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

Electric power overhaul pedestrian multi-target tracking algorithm based on CSSD

A multi-target tracking and power maintenance technology, applied in the field of deep learning target detection and tracking, can solve problems affecting the accuracy of pedestrian targets

Inactive Publication Date: 2020-12-15
SHANDONG LUNENG SOFTWARE TECH +1
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, different lighting levels, differences in the angle and size of human figures in each frame of pictures, and pedestrian occlusion will affect the accuracy of pedestrian target tracking. The current calculation scheme and method cannot solve the above problems well.

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
  • Electric power overhaul pedestrian multi-target tracking algorithm based on CSSD
  • Electric power overhaul pedestrian multi-target tracking algorithm based on CSSD
  • Electric power overhaul pedestrian multi-target tracking algorithm based on CSSD

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0026] Below in conjunction with accompanying drawing and embodiment the present invention will be further described:

[0027] figure 1 Shown is the algorithm model flow chart of the present invention, wherein the part in the box is the content of the traditional SSD algorithm model, and the rest is a CSSD-based power maintenance pedestrian multi-target tracking algorithm, including the following steps:

[0028] S1. The video stream images are decoded, the surveillance images are obtained, and processed into images of continuous frames.

[0029] S2. Input the image data obtained in step S1 into the SSD feature extraction network to obtain a feature map.

[0030] S3. Pass the feature map through the auxiliary convolutional layer to obtain a multi-scale feature map, and perform feature category classification and coordinate regression on the multi-scale feature map.

[0031] S4. Map the pedestrian coordinates to the feature map in step S2 to obtain the pedestrian feature vecto...

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 relates to the field of deep learning target detection and tracking, in particular, relates to an electric power overhaul pedestrian multi-target tracking algorithm based on a CSSD, andprovides the electric power overhaul pedestrian multi-target tracking algorithm based on the CSSD; a CSSD network is used for detecting persons in an image and feature vectors of the personnel in theimage are temporarily stored. Cosine similarity evaluation is performed in combination with the feature patterns of all persons to perform matching of front and back frames of persons so as to realizetracking of the persons in a video stream, so target size and angle change and the like are adapted to the greatest extent and pedestrian target tracking accuracy is effectively improved.

Description

technical field [0001] The invention relates to the field of deep learning target detection and tracking, in particular to a CSSD-based pedestrian multi-target tracking algorithm for power maintenance. Background technique [0002] Safety monitoring in power maintenance scenarios is crucial to production safety, and pedestrian safety monitoring and tracking under surveillance video is an important research direction for large-scale intelligent monitoring systems in power maintenance scenarios. However, different levels of light and darkness, differences in the angle and size of figures in each frame of pictures, and pedestrian occlusion will affect the accuracy of pedestrian target tracking. The current calculation schemes and methods cannot solve the above problems well. Contents of the invention [0003] In order to solve the above problems, based on the pedestrian multi-target tracking technology, on the basis of the original SSD (Single Shot MultiBox Detector) algorith...

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): G06K9/00G06K9/62G06N3/04G06N3/08G06Q10/00G06Q50/06
CPCG06N3/08G06Q10/20G06Q50/06G06V40/10G06V20/41G06N3/045G06F18/22
Inventor 公凡奎沈茂东张俊岭周伟何成高宏马超裴健张波刘海威苏彪田佳
Owner SHANDONG LUNENG SOFTWARE TECH