Deepstream-based road event detection method

A technology of event detection and detection method, applied in the direction of traffic flow detection, electrical components, computer parts, etc., can solve the problems of large influence of meteorological conditions and high capital investment, and achieve the effect of ensuring reliability

Pending Publication Date: 2022-07-08
重庆亲禾智千科技有限公司
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Potentially, the direct AID method is much better than the indirect AID method in terms of discriminative speed, especially in the case of low traffic volume and can also detect sudden traffic events well, but requires a more intensive setup Traffic video detectors require high capital investment to ensure the reliability of discrimination, and weather conditions also have a greater impact on it

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
  • Deepstream-based road event detection method
  • Deepstream-based road event detection method
  • Deepstream-based road event detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] The present invention is a deepstream-based road event detection method. The detection method takes a streaming media camera as an input source, transmits it to a gstreamer pipeline constructed by deepstream in the form of an rtsp stream, utilizes a plug-in in the pipeline, and uses a target detection neural network to detect The position of the object to be detected in each frame is obtained, and the target detected in different frames is detected using an algorithm to assign the relationship of the target, so as to achieve the purpose of tracking the trajectory of the same object.

[0023] In this embodiment, the algorithm expresses the cost between the target and the assigned object as a matrix, and this algorithm expresses the cost between the goal and the assigned object as a matrix, for example, the first row and the third column represent the first goal , the cost of assigning to the third object, if it is not assignable, the value is the maximum value of the 32-b...

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 deepstream-based road event detection method, and belongs to the technical field of road event detection, and the detection method comprises the steps: employing a streaming media camera as an input source, employing a target detection neural network to detect the position of a to-be-detected object of each frame, employing an algorithm to carry out the distribution of the affiliation of targets detected in different frames, and carrying out the detection of the to-be-detected object. Therefore, the purpose of tracking the trajectory of the same object is achieved. According to the method, platformization of the road traffic events can be realized in a targeted manner, and various information of the road traffic events can be obtained in time; related personnel can make corresponding decisions conveniently, and more economic losses are avoided.

Description

technical field [0001] The invention belongs to the technical field of road event detection, in particular to a deepstream-based road event detection method. Background technique [0002] The early traffic incident detection methods mainly detected the traffic incidents that have occurred on the road. Since the 1960s, a variety of traffic incident detection methods have been formed, mainly including the following eight: (1) The driver's mobile phone call; (2) Incident management personnel watching CCTV (Closed Circuit Tv, CCTV) surveillance images; (3) Drivers' help calls or roadside emergency calls; (4) Traffic police patrols; (5) Traffic department or other unit staff Report via walkie-talkie; (6) Traffic status report service; (7) Fleet (bus, truck) report; (8) Automatic detection method based on traffic detection device. These methods can be divided into two categories: manual detection methods and automatic detection methods. [0003] The manual detection method is th...

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): G08G1/01H04L65/65H04L65/60G06K9/62
CPCG08G1/0125G06F18/2321
Inventor 刘宽李悦徐彬峰周积鑫
Owner 重庆亲禾智千科技有限公司
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
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap