Freight train transfinite computer vision inspection system

Abstract

The invention provides a freight train transfinite computer vision inspection system, and belongs to the field of precision optical measurement. The freight train transfinite computer vision inspection system comprises a line laser module, a camera module, a vision processing module and an alarm classification module. The data output end of the camera module is connected with the data input end of the vision processing module; the output end of the vision processing module is connected with the data input end of the alarm classification module; the video processing module comprises an image preprocessing module, a coordinate system transformation module and a high-speed measurement module. The hardware of the freight train transfinite computer vision inspection system consists of a camera and a line-structured optical laser which are arranged on a portal frame; and various transfinite sizes of the top and the two sides of a freight train can be measured automatically and quickly.

Description

technical field [0001] The invention belongs to the fields of precision optical measurement, computer software development, etc., and relates to the geometric dimension technology of online measurement of the outer contour of a large-scale moving object (train) using a computer vision method. Background technique [0002] When railway freight vehicles enter the station, if the goods carried are over-width, over-height, or over-limit, it will cause damage to station facilities and personnel. Therefore, online real-time detection of the appearance of freight trains is conducive to timely judging the safety status of freight trains. [0003] Due to the variety of shapes and types of over-gauge trucks, it is necessary to detect and classify them under a unified standard. There are following deficiencies in the prior art: [0004] 1. The measurement reliability is not high: the more common method is to use optical sensors including lasers. This sensor can only measure the obstru...

AI Technical Summary

Problems solved by technology

[0004] 1. The measurement reliability is not high: the more commonly used method is to use optical sensors including lasers. This sensor can only measure the obstruction of objects within the light passing range. It belongs to point detection and is easy to miss or false alarm.

[0005] 2. Insufficient coverage of the measurement range: both optical sensors and lidar detection will miss some areas

[0006] 3. The ability to discriminate the type of overrun of the measured object is not strong: because it is difficult for the sensor or laser radar to distinguish the specific overrun item, it is impossible to make an accurate conclusion on the overall identification of the item

Images