Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Abnormal event classification model construction method and detection method based on video monitoring

A technology of abnormal events and video monitoring, applied in the direction of biological neural network model, neural architecture, computer components, etc., can solve the problems of limited application range, false abnormal time affecting detection accuracy, etc., and achieve the effect of reducing early warning

Pending Publication Date: 2021-08-27
山东新一代信息产业技术研究院有限公司
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Most of the existing anomaly detection methods learn a model based on outlier detection in the training video, there may be false anomaly time, which affects the accuracy of detection, and the application range is limited

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
  • Abnormal event classification model construction method and detection method based on video monitoring

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] The abnormal event classification model construction method based on video monitoring of the present invention, the motion and appearance of the object are characterized, and the event is classified based on the SVM classifier of multi-classification. The method includes the following steps:

[0037] S100. Detecting the target in the video frame by frame to obtain a bounding box of each object;

[0038] S200. Clip the object based on the bounding box to obtain a clipped image, and convert the clipped image into a grayscale image;

[0039] S300. Construct a learning network model based on the CNN architecture, train the learning network model with the above-mentioned grayscale image as input, obtain the learning network model after training, and output appearance features and motion features, and the above appearance features and motion features form an event data set ;

[0040] S400. After clustering the normal event data set by the K-Means clustering algorithm, select...

Embodiment 2

[0051] The abnormal event detection method based on video surveillance of the present invention comprises the following steps:

[0052] Obtain the learning network model after training and the SVM model after training through the abnormal event classification model construction method based on video monitoring disclosed in embodiment 1;

[0053]After training, the network model is learned to learn the features of the video to be tested, and the appearance features and motion features are obtained. The above appearance features and motion features are used as a data set to input the trained SVM model for event classification and identification.

[0054] For each frame of video image of the video to be tested, the highest score obtained by the SVM model is used as the abnormal score of the video image, and the abnormal score is temporarily smoothed by a Gaussian filter. This method can reduce the early warning of false abnormal events and can be used for real-time monitoring abn...

Embodiment 3

[0058] In the computer-readable medium of the present invention, computer instructions are stored on the computer-readable medium, and when the computer instructions are executed by a processor, the processor executes the method disclosed in Embodiment 1. Specifically, a system or device equipped with a storage medium may be provided, on which a software program code for realizing the functions of any of the above embodiments is stored, and the computer (or CPU or MPU of the system or device) ) to read and execute the program code stored in the storage medium.

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 an abnormal event classification model construction method and a detection method based on video monitoring, belongs to the technical field of video monitoring, and aims to solve the technical problem of how to reduce early warning of false abnormal events and realize real-time monitoring abnormal event detection in various scenes. The method comprises the following steps: detecting a target in a video frame by frame; cutting the object based on the bounding box, and converting the cut image into a grayscale image; training a learning network model by taking the grayscale image as input to obtain a trained learning network model, and outputting appearance features and motion features; after the normal event data sets are clustered through a K-Means clustering algorithm, normal event data located at abnormal points are selected as a false abnormal event data set, and other normal event data sets and abnormal event data sets are combined to form a sample training set; and training the SVM model based on the sample training set to obtain a trained SVM model.

Description

technical field [0001] The invention relates to the technical field of video monitoring, in particular to a video monitoring-based abnormal event classification model building method and detection method. Background technique [0002] Abnormal event monitoring technology in surveillance video is an important branch of intelligent surveillance video analysis. The main problem it solves is to use computers to analyze video data and learn and understand the behaviors and actions that occur in the video. Since the monitored video itself has the characteristics of high resolution, high redundancy, and large capacity, using the method of intelligent abnormal event detection can reduce the cost of manual observation and detection of video, and reduce the safety hazards caused by manual negligence. Therefore, abnormal event detection technology in surveillance video has very important research and application value. [0003] The current abnormal event detection technology uses manu...

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): G06K9/00G06K9/62G06N3/04
CPCG06V20/41G06V20/44G06N3/045G06F18/23213G06F18/2411G06F18/214
Inventor 王雯哲高岩王建华高明
Owner 山东新一代信息产业技术研究院有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com