Electric locomotive pantograph and arc detection method based on deep network and hierarchical testing

Abstract

The invention discloses a method for detecting electric locomotive pantographs and electric arcs based on deep network and hierarchical testing, including: collecting video, extracting pictures of pantographs, electric arcs and suspected electric arcs, and classifying them as neural network training sample sets; The convolutional neural network for real-time pantograph detection and electric arc detection is implemented on the CPU; the convolutional neural network is trained with training samples to obtain pantograph detection models, arc detection models, and arc detection models; the improved YOLO algorithm is used to detect the The pantograph is used as the first-level detection, and the partial picture of the area where the pantograph is located is intercepted, and the adjacent frame prediction method is used to speed up; the improved YOLO algorithm is further used to detect the arc in the partial picture of the pantograph as the second-level detection, and the abnormality is saved picture. The invention can quickly and accurately detect the pantograph arc by using the method of deep learning, give early warning to potential danger, facilitate vehicle failure inspection and ensure operation safety.

Description

technical field [0001] The invention belongs to the technical field of computer digital image processing and pattern recognition, in particular to an electric locomotive pantograph and arc detection method based on deep network and hierarchical testing. Background technique [0002] With the rapid development of electrified railways represented by high-speed railways in China, higher requirements are put forward for the safety of traction power supply systems. As an important part of the traction power supply system, the pantograph-catenary relationship is the main power supply equipment of the electrified railway and plays an extremely important role in the operation of the electrified railway. Under normal conditions, air is a good insulator, but if a large enough voltage is applied across the gas, the air gap will be broken down, resulting in the conduction of current across the gas. This phenomenon is called gas discharge. When the pantograph slides on the contact wire ...

AI Technical Summary

Problems solved by technology

[0004] (1) A high amplitude overvoltage is generated instantly when the arc is extinguished, which affects the safe operation of the electric locomotive;

[0005] (2) Ablation of pantograph carbon slide plate and contact wire, shortening its service life

In severe cases, it will lead to disconnection of the contact wire, resulting in driving accidents;

[0006] (3) Generate electromagnetic radiation and interference

[0011] The Chinese invention patent application with the application number 201310217622.5 published on October 10, 2013, announced a "pantograph automatic detection method based on the active shape model"; the invention is based on the active shape model to complete the pantograph carbon slide The online automatic quantitative detection of thickness has a certain detection and analysis effect, but the defect is that the detection method is based on single image analysis, which does not make full use of the inter-frame correlation of the video collected by the surveillance camera installed in the pantograph system. The arc generated by catenary and pantograph during operation is not analyzed and identified

This method uses high-speed digital imaging equipment to focus on the electric shock position of the electric locomotive pantograph and catenary, and continuously records and shoots images in motion, but lacks the key arc analysis method, and high-speed digital imaging equipment, especially thermal infrared high-speed digital imaging equipment It is expensive and does not belong to the supporting equipment of the pantograph monitoring system of electric locomotives, and its practical application has great limitations

Images