Crowd counting system and method, device and storage medium

A crowd counting and crowd counting technology, applied in the field of computer vision, can solve the problems of large differences in pedestrian scales and aggravate the perspective effect of acquisition equipment, and achieve the effects of accelerating calculation, increasing nonlinear expression ability, and expanding receptive field

Active Publication Date: 2022-02-08
山东力聚机器人科技股份有限公司
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This exacerbates the inherent perspective effect of the collection equipment, which makes the problem that the scale of pedestrians in the collected crowd videos or images more prominent.
The scale of pedestrians varies greatly, which brings great challenges to the model to accurately count the number of people

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
  • Crowd counting system and method, device and storage medium
  • Crowd counting system and method, device and storage medium
  • Crowd counting system and method, device and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] This embodiment proposes a crowd counting system. The system redesigns a brand-new crowd counting network combined with the VGG network, which can dig out rich detailed information in crowd images, which is conducive to improving the accuracy of the counting model. figure 1 It is a schematic structural diagram of a crowd counting system provided by an embodiment of the present invention. like figure 1 As shown, the system includes: an encoder 110 , a multi-scale feature residual fusion module 120 and a decoder 130 .

[0056] The encoder 110 is configured to acquire a crowd image, extract primary features of the crowd image, and obtain a primary feature map of the crowd image.

[0057] The multi-scale feature residual fusion module 120 is connected to the encoder 110, and is configured to iteratively extract the multi-scale features of the crowd image in the form of residual fusion based on the primary feature map to obtain the details of the crowd image feature map. ...

Embodiment 2

[0098] This embodiment provides a crowd counting method. Firstly, the crowd counting system described in Embodiment 1 is trained, and then the trained crowd counting system is used to realize the crowd counting of crowd images, so as to improve the large scale change of pedestrians and crowd occlusion in crowd counting. Counting performance in severe cases. Figure 4 It is a flowchart of a crowd counting method provided by an embodiment of the present invention. like Figure 4 As shown, the method includes steps S10-S40.

[0099] S10: Obtain a plurality of training crowd images; mark the positions of heads in each training crowd image, and use a two-dimensional Gaussian distribution with fixed parameters to fit the heads to obtain a truth map of each training crowd image.

[0100] S20: Construct the crowd counting system described in Embodiment 1.

[0101] S30: Using the plurality of training crowd images as a training set and using the corresponding plurality of ground-tru...

Embodiment 3

[0115] Image 6 It is a schematic structural diagram of a computer device provided by an embodiment of the present invention. like Image 6 As shown, the device includes a processor 610 and a memory 620 . The number of processors 610 may be one or more, Image 6 A processor 610 is taken as an example.

[0116] The memory 620, as a computer-readable storage medium, can be used to store software programs, computer-executable programs and modules, such as program instructions / modules of the crowd counting method described in Embodiment 2 of the present invention.

[0117] Correspondingly, the processor 610 implements the crowd counting training method described in Embodiment 2 of the present invention by running the software programs, instructions and modules stored in the memory 620 .

[0118] The memory 620 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application program required by ...

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 crowd counting system and method, equipment and storage medium. The system includes: an encoder, configured to acquire a crowd image, extract primary features of the crowd image, and obtain a primary feature map of the crowd image; a multi-scale feature residual fusion module, connected to the encoder, is configured to be based on The primary feature map iteratively extracts the multi-scale features of the crowd image in the manner of residual fusion to obtain the detailed feature map of the crowd image; the decoder is connected to the multi-scale feature residual fusion module and set In order to perform resolution recovery and dimensionality reduction on the detailed feature image, a crowd density map of the crowd image is obtained. The present invention uses the detail enhancement module to extract the detailed features of the crowd, and improves the feature expression ability of the network.

Description

technical field [0001] The embodiments of the present invention relate to the field of computer vision, and in particular to a crowd counting system and method, equipment and storage medium. Background technique [0002] With the continuous advancement of urbanization, the population density in cities continues to increase. There are more and more crowds and crowds in public places such as transportation hubs, large-scale event sites, and large shopping malls, and there are more and more hidden dangers caused by excessive crowding and congestion. The density of people is closely related to public safety. The greater the flow of people, the higher the probability of stampedes and other dangerous events. At the same time, the country has continuously increased its emphasis on social and public safety, and various surveillance cameras have been deployed in every corner of public places. A large amount of surveillance video data is coming, but it is very unrealistic to count r...

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): G06K9/00G06K9/62G06N3/04
CPCG06N3/045G06F18/253G06F18/214
Inventor 杨光远孟令灿牛德奎孔妍王子璇张梅玉
Owner 山东力聚机器人科技股份有限公司
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