A Judgment System of Car Crowding Degree Based on Image Recognition

Abstract

The embodiment of the present invention discloses a system for judging the congestion degree of carriages based on image recognition, including: a passenger flow collection unit, which collects passenger flow data at each stop within the collection period; a carriage target image collection unit, which collects passenger flow data within the collection period Collect video image data in the compartment; target tracking unit, which can identify the tracking target and obtain the tracking image data corresponding to each identified tracking target; target modeling unit, which can create a corresponding tracking target for each identified tracking target Three-dimensional tracking model; a single congestion degree judging unit, which can obtain the overlapping degree data of each marked tracking target within a given tracking frame range and other marked tracking targets; the compartment congestion degree judgment unit, which can obtain compartment congestion Degree data and sent to the client. The present invention can be applied to dispatching business scenarios such as real-time congestion degree analysis in vehicles, real-time dispatching and monitoring of lines, and analysis of operating characteristics of historical operation of lines.

Description

technical field [0001] The invention relates to the technical field of public transportation line scheduling, in particular to a system for judging the degree of congestion of carriages based on image recognition. Background technique [0002] Public transportation is the main way for citizens to travel, and publishing the congestion of public transportation is an important way to improve the quality of public transportation services and increase the attractiveness of public transportation travel. With the continuous deepening of the public transportation network, the public transportation methods and routes that passengers can choose are gradually increasing. The difference in travel comfort brought about by the degree of congestion in the car will greatly affect the choice of public transportation routes for passengers. [0003] The existing traditional identification methods have the following problems: 1. Manual investigation method, which has a large workload and a long...

AI Technical Summary

Problems solved by technology

[0003] The existing traditional identification methods have the following problems: 1. Manual investigation method, which has a large workload and a long time, and requires manual configuration of the census; 2. The identification method of swiping card data, the data accuracy is closely related to the passenger's riding habits, Training is needed to guide passengers’ riding habits, and it is necessary to ensure the real-time transmission of card swiping data; the accuracy of passenger flow equipment identification is related to the realization of passenger flow equipment technology, not only additional equipment is required, but also whether the carriage is crowded is based on intuitive image recognition technology, other The difference between accuracy and actual comfort is large

Images