.

An Automatic Meter Reading Method Based on One-dimensional Measuring Curve Mapping

Aiming at the inspection requirements of substation equipment, a pointer instrument recognition and reading algorithm based on the pointer rotation center position was designed. Firstly, the original ORB feature matching algorithm is improved to realize the recognition and searching of the instrument. A similar feature triangle method is proposed to determine the rotation center of the pointer by matching feature points. According to the central coordinate, the gesture of the PTZ is controlled through the serial port instruction to make the target instrument positioned to the center of the image; the pointer instrument is read by the method of Hough transform. The results show that the accuracy and real-time performance of similar feature triangle method for identifying the center of rotation of the pointer is increased by 88.88% and 88.03%, respectively, compared with the gravity method. And the mean deviation of the meter reading is 0.011 MPa, which is better than that without positioning by the rotation center of the pointer. The error of the instrument reading was reduced by 67.65%. The experimental results show that the pointer rotation center can be extracted rapidly and accurately by the algorithm, and the error of instrument reading is reduced more, and the inspection requirements of the substation equipment can be met better.

Research on Algorithm of Pointer Instrument Recognition and Reading Based on the Location of the Rotation Center of the Pointer

Aiming at the problem that the reading speed of pointer instruments is slow and the precision is low, this paper proposes a new automatic reading algorithm. The paper improves the algorithm in instrument center recognition, pointer recognition and 0 scale line recognition. First, round-degree is used to recognize the instrument center. Then the algorithm removes the interferential images outside the pointer area through the polar angle frequency distribution histogram and uses the improved least square method to fit the pointer line. It overcomes the problem that the fitted line is off the instrument center. Next, the proposed method removes the interferential images around the 0 scale line by the polar radius frequency distribution histogram and uses the improved least square method to recognize the 0 scale line. It solves the problem that Hough line transform recognizes the short line under the influence of interferential images difficultly. Finally, the method calculates the reading according to the relationship between instrument scales and pointer angles. Experimental results prove that the algorithm is fast and accurate.

An automatic reading method of pointer instruments

An improved various meters recognition algorithm was proposed which was adapted to suit the equipment inspection robot works outdoor in the smart substation.First,the template images were made for each style of meter and the "min" and "max" values were illustrated for each template image.SIFT method was utilized to match with the template image for detecting the meter device region from the real-time image captured by the robot.And then,a fast Hough transform method was proposed to detect the line in the preprocessed image as the pointer of the meter.Comparing the angle with the "min" or "max" value the result of the meter can be recognized.This algorithm has been applied in one 500 kV smart substation in China.Various instruments recognition rate of over 99.2%,with high precision and high robustness of the meter readings.Experimental results of various meters have shown good performance,which proves that the proposed method was both effective and accurate.

Li Jian

Papers

Meter Recognition Algorithm for Equipment Inspection Robot