A multi-stage model joint detection method for super-complex category defects of power grid voltage equalizing ring

Abstract

The present application relates to a combined detection method, device, computer equipment, storage medium and computer program product of a multi-stage model for super-complex category defects of a power grid voltage balancing ring. The method includes: performing joint reasoning on the first model, the second model and the third model to obtain the first training result, the second training result and the third training result. A first optimized first model is determined based on the first training results and the image labels corresponding to the corresponding first training results. Based on the second training results and the image labels corresponding to the corresponding second training results, a second optimized first model and a second optimized second model are determined. Based on the third training results, and the image labels corresponding to the corresponding third training results, the optimized first model, the optimized second model, and the optimized third model are determined. Through multiple iterations of training, it is determined that the first target model, the second target model and the third target model obtained after the training is completed, which improves the detection accuracy of the voltage equalizing ring defect.

Description

technical field [0001] The present application relates to the technical field of defect detection, and in particular, to a method, device, computer equipment, storage medium and computer program product for joint detection of super-complex category defects of power grids with multi-stage models. Background technique [0002] With the development of defect detection technology, it is often necessary to perform defect detection on the voltage equalizing ring in the transmission line of the power grid to maintain the stable operation of the power grid. Among them, in order to detect the defects of the pressure equalizing ring, the YOLOv4 (You Only Look Once 4th version) model corresponding to the defect category is often constructed. In this way, the images to be detected are respectively input into the YOLOv4 models corresponding to each defect category to determine the category of the voltage grading ring defect. [0003] In the traditional technology, in the training proces...

AI Technical Summary

Problems solved by technology

[0004] Based on this, it is necessary to provide a multi-stage model joint detection method, device, computer equipment, computer-readable storage medium and computer program product for the above-mentioned technical problems.

Images