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

Multi-view pedestrian detection method based on feature fusion

A technology of feature fusion and pedestrian detection, applied in the field of image processing, can solve the problems of many restrictions, occupy time, occupy memory and time resources, etc., to achieve the effect of high efficiency and avoid accuracy loss

Pending Publication Date: 2022-04-29
NORTHEASTERN UNIV
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the reconstruction process takes up a lot of time, and the matching operation also takes up a lot of memory and time resources. There are many restrictions in actual use, so how to use multi-view information without taking up too much time has become our main consideration.

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
  • Multi-view pedestrian detection method based on feature fusion
  • Multi-view pedestrian detection method based on feature fusion
  • Multi-view pedestrian detection method based on feature fusion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The invention will be further described below in conjunction with the accompanying drawings and specific implementation examples.

[0039] Such as figure 1 As shown, a multi-view pedestrian detection method based on feature fusion, including:

[0040] Step 1: Collect image data of pedestrians from multiple angles as a sample set; including:

[0041] Step 1.1: Collect image data from N angles of view, and set N cameras to collect image data of pedestrians passing through a certain intersection within a period of time;

[0042] Step 1.2: Take N images collected at the same time as a set of sample data;

[0043] Convert the pixel coordinates (upper left, lower right) corresponding to each target in the N viewing angles to the corresponding map coordinates; and determine the internal and external parameters of the camera.

[0044] Step 2: If image 3 As shown, construct a neural network model integrating feature fusion and key point detection, and use the sample set to ...

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 provides a multi-view pedestrian detection method based on feature fusion, and relates to the technical field of image processing, a double-branch feature fusion structure is introduced on the basis of a lightweight network Resnet18, feature extraction is further refined, a direct addition mode is adopted for fusion, the number of channels in the fusion process is not changed, and meanwhile, the idea of key points is introduced, so that the feature extraction efficiency is improved. The central point is used for replacing a positive sampling point, and a large number of candidate frames do not need to be generated, so that subsequent non-maximum suppression operation does not exist, the model is more flexible, the parameter quantity is reduced, the operation speed of the model is increased, and the negative influence of an anchor frame mechanism on the model is avoided; and aggregating the obtained feature maps together, so that the finally generated aerial view can aggregate multi-camera pedestrian information.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a multi-view pedestrian detection method based on feature fusion. Background technique [0002] Whether it is for Advanced Driving Assistance System (ADAS, Advanced Driving Assistance System) or automatic driving, pedestrian detection has always been an unavoidable problem. During the operation of the vehicle, if the position of each pedestrian can be accurately detected, the safety of the autonomous driving field will be greatly guaranteed. Therefore, pedestrian detection has been a research hotspot in the field of autonomous driving for the past 20 years. The mobility of pedestrians makes them less predictable than that of vehicles ahead. In the task of pedestrian detection, occlusion has always been the focus of attention. For example, at an intersection, pedestrians may be blocked by passing vehicles or other pedestrians, making it difficult to predict the location...

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): G06V20/56G06K9/62G06N3/04G06N3/08G06V10/774G06V10/80G06V10/82
CPCG06N3/084G06N3/045G06F18/214G06F18/253
Inventor 刘宇红韩春燕任涛
Owner NORTHEASTERN UNIV
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com