High-speed railway rockfall real-time detection method based on aggregation channel features and texture features

Abstract

The invention discloses a high-speed railway rockfall real-time detection method based on aggregation channel features and texture features, and belongs to the field of image target detection. The method includes collecting a sample image without falling rocks, performing image preprocessing, and establishing an initial background model; determining a detection area by using the established background model, and extracting aggregation channel features in the image; based on a background subtraction method, obtaining a binary image as a preliminary detection result; introducing an HSV color space, removing a virtual scene in the binarized image based on texture features, taking the binarized image with the virtual scene removed as a rockfall detection result at the current moment, and marking the detection result in the current image; and updating the background model. According to the invention, the rail area can be rapidly divided from the RGB image, prior information is not needed, a large amount of feature matching calculation is not needed, meanwhile, aggregation channel features and texture features are utilized, interference caused by outdoor factors such as illumination is effectively solved, and meanwhile falling rocks on the high-speed railway can be effectively detected.

Description

technical field [0001] The invention belongs to the field of image target detection, and more specifically relates to a real-time detection method for high-speed railway rockfall based on aggregated channel features and texture features. Background technique [0002] The geological conditions along the railway are complex and changeable, and the track structure also faces the risk of foreign object invasion caused by various natural disasters such as landslides, collapses, mudslides, and earthquakes. As a result, frequent railway accidents occur every year, posing a great safety hazard to railway transportation safety. Therefore, the problem of track structure safety monitoring has become a key problem to be solved urgently. [0003] At present, although my country has many years of engineering practice experience in the construction and operation of railway track structure anti-foreign object intrusion, and has accumulated certain theoretical research results in the field o...

AI Technical Summary

Problems solved by technology

[0009] In view of the existing technology in the detection of falling rock targets, neither the traditional feature matching method can be used for detection, nor can the deep learning method be used for feature learning, and effective detection cannot be carried out; in the complex outdoor background, the target detection is vulnerable to rain and light. The influence of outdoor factors such as changes causes problems such as false detection. The present invention proposes a high-speed railway rockfall detection method based on aggregated channel features and texture features to solve these problems.

Images