It is the algorithm flow of the whole bucket tooth fault detection system of excavator. This algorithm needs to detect the target of the template image first. After obtaining the position of the bucket teeth of the normal excavator, the bucket teeth of the normal excavator are separated from the image for image segmentation. After segmentation, the contour feature can be extracted. Through the above steps, the number of bucket teeth, the position information of bucket teeth, the position relationship between bucket teeth, and the contour feature information can be obtained.
A similar process is carried out for the image to be detected. First, the number of bucket teeth of the excavator is counted a priori through target detection. If it is consistent with the number counted in the template image, the next image detection will continue; If the numbers do not match, the subsequent image segmentation and contour feature extraction operations are carried out to obtain the number of bucket teeth to be detected, the position information of bucket teeth, the position relationship between bucket teeth of excavator and the contour feature information. Through the calculation of feature similarity, the feature similarity between the template bucket teeth and the bucket teeth to be detected is calculated, and combined with the position relationship between the bucket teeth, the bucket tooth sequence position of the bucket teeth to be detected can be determined. At the same time, the fault bucket teeth can also be located. Then the image segmentation of the fault excavator bucket teeth is carried out to obtain the segmentation diagram, and finally the wear degree is calculated.
In each detection process, the target detection results, image segmentation results, bucket tooth information and fault information of the excavator are output to the graphical user interface to visualize the results.