A Machine Learning-Based Video Anomaly Detection Method

An anomaly detection and machine learning technology, applied in instruments, closed-circuit television systems, computer components, etc., can solve the problems of insufficient video anomaly detection methods, lack of self-learning and improvement capabilities, etc., to achieve comprehensive anomaly detection types and functional expansion. Good performance, reducing the effect of video misjudgment

Active Publication Date: 2017-08-11
ZHEJIANG UNIV OF TECH
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In view of the fact that the existing video anomaly detection methods are still not comprehensive enough and lack the ability of self-learning and improvement, the present invention proposes a video anomaly detection method based on machine learning. The method first reads the video file, decomposes the video into frame-by-frame images, and then Detect the decomposed image data; if the user makes a misjudgment during use, the misjudgment video can be classified and stored according to the abnormal type as a learning memory. When the same type of abnormality occurs again, it will be called for comparison to prevent it from happening again. The same misjudgment occurred

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
  • A Machine Learning-Based Video Anomaly Detection Method
  • A Machine Learning-Based Video Anomaly Detection Method
  • A Machine Learning-Based Video Anomaly Detection Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0027] The present invention will be further described below in conjunction with the accompanying drawings.

[0028] refer to figure 1 and figure 2 , a kind of video anomaly detection method based on machine learning, described method comprises the following steps:

[0029] 1) Video file read-in: the video file is read-in method with frame-by-frame bmp image data;

[0030] 2) Video anomaly detection: detect the decomposed image, and the types of anomalies that can be detected are

[0031] a) The picture is too bright and the picture is too dark: The overall brightness and darkness of the image picture depends on the lens grayscale of the image. First grayscale the image f(x,y):

[0032] f(x,y)=0.114*B(x,y)+0.587*G(x,y)+0.299*R(x,y),

[0033] Among them, B(x,y), G(x,y), and R(x,y) correspond to the blue, green and red component values ​​of the image at the pixel (x,y) respectively. Set the brightness threshold light and the darkness threshold dark. When f(x,y)>light, the...

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

A video anomaly detection method based on machine learning, comprising the following steps: 1) video file reading: read the video file as bmp image data frame by frame; 2) video anomaly detection: detect the decomposed image , the process is as follows: a) the picture is too bright, the picture is too dark; b) the gain is disordered; c) blurred and blocked; d) striped interference and scrolling; e) snowflake interference; f) jitter; g) color cast; h ) freeze; i) blue, black screen; 3) machine learning: compare the similarity of the video judged as abnormal with the same type of misjudged video in the database according to the abnormal code, and judge whether the current video is misjudged. The invention has comprehensive detection, self-learning and improving ability, and high accuracy.

Description

technical field [0001] The invention relates to the technical fields of image processing, video similarity calculation, machine learning, etc., and the main content is a video anomaly detection method based on machine learning. Background technique [0002] With the development of our society and the rise of the electronic field, video surveillance systems have been widely used in all walks of life, and are no longer limited to special fields such as public security, finance, banking, transportation, military and ports in the past. Communities, office buildings, hotels, public places, factories, shopping malls, communities, and even homes within easy reach of life have all installed video surveillance systems. However, as the number of monitoring cameras in the application continues to increase and the monitoring time continues to prolong, it is becoming more and more difficult to maintain the monitoring software manually in real time, so the intelligence of the monitoring s...

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): H04N7/18G06K9/00
Inventor 张永良张智勤董灵平阮盛鹏肖刚
Owner ZHEJIANG UNIV OF 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