A pantograph image extraction method and device based on a line array camera

Abstract

The invention relates to the field of pantograph operation state monitoring. Aiming at the problems existing in the prior art, the present invention provides a pantograph image extraction method and device based on a line array camera. According to the extracted pantograph image, it is monitored whether the working state of the pantograph is abnormal, so as to guide the maintenance. According to the established pantograph classification model, the present invention performs feature detection and initial pantograph positioning on the image to be inspected. If there is a pantograph in the image to be inspected, the pantograph is repositioned to judge According to the position of the pantograph slide plate and the contact line, whether there is a real pantograph in the image to be inspected is judged according to the relative position of the pantograph slide plate and the contact line; otherwise, the pantograph initial positioning step is performed on the next image to be inspected.

Description

technical field [0001] The invention relates to the field of pantograph operation state monitoring, and relates to a method and device for extracting pantograph images based on a line array camera. Background technique [0002] The pantograph sliding plate monitoring device mainly detects the running state of the pantograph. Install the camera at the position where the train passes, and take pictures of the pantograph at fixed points, so as to monitor the abnormal state of the carbon slide of the pantograph, so as to guide the maintenance. [0003] Previously, pantograph images were mainly taken by area array cameras, which mainly had the following disadvantages: 1. The area array cameras were triggered by fixed points, and only took one or two photos, while the hardware trigger device was prone to false triggering, resulting in the There is no pantograph or only part of the pantograph in the image; 2. Because the pantograph monitoring is real-time online monitoring, the hi...

AI Technical Summary

Problems solved by technology

[0003] Previously, pantograph images were mainly taken by area array cameras, which mainly had the following disadvantages: 1. The area array cameras were triggered by fixed points, and only took one or two photos, while the hardware trigger device was prone to false triggering, resulting in the There is no pantograph or only part of the pantograph in the image; 2. Because the pantograph monitoring is real-time online monitoring, the high-speed railway train runs too fast, and the frame rate of the area array camera cannot keep up with the train running speed, resulting in the Pantograph pictures are prone to smear

[0005] The pantograph image captured by the line array camera has the following characteristics, such as: 1. The illumination conditions are different. The pantograph monitoring device is an all-weather detection device. It is affected by sunlight during the day, and it is prone to serious reflections of the pantograph, resulting in The pantograph photographed is too bright; some normal-speed trains are coal freight trains, and the roof is dark, so the pantograph is also dark; 2. The line frequency of the line array camera does not match the speed of the train. Usually, before the monitoring device works, it will Set the parameters of the camera line frequency to match the experienced speed. However, sudden acceleration and deceleration will inevitably occur during the train operation. The mismatch between the line frequency and the speed will cause the captured pantograph image to be distorted and the size to change. bigger or smaller

3. The objective situation of each monitoring device installation site is different. Due to the different installation personnel and different installation positions of different monitoring devices, it is easy to make the positions of the pantographs in the image different.

In view of the characteristics of the images captured by the above line array cameras, the traditional image processing method needs to set too many algorithm parameters, which is sensitive to the environment. Different installation points may need to set different parameters. The workload is huge, and the parameter setting is experienced. In addition to image processing professionals, the rest of the staff cannot set the parameters well

Images