A multi-level passenger flow monitoring method for rail transit networks based on multi-source data fusion
In the field of intelligent transportation,passenger flow monitoring is a research hotspot,and the data obtained has im-portant guiding significance for timely scheduling and reasonable allocation of transportation resources.The existing methods for moni-toring rail transit passenger flow have the problem of strong predictability and inaccurate monitoring results.A multi-level rail transit network passenger flow monitoring method based on multi-source data fusion is designed to achieve monitoring of rail transit network passenger flow under large passenger flow.A large number of high-definition network cameras and infrared cameras are deployed in multi-level rail transit networks to collect passenger flow monitoring images.Design an image denoising algorithm based on sparse representation of image signals for the two collected external images,and apply denoising processing to them.Design a pedestrian de-tection algorithm based on the DPM model,and implement pedestrian motion target detection in visible and infrared images.For the detection results of two types of images,design a multi-source data fusion algorithm based on dense cross network,implement multi-source data fusion and counting,and achieve the purpose of multi-level rail transit network passenger flow monitoring.The test re-sults show that the passenger flow monitoring errors of the high-speed rail network,intercity trajectory network,suburban railway net-work,and urban rail network of this method are 44 people,40 people,19 people,and 4 people,respectively,with low passenger flow monitoring errors.
万方数据
维普
