Method and device for detecting and identifying foreign matters of rail wagon

Abstract

The invention discloses a method for detecting and identifying foreign matters of a rail wagon, which comprises the following steps of: performing sliding window blocks on an input image, and obtaining a corresponding number of local detection areas according to the sizes of the sliding windows; carrying out foreign matter suspected candidate region positioning on all the local detection regions one by one, and screening and obtaining foreign matter candidate regions suspected to have foreign matters; carrying out foreign matter secondary classification and identification on the foreign mattercandidate area, and judging whether the foreign matter is a foreign matter and a specific category to which the foreign matter belongs; and outputting foreign matter information according to the category of the foreign matter category and a preset severity level. The method has the beneficial effects that the high-real-time and high-accuracy automatic detection and identification of the foreign matters of the rail wagon can be realized, the method has very high detection precision and relatively low false identification rate, meanwhile, the relatively high detection speed also meets the actual engineering requirements, the missing detection and false alarm conditions are greatly reduced, the burden of workers is reduced, and the potential safety hazard is reduced.

Description

technical field [0001] The invention relates to the technical field of rail transit foreign object detection, in particular to a deep learning-based method and device for detecting and identifying foreign objects on railway wagons. Background technique [0002] During the high-speed operation of the train, the foreign matter suspended on the car body will exceed a certain range of the car body under the influence of the airflow, which will seriously affect the safety of the peripheral equipment of the railway. Accidents that damage peripheral equipment due to foreign matter occur from time to time. If the foreign matter suspended by the car body falls off during operation, falling on the rails will greatly increase the possibility of derailment of this train or the next train. Detecting foreign objects in railway wagons is an important measure to ensure train safety and cargo safety. [0003] At this stage, the traditional algorithm uses grayscale changes to detect foreign ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the defects in the prior art, provide a method for foreign object detection and recognition of railway wagons based on deep learning, and solve the problem that the existing traditional foreign object detection algorithm has a poor detection effect on foreign objects with a gray scale similar to that of the car body, and the False alarms caused by body reflections and painted parts of icon fonts

Images