Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Improved double-flow traffic accident detection method

A traffic accident and detection method technology, applied in the direction of instruments, biological neural network models, character and pattern recognition, etc., can solve the problems of high false positive rate and missed detection rate of models, limited accuracy of accident detection, etc.

Pending Publication Date: 2021-08-31
SOUTHEAST UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the existing feature fusion detection technology is difficult to distinguish between accident-related traffic jams and accident-unrelated traffic jams, so the false alarm rate and missed detection rate of the model are high, and the accuracy of accident detection is limited.

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
  • Improved double-flow traffic accident detection method
  • Improved double-flow traffic accident detection method
  • Improved double-flow traffic accident detection method

Examples

Experimental program
Comparison scheme
Effect test

example

[0064] Such as Figure 4 As shown, the ROC curves of the basic model and the three extended models are compared. The appearance feature extraction network in the basic model uses MobileNetV3, and does not introduce triple loss and visual attention mechanisms. Extended models 1-3 are "basic model + triple loss", "basic model + visual attention mechanism" and "basic model + triple loss + visual attention mechanism". From the ROC curve comparison chart, it can be seen that the basic model and the three extended models all maintain a high AUC value. Among them, the AUC value of the basic model was the lowest (AUC=0.93), and the AUC value of the extended model 3 was the highest (AUC=0.97). The AUC value of the extended model 1 is slightly higher than that of the basic model (AUC=0.94), but there is still a gap with the extended model 2 (AUC=0.96), which reflects that the introduction of the triplet loss is more accurate than the visual attention module. The effect is greater. F...

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 an improved double-flow traffic accident detection method, which includes the following steps: collecting accident and non-accident video data sets, an accident-related traffic jam picture data set and an accident-unrelated traffic jam picture data set, and dividing the video data sets into a training set and a test set; extracting key frames and optical flow data of each video in the training set and the test set; building a double-branch accident detection model, wherein the double-branch accident detection model comprises an appearance feature extraction network taking a key frame as input and a motion feature extraction network taking optical flow data as input; constructing a joint loss function containing triple loss and cross entropy loss to train an accident detection model; and performing validity evaluation on the trained accident detection model by using video data in the test set. According to the invention, the false alarm rate and the omission ratio of the model are reduced, and the model accident detection precision is improved.

Description

technical field [0001] The invention belongs to the field of traffic accident detection, and in particular relates to a double-flow traffic accident detection method. Background technique [0002] Traffic accidents are one of the most challenging and difficult problems in the field of transportation. Traffic accidents may cause property damage and casualties, which brings great challenges to the traffic management department. In order to effectively improve the response speed of emergency response to accidents and reduce casualties and traffic congestion caused by untimely rescue, it is of great practical significance to study the rapid detection technology of traffic accidents. [0003] Traditional accident detection techniques mainly rely on traffic flow data modeling or manual video detection. Traffic flow data modeling relies on the data quality of detectors, and it is difficult to distinguish traffic accidents from traffic jams, so the detection accuracy is low. Manu...

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 Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04
CPCG06V20/46G06N3/045G06F18/214Y02T10/40
Inventor 王晨周威夏井新陆振波许跃如
Owner SOUTHEAST UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com