Method for extracting suspected smoke area from dynamic smoke

A smoke and area technology, applied in image enhancement, image analysis, instruments, etc., can solve the problem of inaccurate video smoke detection algorithm, and achieve the effect of accurate extraction and reduced missed detection rate

Pending Publication Date: 2021-01-22
HARBIN UNIV OF SCI & TECH
View PDF1 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem that the existing video smoke detection algorithm is not accurate enough, and propose a method for extracting suspected smoke areas in dynamic smoke

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
  • Method for extracting suspected smoke area from dynamic smoke
  • Method for extracting suspected smoke area from dynamic smoke
  • Method for extracting suspected smoke area from dynamic smoke

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0034] A method for extracting suspected smoke areas in dynamic smoke in this embodiment, the method is implemented through the following steps: figure 1 as shown,

[0035] Step 1, preprocessing the input video image;

[0036] Perform denoising processing on the input video image, and improve the anti-interference ability of the target area by selecting the color space and extracting key frame processing steps;

[0037] Specifically:

[0038] First, input a video image and process it frame by frame;

[0039] After that, the video data set is screened and cropped to obtain the format of the data video file and a video image of uniform size;

[0040] Afterwards, the video image is normalized based on the size of the smoke data set, that is, according to the size of the smoke data set is 32x24, it is normalized into a video file with a size of 320x240. and sent to the recognition model;

[0041] Afterwards, utilize filter to carry out denoising processing to video image; Des...

specific Embodiment approach 2

[0053] The difference from the specific embodiment 1 is that in this embodiment, a method for extracting suspected smoke areas in dynamic smoke, the step of detecting the corners of moving objects described in step 2 is specifically:

[0054] A corner point is a feature point where the gray value of the pixel point changes drastically when moving in the horizontal and vertical directions. Input the preprocessed smoke video image frame by frame, and go through corner detection frame by frame to find corner points:

[0055] exist image 3 , 4 It is obvious that there will be corner points in the rectangular frame area due to smoke in the area. The following methods are used to calculate and judge the corner points of the smoke area in the image:

[0056] Record the image as I(x, y), and the similarity after translating (Δx, Δy) at point (x, y) is:

[0057]

[0058] ω(x,y) is a window centered on point (x,y), that is, a weighting function. For example, the Gaussian weightin...

specific Embodiment approach 3

[0073] The difference from the second specific embodiment is that in this embodiment, a method for extracting a suspected smoke area in dynamic smoke, the step of performing optical flow vector estimation on the corner points of the moving object described in step two is specifically:

[0074] Perform optical flow vector estimation on the corner points detected in the smoke video image, formula (4) to formula (11), can obtain the corner point optical flow vector diagram, see Figure 5 as shown,

[0075] The optical flow is defined as the instantaneous speed of pixel motion on the image plane, and the distance between the previous frame and the current frame is found according to the change of pixels in the image sequence in the time domain and the correlation between adjacent frames. The corresponding relationship exists, so as to calculate the motion information of the object between adjacent frames;

[0076] According to the degree of density of the vector in the optical fl...

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 method for extracting a suspected smoke area from dynamic smoke. An existing video smoke detection algorithm has the problem that smoke detection is not accurate enough. Theinvention belongs to the field of image recognition. The invention discloses a method for extracting a suspected smoke area from dynamic smoke, and the method comprises the following steps: firstly inputting a video image, and carrying out the frame-by-frame processing of the video image; and screening and cutting the video data set; normalizing the standard with the smog data set size of 32 * 24into a video file with the size of 320 * 240, so as to select video image blocks beneficial to the later period, and sending the video image blocks into an identification model; after normalization, selecting a proper filter to carry out denoising processing, then performing angular point detection on a moving object after partitioning is carried out, judging the moving direction, and extracting asuspected smoke area; and finally, extracting features from the area for identification. Accurate extraction of a suspected smoke area is realized, and the omission factor is reduced by 25 times. Andthe smoke identification accuracy reaches 94-96%.

Description

technical field [0001] The invention relates to a method for extracting suspected smoke areas in dynamic smoke. Background technique [0002] In video smoke detection, the accurate extraction of dynamic smoke areas and how to better preserve the integrity and irregularity of smoke have a crucial impact on the accuracy of smoke recognition. The conventional moving object extraction method is Gaussian mixture model And inter-frame difference method and optical flow method. In the smoke video image, from the perspective of algorithm performance, the inter-frame difference method is greatly affected by the environment, and the extraction effect of the smoke motion area is not obvious, and there is a problem of omission. Other algorithms are less affected. Although the dense optical flow method extracts a complete area and retains the irregular characteristics of the smoke as much as possible, the calculation speed is relatively slow, and it is difficult to achieve the needs of...

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): G06T5/00G06T7/11G06T7/136G06T7/90G06K9/46G06K9/00
CPCG06T5/002G06T7/90G06T7/11G06T7/136G06V20/40G06V10/44
Inventor 刘明珠贺雅楠
Owner HARBIN UNIV OF SCI & 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