Improved DeepSort target detection tracking method based on YOLOv4

A target detection and detection frame technology, applied in the field of computer vision, can solve the problems of poor target tracking effect, achieve good tracking effect, reduce missed detection, and improve robustness

Inactive Publication Date: 2021-07-23
GUILIN UNIV OF ELECTRONIC TECH
View PDF6 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide an improved DeepSort target detection and tracking method based on YOLOv4, which aims to solve the problem of poor target tracking effect under weak light and occlusion

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 DeepSort target detection tracking method based on YOLOv4
  • Improved DeepSort target detection tracking method based on YOLOv4
  • Improved DeepSort target detection tracking method based on YOLOv4

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] Embodiments of the present invention are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference numerals designate the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

[0025] see Figure 1 to Figure 4 , the present invention provides an improved DeepSort target detection and tracking method based on YOLOv4, comprising:

[0026] S101 input data to obtain the detection frame of the current frame;

[0027] After the data is input, it is detected by the YOLOv4 algorithm to obtain the depth features of the detection frame and image;

[0028] The specific steps are:

[0029] S201 input target picture;

[0030] S202 performs an undistorted operation on the target image, and normalizes it, and inpu...

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 DeepSort target detection tracking method based on YOLOv4, and the method comprises the steps: inputting data, and obtaining a detection frame of a current frame; based on the detection frame of the current frame, carrying out track prediction through a Kalman filtering algorithm, and obtaining a prediction frame; performing cascade matching on the prediction frame and the detection frame of the next frame based on a Hungary algorithm; performing GIOU association matching on the tracks which fail in cascade matching; updating the trajectory based on a Kalman filtering algorithm, adding 1 to the tracking times if the target tracking succeeds, and not counting if the tracking fails; and repeating the steps, and determining that the tracking is successful if the tracking times are equal to the set times. Under the conditions of weak illumination and shielding, the method is better in tracking effect, the missing detection phenomenon is reduced, and the robustness of the system is improved.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to an improved DeepSort object detection and tracking method based on YOLOv4. Background technique [0002] The main purpose of vehicle detection and tracking is to identify vehicles in the front area of ​​interest and judge their status, and then track them. [0003] The core idea of ​​YOLO is to transform the target detection into a regression problem, use the whole picture as the input of the network, pass through a neural network, get the position of the boundingbox (bounding box) and its category, and realize end-to-end (end-to-end) -end) detection method. YOLOv1 divides a picture into S×S grids on average, and each grid is responsible for predicting the target whose center point falls within the grid. The detection speed is fast and the migration ability is strong, but the detection effect on small targets is not good; YOLOv2 Using darknet-19 as the feature extractio...

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): G06T7/246G06T7/277G06K9/62G06K9/00G06K9/32
CPCG06T7/246G06T7/277G06T2207/10016G06T2207/20081G06T2207/20084G06V20/584G06V10/25G06V2201/08G06V2201/07G06F18/22
Inventor 陈紫强张雅琼晋良念谢跃雷
Owner GUILIN UNIV OF ELECTRONIC TECH
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