Foreground target extraction method of dynamic background video image

A video image and dynamic background technology, applied in the field of computer vision, can solve the problem of noise interference of foreground objects, achieve the effect of simplifying the algorithm, facilitating real-time application, and reducing correlation errors

Inactive Publication Date: 2018-05-01
SHANGHAI UNIVERSITY OF ELECTRIC POWER
View PDF4 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The Gaussian modeling method is simple, but its disadvantage is that the extracted foreground target has a lot of noise interference

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
  • Foreground target extraction method of dynamic background video image
  • Foreground target extraction method of dynamic background video image
  • Foreground target extraction method of dynamic background video image

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0044] In order to more accurately extract the foreground target under the dynamic background, the present application proposes a video tracking method based on a continuous observation hidden Markov model. The main idea of ​​video tracking is: use a rectangular frame to delineate the target in the first frame, such as a pedestrian. In the second frame, select rectangular boxes of different sizes at different positions, compare with the target in the box in the first frame, if the content of which box in the second frame is most similar to the content in the first frame, then Just select the content in this box as the target tracked in the second frame. Then start the third frame, and so on. The core issue of video tracking is data association, the essence of which is to establish a hidden Markov model (HMM). Afterwards, the problem of data association is transformed into the problem of identification and distribution of motion trajectories, and the hidden Markov model is us...

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 a foreground target extraction method of a dynamic background video image. The method comprises that S1) human bodies in a video frame are modeled via a continuous observed value hidden Markov model to obtain multiple human body models; S2) observation vectors of each the human body model in a set time period form an observation sequence; and S3) the conditional probability that each human body model generates the observation sequence is calculated via a forward-backward algorithm, and the observation sequence with the largest conditional probability is selected as a foreground target. Compared with the prior art, a Gaussian model and the continuous observed value hidden Markov model are combined and applied to modeling of the observation target in the video, and the condition that background modeling is much different from the practical background is avoided.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a method for extracting a foreground object from a dynamic background video image. Background technique [0002] Video surveillance is the most important means of information acquisition in China's security industry. The monitoring system monitors multiple production sites in real time through cameras installed at different production sites, improving the efficiency of supervision. Motion detection plays an integral role in video surveillance as a preprocessing step in many computer vision applications. The moving object in motion detection is called the foreground, which is the area of ​​interest. The purpose of motion detection is to extract the foreground area in a video stream for further processing such as target recognition, tracking, and behavior analysis. How to efficiently, quickly and accurately extract the foreground target information in the surveillance video is a v...

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): G06T7/194G06T7/215G06T7/277
CPCG06T2207/10016G06T7/194G06T7/215G06T7/277
Inventor 吕学勤王裕东钦超瞿艳
Owner SHANGHAI UNIVERSITY OF ELECTRIC POWER
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