Unlock instant, AI-driven research and patent intelligence for your innovation.

A traffic vehicle information acquisition method based on mask R-CNN

A technology of vehicle information and acquisition method, applied in the field of traffic vehicle information acquisition based on MaskR-CNN, can solve the problems of inability to meet high-level intelligent traffic, lack of robustness of vehicle information, etc., and achieve low equipment cost and high degree of intelligence. Effect

Active Publication Date: 2021-09-24
SOUTHEAST UNIV +1
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the existing traffic vehicle information recognition based on video detection is often limited to single tasks such as identifying vehicle types, vehicle speeds, and traffic volumes, and the acquired vehicle information is not robust enough to fully utilize the advantages of video methods and cannot meet high-level requirements. The need for smart traffic

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A traffic vehicle information acquisition method based on mask R-CNN
  • A traffic vehicle information acquisition method based on mask R-CNN
  • A traffic vehicle information acquisition method based on mask R-CNN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0024] like Figure 1-Figure 5 A traffic vehicle information acquisition method based on Mask R-CNN is shown. Taking a bridge deck traffic scene as an example, the information of passing vehicles is obtained through the traffic monitoring lens arranged beside the road. The overall method framework is as figure 1 As shown, it contains the following content:

[0025] 1. Establish a traffic image database containing vehicles, select a skeleton network structure for extracting image features, and use image segmentation and labeling tools to classify vehicles in the database into multiple types, and at the same time classify wheels into one category separately. Afterwards, the Mask R-CNN network is trained, the training iteration is 30,000 times, and the learning rate is set to 2×10 before the iteration 10,000 times -3 , 2×10 between 10,000 and 20,000 times -4 , 2×10 between 20,000 and 30,000 times -4 . After training, the network has the ability to recognize vehicles and whee...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a method for obtaining traffic vehicle information based on Mask R-CNN, which can simultaneously obtain statistical information on the type, number of axles, length, vehicle speed, driving lane and vehicle number of vehicles in a traffic scene. This method first establishes a vehicle virtual detection area within the field of view of the traffic surveillance lens, and then detects the video frame by frame based on the Mask R‑CNN network. The SORT target tracking method is used to track the vehicle target entering the detection area. When the vehicle leaves the detection area, the identification value with the highest frequency in the information sequence of the vehicle type, the number of axles, and the lane obtained from the multiple frames during the vehicle tracking process in the detection area is used as the final vehicle parameter. The obtained vehicle length is averaged as the vehicle length, and then the vehicle speed is calculated according to the driving distance and time of the vehicle in the detection, and the number of passing vehicles on the corresponding lane is accumulated. The traffic vehicle information acquisition method proposed by the invention has a high degree of intelligence and can be used as an important component of smart traffic.

Description

technical field [0001] The invention relates to the field of computer vision technology and intelligent transportation, and specifically relates to a method for acquiring traffic vehicle information based on Mask R-CNN. Background technique [0002] Traffic vehicle information provides important information support for traffic planning, urban management, autonomous driving and infrastructure maintenance. The acquisition of current traffic vehicle information is mainly based on embedded sensing with embedded sensors as the core and non-contact sensing with radar, infrared, video and other technologies as the core. The advantages of the embedded sensing method are high measurement accuracy, strong stability, and less susceptible to external interference, but the corresponding equipment is expensive, difficult to replace, and cannot obtain information such as vehicle models. Non-contact sensing, especially methods based on video technology, has been extensively studied in rece...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G08G1/015G08G1/04G08G1/052G08G1/065
CPCG08G1/015G08G1/04G08G1/052G08G1/065
Inventor 张建张博许肇峰
Owner SOUTHEAST UNIV