An aerial photograph image tower identification card fault diagnosis method based on depth learning

Abstract

The invention relates to an aerial photograph image tower identification card fault diagnosis method based on depth learning. The method comprises the following steps of: establishing a pole and toweridentification card detection image library and a tag library thereof; estalbihsing the image library and tag library of tower leg detection; establishing a tower identification card status classification image library and its label library; establishing the deep learning object detection mode Faster R-CNN, including the basic network NasNet, a zone Proposal Network and a Fast R-CNN detection network; establishing a depth-learning image classification model ResNet; traninig the established detection model or classification model on the prepared galleries, and performing data enhancement on the input data of each iteration in the training process, including random rotation, random fill-in cropping and random grayscale. According to the relative position of the detected tag and the tower leg, the fault of tower tag falling off is diagnosed. If there is no tag falling off, the state of the tag is diagnosed by using the classification model.

Description

technical field [0001] The invention belongs to the fields of high-voltage transmission line inspection technology, image recognition technology, and machine learning technology, and in particular relates to a method for diagnosing faults of tower identification plates based on aerial photography images based on deep learning. Background technique [0002] The tower identification plate is an important component in the transmission line operation and maintenance management system, which is installed on each base tower of the overhead transmission line. The pole number plate in the identification plate shows the line name, line number and pole tower number. The inspection and maintenance personnel can know the basic information of the pole tower through the pole number plate, which is convenient for subsequent maintenance work. The warning signs in the signboards display prohibited behaviors and safety reminders, warn staff and non-staff of relevant safety information, and co...

AI Technical Summary

Problems solved by technology

[0003] In recent years, unmanned aerial vehicle inspection has gradually become one of the main means of inspection and maintenance of transmission lines. The necessity of automatically detecting tower signs in aerial images and performing fault diagnosis has become increasingly prominent: sign records the information of current poles and towers , is an effective auxiliary positioning method. If the identification plate can be detected during the inspection process of the UAV, the inspection image of the transmission line and the image of the number plate can be recorded synchronously, so that the inspection and maintenance center can store The inspection image data of the information is convenient for subsequent fault location; the identification plate is located outdoors, and is affected by factors such as wind and rain all the year round, some long-term years, and individual line tower changes, which are prone to serial number blurring, corrosion, damage, bending, etc. Folding and other phenomena, leaving potential safety hazards for line operation, maintenance and management

Images