Research On Detection And Recognition Method Of Circular Pointer Instrument With Uniform Scale

In the past few decades,with the development of computer technology,computer vision and digital image processing technology have been widely used in the detection and identification of pointer meters.Aiming at the round pointer type instrument image with uniform distribution of scales collected by the substation inspection robot,this paper proposes a method based on computer vision for automatic detection and identification of pointer instruments,The method mainly includes instrument target detection,instrument image registration and instrument representation three steps:Firstly,the improved YOLO algorithm is used to train the self-made pointer instrument dataset to obtain the convolutional neural network model.The trained convolutional neural network model is used to detect the bounding box position of the instrument target in the current field of view,and the ROI area of the instrument target is obtained,then the detection of the instrument target is completed.Then,the ORB algorithm is used to extract the feature points of the detected instrument target image and the pre-acquired high-quality instrument template image,and the feature points are matched by the FLANN algorithm,Eliminating erroneous matching pairs by RANSAC algorithm combined with improved feature matching strategy,Further,the parameter matrix of the perspective transformation parameter model is obtained,and the high-quality instrument image obtained by the perspective transformation of the instrument image to be registered is obtained,then complete the registration of the instrument image.Finally,using the improved EAST algorithm proposed in this paper,the text of the high-quality instrument image after registration is detected,all the texts in the instrument image are detected,and the detected text image is identified by the designed digital recognition model,then the positioning and identification of the scale number of the instrument are completed.On the basis of this,the meter scale interpretation method,the center of the dial and the straight line of the meter pointer are extracted by the instrument representation number interpretation method proposed in this paper,the meter reading can be completed by combining the angle method and the distance method.The proposed pointer meter reading method is verified by a large number of experiments.The experimental results show that the average accuracy of the instrument identification method proposed in this paper is higher than 97.5%.The automatic detection and recognition task of the circular pointer instrument with the scale mean line in complex background has good accuracy and stability,and can meet the practical application requirements of the intelligent substation.