Railway freight open wagon carriage number and loading state identification method based on airborne vision

Abstract

The invention discloses a railway freight open wagon carriage number and loading state identification method based on airborne vision, and belongs to the field of machine vision. The method has the steps of using an unmanned aerial vehicle carrying a high-definition camera to collect image data of a freight open wagon carriage located at a loading station in the flight process; preprocessing the acquired image data, positioning a carriage character region according to priori knowledge such as a character width / height ratio, segmenting the character region by using a projection method, and constructing a CNN convolutional neural network model to identify segmented digital characters; locating an area where a carriage boundary is positioned through Hough linear transformation, designing gridfeatures, and building a BP neural network to recognize the carriage loading state; by using the method, the loading state of the carriage is judged while the carriage number of the freight open wagon is recognized, the recognition precision is high, the workload of workers is greatly reduced, and the method is suitable for recognizing the carriage number and the loading state of the freight openwagon along a railway.

Description

technical field [0001] The invention relates to an airborne vision-based method for identifying the numbering and loading state of railway freight gondola cars, and belongs to the field of machine vision. Background technique [0002] As the main fuel for thermal power plants, coal is still mainly transported by train. In order to track and manage the coal carriages, it is necessary to identify and register the number and loading status of the coal carriages. Whether coal carriages are loaded with coal directly affects coal transportation management. The traditional identification of the loading state of the coal carriage mainly relies on manual work. The workers go to the corresponding position of the carriage to check, and then record the carriage number and loading state. Due to the large number of coal carriages at a time, manual records are prone to errors and require a lot of labor. [0003] The current compartment number identification usually adopts radio frequency...

AI Technical Summary

Problems solved by technology

However, the above prior art does not automatically identify the loading state of the compartment, but only recognizes whether the compartment is overloaded

To sum up, the existing research results are mainly used to identify the number of freight open car compartments, and it is impossible to determine whether the compartments are loaded at the same time.

Images