Automatic detection and defect identification method for power grid wire clamp and device

Abstract

The invention discloses an automatic detection and defect identification method for a power grid wire clamp. The method comprises the following steps: acquiring a wire clamp picture through an image acquisition module; training the wire clamp picture to obtain a wire clamp type; classifying, deleting and selecting the wire clamp pictures; and identifying and outputting the fault wire clamp. The invention also discloses an automatic detection and defect identification device for the power grid wire clamp. According to the method and the device, image acquisition is carried out on the wire clampon the power transmission line through the image acquisition module; the Faster R-CNN technology is adopted to train the wire clamp pictures, the augmented training technology is adopted, the defectsthat the time is long and the model updating period is slow due to the fact that new wire clamp features need to be retrained are overcome, fault detection is conducted on the wire clamp, the maintenance cost is greatly saved, and a line patrol system is more efficient and intelligent.

Description

technical field [0001] The invention relates to the technical field of safety of overhead power transmission lines, in particular to a method and device for automatic detection and defect identification of power grid clamps. Background technique [0002] In the overhead transmission line, various transmission lines are intertwined, and the transmission lines may be entangled, pulled, and vibrated with each other, which can easily lead to strand breakage and seriously threaten the safe and reliable operation of the overhead transmission line. As an important device in overhead transmission lines, wire clips can wrap a large number of transmission lines, protect the transmission lines and assist vibration reduction, and greatly extend the service life of the transmission lines. However, overhead transmission lines are used for long-distance power transmission, and they must cross different natural environments during the period. After a long period of exposure to wind and sun...

AI Technical Summary

Problems solved by technology

However, in practical applications, a large number of clips with different characteristics, such as appearance and texture, are difficult to collect at one time. Then there are clips with new features. The current solution is to combine the collected data with the previous ones. The collected data is unified and retrained. This solution has a flaw. With the addition of newly collected data, the overall data becomes more and more, the model training becomes slower and slower, and the model update cycle becomes longer and longer. , seriously affecting work efficiency

Images