CN109544464A - A kind of fire video image analysis method based on contours extract - Google Patents
A kind of fire video image analysis method based on contours extract Download PDFInfo
- Publication number
- CN109544464A CN109544464A CN201811219347.XA CN201811219347A CN109544464A CN 109544464 A CN109544464 A CN 109544464A CN 201811219347 A CN201811219347 A CN 201811219347A CN 109544464 A CN109544464 A CN 109544464A
- Authority
- CN
- China
- Prior art keywords
- image
- scene
- video
- gray level
- daytime
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003703 image analysis method Methods 0.000 title claims abstract description 6
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000012544 monitoring process Methods 0.000 claims abstract description 6
- 238000011084 recovery Methods 0.000 claims abstract description 4
- 239000000203 mixture Substances 0.000 claims abstract description 3
- 238000003708 edge detection Methods 0.000 claims description 8
- 238000009499 grossing Methods 0.000 claims description 3
- 238000000034 method Methods 0.000 description 7
- 238000011835 investigation Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 206010000369 Accident Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The fire video image analysis method based on contours extract that the present invention relates to a kind of, comprising: 1) gray processing processing is carried out to scene monitoring video on daytime;2) denoising is filtered to gray level image;3) brightness step of gray level image is calculated;4) edge thinning processing is carried out, by each pixel compared with neighbouring two pixels on gradient direction, retains local maximum;5) screen, track and connect edge: setting two threshold values of height filter out the important edges section in image by high threshold, then track remaining marginal portion on gradient direction by Low threshold, then by scanning to entire image, connect edge section;6) carry out image blend overlap-add operation, scene contour images on daytime are superimposed with the video image of night scene by pixel linear weighted function, thus realize daytime scene night video recovery.
Description
Technical field
The invention belongs to image procossings and computer vision field, provide a kind of fire video image based on contours extract
The problem of parser is substantially image segmentation.
Background technique
In today's society, fire is always one of major casualty of facing mankind.Serious fire incident can not only be made
It, can also threat to life at a large amount of casualties.In recent years, security against fire field is there has also been significant progress, however particularly serious
Fire incident still happens occasionally.Thus, avoid fire from not only needing the monitoring, early warning fire incident in time before fire occurs,
Also need the reason of accurately investigation and analysis fire occurs after fire occurs.
Traditional fire accident investigation mainly with site inspection and is collected evidence as main means, is recognized carrying out cause of fire
Periodically, often assert according to the remaining Fire Trace in scene, residue and in conjunction with witness's interrogation record by reasoning from logic
Conclusion.This conventional fire causal investigation method has very big subjectivity and limitation.With digital image processing techniques
Development, throughout major place, these monitor video images objective can specifically reflect existing safety monitoring equipment
Field situation has played good effect to identification cause of fire.But since some objective elements limit, for example fire occurs
Night, then only by monitor video, fiery mediator person has no way of determining specific location on fire.It does not still restore well at present
The method of night fire scenario.
Summary of the invention
Realize that fire scenario restores for the ease of fire investigation personnel, the present invention proposes a kind of fire based on contours extract
Video image analysis method, technical scheme is as follows:
A kind of fire video image analysis method based on contours extract, includes the steps that following:
1) gray processing processing is carried out to scene monitoring video on daytime: is weighted according to the sampled value in each channel of image flat
, the color image of triple channel is converted into single pass gray level image;
2) denoising is filtered to gray level image: initial data and Gaussian smoothing template is made into convolution, more obscured
Original image, removal can interfere with the high-frequency noise information of edge detection;
3) it calculates the brightness step of gray level image: making planar convolution using sobel operator and gray level image, from lateral and vertical
To the gradient value approximation for calculating every pixel luminance function, the brightness step of gray level image is obtained;
4) edge thinning processing is carried out, by each pixel compared with neighbouring two pixels on gradient direction, is retained
Local maximum;
5) screen, track and connect edge: setting two threshold values of height filter out the important side in image by high threshold
Rim segment, then remaining marginal portion is tracked on gradient direction by Low threshold, then by scanning to entire image, connect edge
Section;
6) image blend overlap-add operation is carried out, the video image of scene contour images on daytime and night scene is passed through into pixel
Linear weighted function is superimposed, thus realize daytime scene night video recovery.
Detailed description of the invention
Algorithm flow chart Fig. 1 of the invention
Fig. 2 is the monitor video image of scene on daytime
Fig. 3 is the monitor video image of night-time scene
Fig. 4 is reproduction image of the scene on daytime in night video
Specific embodiment
In view of the profile of video scene stores the important information of image, these information best embody the knot of fire scenario
Structure attribute, therefore the present invention proposes that a kind of fire video image analysis algorithm based on contours extract will be white by image segmentation
It scene contours extract, which comes out, to be reappeared onto the monitor video at night, and specific location on fire is positioned.Since profile is image
The important feature condition of edge detection, therefore realize scene contours extract, it needs to use edge detection algorithm.
Edge detection is the basic problem in image procossing and computer vision, it is therefore an objective to identify brightness in digital picture
Change apparent point, orients the position of image outline.Pass through boundary lines in the available video scene of edge detection algorithm
Hierarchical relationship and surround relationship, realize scene boundary profile extraction.It is mixed using image in the subsequent operation of contours extract
Superposition algorithm is closed, the lines of outline extracted is added on the monitor video image at night, to realize answering for fire scenario
Former and auxiliary positioning fire location lays the foundation for the identification of the fire origin cause of formation.
Fire video image analysis algorithm based on contours extract of the invention, technical solution are as follows:
1) gray processing processing is carried out to the Three Channel Color video image of input first, converts grayscale image for color image
Picture.Each pixel of color image possesses tri- components of R, G, B, and the value range of each component is from 0 to 255, to meter
Calculation machine post-treatment operations bring burden.Although and only one channel of gray level image, as color image still
It can reflect the entirety and local chroma luminance distribution characteristics of image, this is greatly lowered the time complexity of algorithm.With RGB
For format, common image gray processing method has:
Gray=(R+G+B)/3
Gray=0.299R+0.587G+0.114B
2) denoising is filtered to gray level image secondly by Gaussian Blur.Gaussian filtering is exactly to add to entire image
The process of weight average, the value of each pixel are by obtaining after other pixels weighted average in itself and field." fuzzy " is
Refer to by Gassian low-pass filter, by brightness value " smoothing ".Since noise also focuses on high-frequency signal, it is easy to be identified as puppet
Edge.It is excessive or miss one it should be noted that the selection of blur radius so the interference of noise can be removed using Gaussian Blur
A little weak boundary information.
3) brightness step of gray level image is then calculated.Using sobel operator, (Sobel Operator is mainly used at image
Edge detection in reason) with image make planar convolution, can obtain horizontal and vertical height difference approximation respectively, if with
A is original image, take Gx and Gy as the gray value of image of horizontal and vertical edge detection, and G indicates the knot of transverse and longitudinal direction gray value
Hop algorithm represents this gray value size, and formula is as follows:
Gradient angle, θ range from radian-π to π, then it is approximate arrive four direction, respectively represent level, it is vertical and two
A diagonal (generally 0 °, 45 °, 90 °, 135 °) falls in the gradient angle in each region to a particular value, represents four
One of direction.
4) micronization processes then are carried out to obtained a group gradient edge.By non-maxima suppression algorithm, for it is horizontal,
Vertically, two diagonal line four directions retain part by each pixel compared with neighbouring two pixels on gradient direction
Greatest gradient value.The rule for the two neighboring pixel that non-maxima suppression compares are as follows: for 0 °, compare the left side and the right;45 °,
Compare upper right and lower-left;90 °, compare bottom and upper segment;135 °, compare upper left and bottom right.
5) marginal position is determined by threshold value, screen, track and connects edge.In order to reduce false edge, using dual threashold
Value method, setting two threshold values of height, high threshold 200, Low threshold 50.If edge pixel is higher than high threshold, it is considered as side
Boundary's point can make image contain less false edge in this way, but since threshold value is excessively high, the image border of generation is likely to not close
It closes, therefore also needs to be arranged a Low threshold.Edge is linked to be profile in high threshold image, whenever reaching profile endpoint,
Meet the point of Low threshold around tracking endpoint in 8 neighborhood territory pixel points, then this point is included in new edge, until entire edge contour
Closure.
6) finally, the frame per second for saving original video saves each frame contour edge image, and with the monitoring of corresponding night scene
Video image carry out mixed linear superposition, thus realize daytime scene night video recovery.Linearity mixing mathematics is former
Reason:
G (x)=(1-a) F (x)+aQ (x)
Wherein, the value range of a is 0 to 1, and with the change of a, the transparency of two images also constantly changes, in this hair
In bright, in order not to change the original effect of lines of outline, its weight is set as 1.F (x) and Q (x) is the two width figures for participating in mixing
Picture, G (x) indicate output image, obtain to the end folded by carrying out linear weighted function to each pixel value corresponding in two images
Add image.
Claims (1)
1. a kind of fire video image analysis method based on contours extract, includes the steps that following:
1) gray processing processing is carried out to scene monitoring video on daytime: is weighted and averaged according to the sampled value in each channel of image, it will
The color image of triple channel is converted to single pass gray level image.
2) denoising is filtered to gray level image: initial data and Gaussian smoothing template is made into convolution, the original more obscured
Beginning image, removal can interfere with the high-frequency noise information of edge detection;
3) it calculates the brightness step of gray level image: making planar convolution using sobel operator and gray level image, from horizontal and vertical meter
The gradient value approximation for calculating every pixel luminance function, obtains the brightness step of gray level image;
4) edge thinning processing is carried out, by each pixel compared with neighbouring two pixels on gradient direction, retains part
Maximum value;
5) screen, track and connect edge: setting two threshold values of height filter out the important edges in image by high threshold
Section, then remaining marginal portion is tracked on gradient direction by Low threshold, then by scanning to entire image, connect edge
Section;
Carry out image blend overlap-add operation, by daytime scene contour images and night scene video image pass through pixel linearly add
Power to be superimposed, thus realize daytime scene night video recovery.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811219347.XA CN109544464A (en) | 2018-10-19 | 2018-10-19 | A kind of fire video image analysis method based on contours extract |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811219347.XA CN109544464A (en) | 2018-10-19 | 2018-10-19 | A kind of fire video image analysis method based on contours extract |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109544464A true CN109544464A (en) | 2019-03-29 |
Family
ID=65844453
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811219347.XA Pending CN109544464A (en) | 2018-10-19 | 2018-10-19 | A kind of fire video image analysis method based on contours extract |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109544464A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110223245A (en) * | 2019-05-16 | 2019-09-10 | 华南理工大学 | Blurred picture clearness processing method and system based on deep neural network |
CN110232380A (en) * | 2019-06-13 | 2019-09-13 | 应急管理部天津消防研究所 | Fire night scenes restored method based on Mask R-CNN neural network |
CN110473222A (en) * | 2019-07-02 | 2019-11-19 | 清华大学 | Image-element extracting method and device |
CN110774055A (en) * | 2019-10-10 | 2020-02-11 | 华中科技大学 | Cutter breakage monitoring method and system based on image edge detection |
CN111476807A (en) * | 2020-03-30 | 2020-07-31 | 迈克医疗电子有限公司 | Edge processing method and device for segmenting image and analysis instrument |
CN111583157A (en) * | 2020-05-13 | 2020-08-25 | 杭州睿琪软件有限公司 | Image processing method, system and computer readable storage medium |
CN111612776A (en) * | 2020-05-22 | 2020-09-01 | 福州数据技术研究院有限公司 | Automatic pathological gross specimen size measuring method based on image edge recognition |
CN111739252A (en) * | 2020-07-03 | 2020-10-02 | 徐州鑫科机器人有限公司 | Fire monitoring and automatic fire extinguishing system and working method thereof |
CN115170992A (en) * | 2022-09-07 | 2022-10-11 | 山东水发达丰再生资源有限公司 | Image identification method and system for scattered blanking of scrap steel yard |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040037465A1 (en) * | 2002-08-21 | 2004-02-26 | Krause Larry G. | System and method for detection of image edges using a polar algorithm process |
CN107452007A (en) * | 2017-07-05 | 2017-12-08 | 国网河南省电力公司 | A kind of visible ray insulator method for detecting image edge |
-
2018
- 2018-10-19 CN CN201811219347.XA patent/CN109544464A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040037465A1 (en) * | 2002-08-21 | 2004-02-26 | Krause Larry G. | System and method for detection of image edges using a polar algorithm process |
CN107452007A (en) * | 2017-07-05 | 2017-12-08 | 国网河南省电力公司 | A kind of visible ray insulator method for detecting image edge |
Non-Patent Citations (1)
Title |
---|
陈功: "鲁棒的智能视频监控方法研究" * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110223245A (en) * | 2019-05-16 | 2019-09-10 | 华南理工大学 | Blurred picture clearness processing method and system based on deep neural network |
CN110232380A (en) * | 2019-06-13 | 2019-09-13 | 应急管理部天津消防研究所 | Fire night scenes restored method based on Mask R-CNN neural network |
CN110232380B (en) * | 2019-06-13 | 2021-09-24 | 应急管理部天津消防研究所 | Fire night scene restoration method based on Mask R-CNN neural network |
CN110473222A (en) * | 2019-07-02 | 2019-11-19 | 清华大学 | Image-element extracting method and device |
CN110774055A (en) * | 2019-10-10 | 2020-02-11 | 华中科技大学 | Cutter breakage monitoring method and system based on image edge detection |
CN111476807A (en) * | 2020-03-30 | 2020-07-31 | 迈克医疗电子有限公司 | Edge processing method and device for segmenting image and analysis instrument |
CN111583157A (en) * | 2020-05-13 | 2020-08-25 | 杭州睿琪软件有限公司 | Image processing method, system and computer readable storage medium |
CN111612776A (en) * | 2020-05-22 | 2020-09-01 | 福州数据技术研究院有限公司 | Automatic pathological gross specimen size measuring method based on image edge recognition |
CN111739252A (en) * | 2020-07-03 | 2020-10-02 | 徐州鑫科机器人有限公司 | Fire monitoring and automatic fire extinguishing system and working method thereof |
CN111739252B (en) * | 2020-07-03 | 2022-03-01 | 徐州鑫科机器人有限公司 | Fire monitoring and automatic fire extinguishing system and working method thereof |
CN115170992A (en) * | 2022-09-07 | 2022-10-11 | 山东水发达丰再生资源有限公司 | Image identification method and system for scattered blanking of scrap steel yard |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109544464A (en) | A kind of fire video image analysis method based on contours extract | |
CN110246108B (en) | Image processing method, device and computer readable storage medium | |
CN109507192B (en) | Magnetic core surface defect detection method based on machine vision | |
CN110264459A (en) | A kind of interstices of soil characteristics information extraction method | |
CN102855617B (en) | Method and system for processing adaptive images | |
CN111260616A (en) | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization | |
CN107505546A (en) | A kind of method that corona discharge is monitored using ultraviolet imager | |
US9811746B2 (en) | Method and system for detecting traffic lights | |
CN105844655B (en) | A kind of laser rays stripe edge extracting method | |
WO2022027931A1 (en) | Video image-based foreground detection method for vehicle in motion | |
CN103606132A (en) | Multiframe digital image denoising method based on space domain and time domain combination filtering | |
CN106096491B (en) | Automatic identification method for microaneurysms in fundus color photographic image | |
CN111402247A (en) | Machine vision-based method for detecting defects of suspension clamp on power transmission line | |
CN108769641B (en) | Bayer color reconstruction method of video satellite image | |
CN108146093A (en) | A kind of method for removing bill seal | |
CN109214380A (en) | License plate sloped correcting method | |
CN106127124A (en) | The automatic testing method of the abnormal image signal in region, taxi front row | |
CN104021527A (en) | Rain and snow removal method in image | |
CN106228541A (en) | The method and device of screen location in vision-based detection | |
CN106657816A (en) | ORB algorithm based multipath rapid video splicing algorithm with image registering and image fusion in parallel | |
CN107292897A (en) | Image edge extraction method, device and terminal for YUV domains | |
CN106408533A (en) | Card image extraction method and card image extraction system | |
CN105787955A (en) | Sparse segmentation method and device of strip steel defect | |
CN109492587A (en) | Image processing method and device | |
CN112446889A (en) | Medical video reading method based on ultrasound |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190329 |