Device and method for non-contact detection of rail vehicle compartment state

Abstract

The invention discloses a non-contact detection device and method for the state of a rail vehicle carriage. The device includes three groups of data acquisition units composed of laser scanners and high-definition cameras, and a control unit for image recognition processing on the data acquired by the data acquisition units. The data acquisition unit collects the relevant data of the rail vehicle and sends it to the control unit. The control unit uses the image recognition algorithm based on the deep belief network model, and uses the surface contour and image data of the carriage as samples for training. Fast and accurate detection. From this, it can judge the arrival of the vehicle, measure the speed of the vehicle, identify the type of car, restore the three-dimensional image of the car, judge the overrun and load situation of the car, identify the type of cargo and cover information in the car, detect the opening and closing status of the car door and the degree of damage to the door lock and other car states. , which greatly improves the detection efficiency and frees the operator from cumbersome operations.

Description

technical field [0001] The invention relates to a non-contact detection device and a detection method, in particular to a non-contact detection device and method for detecting the state of a rail vehicle compartment. Background technique [0002] At present, the carriages required for rail transportation in my country mainly include box cars, gondola cars, tank cars, and air-conditioned passenger vehicles. Whether the lock is in good condition, recording the type of goods and covering information, estimating the loading capacity of the compartment, judging whether the height of the goods exceeds the limit and the position of the overrun, etc. [0003] At present, the detection and recording of the status of the truck compartments are mainly done manually, but due to the large number of compartments and the long distances between them, it is difficult to perform effective manual detection in a relatively short period of time. Several electronic pre-inspection systems are also...

AI Technical Summary

Problems solved by technology

However, it is still necessary to manually participate in the identification and pre-inspection of the vehicle when the vehicle passes. It cannot automatically identify the type of compartment and the type of cargo, and cannot check the state of the cover of the open car, so there is a lack of real-time performance.

And it is impossible to accurately calculate the speed of the vehicle, automatically identify the front and rear of the vehicle, automatically identify the joints of the carriages, calculate the number of carriages, and restore the three-dimensional digital images of the carriages. There are still certain difficulties in management

Images