CN106447685A - Infrared tracking method - Google Patents
Infrared tracking method Download PDFInfo
- Publication number
- CN106447685A CN106447685A CN201610803370.8A CN201610803370A CN106447685A CN 106447685 A CN106447685 A CN 106447685A CN 201610803370 A CN201610803370 A CN 201610803370A CN 106447685 A CN106447685 A CN 106447685A
- Authority
- CN
- China
- Prior art keywords
- template
- point
- edge
- region
- field picture
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000003044 adaptive effect Effects 0.000 claims abstract description 24
- 238000010168 coupling process Methods 0.000 claims description 17
- 230000008878 coupling Effects 0.000 claims description 16
- 238000005859 coupling reaction Methods 0.000 claims description 16
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 230000002776 aggregation Effects 0.000 claims description 3
- 238000004220 aggregation Methods 0.000 claims description 3
- 238000003708 edge detection Methods 0.000 abstract description 7
- 230000002146 bilateral effect Effects 0.000 abstract 1
- 238000001514 detection method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 1
Classifications
-
- 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
- G06T2207/10048—Infrared image
Landscapes
- Image Analysis (AREA)
Abstract
The embodiments of the invention disclose an infrared tracking method which is based on Hausdorff distance matching and is adaptive to templates. The method includes the following steps: firstly, conducting edge detection on an infrared image that is input by applying improved sobel edge detection operator, extracting profile characteristics of an object; secondly, using rapid bilateral Hausdorff distance method to conduct template matching on an object template and a to-be-searched area in tracking; finally, conducting self-adaptive template updating on an optimal matching result. According to the invention, the method can conduct stable and accurate tracking on the infrared object in a long time, has rapid operation speed, and has strong real-timeness and robustness.
Description
Technical field
The present invention relates to digital image processing field, especially relate to one kind and be based on Hao Siduofu (Hausdorff) distance
The infrared track method of the adaptive template of coupling.
Background technology
Nowadays, with the continuous lifting of information technology and computing power, increasing people has started to computer
Target motion problems in vision are studied.In infrared image processing field, infrared object tracking intelligent security-protecting and monitoring with
And infrared guidance technology research field develops rapidly, exigent prison in the real-time particularly followed the tracks of at some and precision
In control and operational environment, the research to new algorithm has great meaning.
Infrared object tracking is in Engineering Control, traffic monitoring, medical image research, automated navigation system, astronomical monitoring etc.
There is critically important practical value in field.Particularly in military aspect, infrared guidance become more and more important main battle weaponry it
One.The algorithm of most of infrared object trackings also has the weak point of many at present, and more complicated algorithm is in real-time target
Tracking aspect does not reach requirement, and single algorithm can not carry out steady in a long-term being tracked again, and this is for modernization information war
Striving is totally unfavorable impact, so the current people research that becomes of the infrared object tracking algorithm of research real-time high-efficiency is important
Problem.
Infrared object tracking algorithm based on Hausdorff distance coupling is that contour feature based on infrared target carries out mesh
Mark modeling, can be very good to carry out robust tracking for specific target appearance feature, adds adaptive To Template more
Newly, the interference that can effectively reduce noise is to the impact followed the tracks of.But when current infrared target be blocked, dimensional variation the problems such as
Also need to deeper into research.
Content of the invention
It is an object of the invention to provide a kind of infrared track side of the adaptive template based on Hausdorff distance coupling
Method, for infrared object tracking real-time good, follow the tracks of stable and strong robustness.
Technical scheme disclosed by the invention is：Using improved Sobel edge edge detective operators, the infrared image of input is entered
Row rim detection, extracts the contour feature of target, when following the tracks of using quickly two-way Hausdorff distance method to target mould
Plate and region to be searched carry out template matching, adaptive template renewal is carried out to optimal matching result thus carry out target with
Track, comprises the following steps that：
Step 1：Input infrared video, is carried out artificial selected target region, is entered using improved Sobel operator to initial frame
Row binary conversion treatment, obtains the edge contour information of target and sets up To Template；
Step 2：Read next two field picture, the Sobel operator that region to be detected is improved is processed, after binaryzation,
Obtain edge contour information；
Step 3：The two-way Hausdorff distance that To Template after Sobel and region to be detected are improved
Coupling, finds out the position of best match；
Step 4：Using the renewal of adaptive template conditional judgment To Template, if meeting condition, update To Template
And enter next frame, otherwise do not update To Template and enter next frame until Infrared video image terminates.
Further, in step 1, binary conversion treatment is carried out using improved Sobel operator, will edge pixel gray value
It is set to 1, rest of pixels value is set to 0, obtain the edge contour information of target and set up To Template, specific as follows：
The Sobel edge edge detection process that improve of infrared image gray value f (i, j) to input, wherein i, j divide and are changed into
Image pixel is horizontal, ordinate value, and Sobel operator is the edge detection operator of first derivative, and formula is as follows：
Sx=(Z1+2Z2+Z3)-(Z7+2Z8+Z9)
Sy=(Z1+2Z4+Z7)-(Z3+2Z6+Z9)
Wherein, Z1,Z2,Z3,Z4,Z5,Z6,Z7,Z8,Z9It is respectively 8 neighborhood territory pixel gray values of pixel f (i, j), Sx,SyPoint
Wei not level, the gradient magnitude of vertical direction.
Improved Sobel operator increased the gradient calculation in positive and negative 45 ° of directions, and formula is as follows：
S45°=(Z2+2Z3+Z6)-(Z4+2Z7+Z8)
S-45°=(Z2+2Z1+Z4)-(Z6+2Z8+Z9)
Work as Sx,Sy,S45°,S-45°In any one is more than during default threshold value T for as edge, the side of record target area
The coordinate figure of edge pixel sets up To Template.
Further, in step 3, the To Template after Sobel and region to be detected are improved is two-way
Hausdorff distance coupling, specific as follows：
To the marginal information set A={ a in To Template1,a2,...,amAnd region candidate template to be measured edge letter
Breath set B={ b1,b2,...,bnCarrying out two-way Hausdorff distance coupling, formula is as follows：
H (A, B)=max (h (A, B), h (B, A))
Wherein, a1,a2,...,amFor the coordinate information of To Template edge aggregation, b1,b2,...,bmFor candidate's mould to be measured
The edge coordinate of plate,
H (A, B)=max (a ∈ A) min (b ∈ B) | | a-b | |
H (B, A)=max (b ∈ B) min (a ∈ A) | | b-a | |
In above formula (6), h (A, B) represents the point a from set AiDistance set to the point of the set B nearest apart from this point enters
Row sequence, takes maximum therein, and in above formula, h (B, A) represents the point b from set BiTo the point putting nearest set A apart from this
Distance set be ranked up, take maximum therein, | | | | be apart from normal form.
When the Hausdorff improving is apart from matching primitives, to current point ai8 contiguous range in look for another set
Whether there is the point of this scope, if a in coordinateiThe corresponding set B of four neighborhood position up and down in have at least one point corresponding,
Then calculate thisiPoint is 1 to set B point Hausdorff distance, otherwise continually looks for aiHave corresponding in the corresponding set B of eight neighborhood point
At least one point, then calculate thisiPoint is 2 to set B point Hausdorff distance, and otherwise distance is just designated as 10.
Further, the renewal of the use adaptive template conditional judgment To Template in step 4, detailed process is as follows：
By calculating the edge point set sum S in To TemplateAAnd the marginal point of best match position region template
The total S of setB, calculate Adaptive template-updating conditional parameter P1=SA/SB, and the gray scale of To Template region original image
Average hAOriginal image gray average h with best match position region templateB, calculate Adaptive template-updating conditional parameter P2
=hA/hB, and if only if meets condition α ＜ P1＜ β and condition γ ＜ P2During ＜ λ, renewal current region template is To Template,
Otherwise do not update, wherein α, beta, gamma, λ is constant.Thus reducing the interference of noise, carry out real-time, stable infrared object tracking.
Brief description
Fig. 1 is the stream of the infrared track method of the adaptive template based on Hausdorff distance coupling of the inventive method
Cheng Tu.
Fig. 2 is the Sobel filter template schematic diagram both horizontally and vertically of the inventive method.
Fig. 3 is the Sobel filter template schematic diagram in positive and negative 45 degree of directions of the inventive method.
Specific embodiment
Describe the adaptive template based on Hausdorff distance coupling of the inventive method below in conjunction with accompanying drawing in detail
Infrared track method specific implementation process.
As shown in figure 1, in the inventive method, a kind of adaptive template based on Hausdorff distance coupling infrared with
Track method carries out rim detection using improved Sobel edge edge detective operators to the infrared image of input, extracts the profile of target
Feature, carries out template using quickly two-way Hausdorff distance method to To Template and region to be searched when following the tracks of
Join, adaptive template renewal being carried out to optimal matching result thus carrying out target following, comprising the following steps that：
Step 1：Input infrared video, is carried out artificial selected target region, is entered using improved Sobel operator to initial frame
Row binary conversion treatment, obtains the edge contour information of target and sets up To Template；
Step 2：Read next two field picture, the Sobel operator that region to be detected is improved is processed, after binaryzation,
Obtain edge contour information；
Step 3：The two-way Hausdorff distance that To Template after Sobel and region to be detected are improved
Coupling, finds out the position of best match；
Step 4：Using the renewal of adaptive template conditional judgment To Template, if meeting condition, update To Template
And enter next frame, otherwise do not update To Template and enter next frame until Infrared video image terminates.
Further, in step 1, binary conversion treatment is carried out using improved Sobel operator, will edge pixel gray value
It is set to 1, rest of pixels value is set to 0, obtain the edge contour information of target and set up To Template, specific as follows：
The Sobel edge edge detection process that improve of infrared image gray value f (i, j) to input, wherein i, j divide and are changed into
Image pixel is horizontal, ordinate value, and Sobel operator is the edge detection operator of first derivative, and formula is as follows：
Sx=(Z1+2Z2+Z3)-(Z7+2Z8+Z9)
Sy=(Z1+2Z4+Z7)-(Z3+2Z6+Z9)
Wherein, Z1,Z2,Z3,Z4,Z5,Z6,Z7,Z8,Z9It is respectively 8 neighborhood territory pixel gray values of pixel f (i, j), Sx,SyPoint
Wei not level, the gradient magnitude of vertical direction.
Improved Sobel operator increased the gradient calculation in positive and negative 45 ° of directions, and formula is as follows：
S45°=(Z2+2Z3+Z6)-(Z4+2Z7+Z8)
S-45°=(Z2+2Z1+Z4)-(Z6+2Z8+Z9)
Work as Sx,Sy,S45°,S-45°In any one is more than during default threshold value T for as edge, the side of record target area
The coordinate figure of edge pixel sets up To Template.
Further, in step 3, the To Template after Sobel and region to be detected are improved is two-way
Hausdorff distance coupling, specific as follows：
To the marginal information set A={ a in To Template1,a2,...,amAnd region candidate template to be measured edge letter
Breath set B={ b1,b2,...,bnCarrying out two-way Hausdorff distance coupling, formula is as follows：
H (A, B)=max (h (A, B), h (B, A))
Wherein, a1,a2,...,amFor the coordinate information of To Template edge aggregation, b1,b2,...,bmFor candidate's mould to be measured
The edge coordinate of plate,
H (A, B)=max (a ∈ A) min (b ∈ B) | | a-b | |
H (B, A)=max (b ∈ B) min (a ∈ A) | | b-a | |
In above formula, h (A, B) represents the point a from set AiDistance set to the point of the set B nearest apart from this point is carried out
Sequence, takes maximum therein, and in above formula, h (B, A) represents the point b from set BiTo the point putting nearest set A apart from this
Distance set is ranked up, and takes maximum therein, | | | | it is apart from normal form.
When the Hausdorff improving is apart from matching primitives, to current point ai8 contiguous range in look for another set
Whether there is the point of this scope, if a in coordinateiThe corresponding set B of four neighborhood position up and down in have at least one point corresponding,
Then calculate thisiPoint is 1 to set B point Hausdorff distance, otherwise continually looks for aiHave corresponding in the corresponding set B of eight neighborhood point
At least one point, then calculate thisiPoint is 2 to set B point Hausdorff distance, and otherwise distance is just designated as 10.
Further, the renewal of the use adaptive template conditional judgment To Template in step 4, detailed process is as follows：
By calculating the edge point set sum S in To TemplateAAnd the marginal point of best match position region template
The total S of setB, calculate Adaptive template-updating conditional parameter P1=SA/SB, and the gray scale of To Template region original image
Average hAOriginal image gray average h with best match position region templateB, calculate Adaptive template-updating conditional parameter P2
=hA/hB, and if only if meets condition α ＜ P1＜ β and condition γ ＜ P2During ＜ λ, renewal current region template is To Template,
Otherwise do not update, wherein α, beta, gamma, λ is constant.Thus reducing the interference of noise, carry out real-time, stable infrared object tracking.
As shown in Fig. 2 in the inventive method, a kind of adaptive template based on Hausdorff distance coupling infrared with
Track method carries out detecting and setting up contour mould of objective contour using Sobel operator filtering template both horizontally and vertically,
Formula is as follows：
Horizontal gradient amplitude Sx=(Z1+2Z2+Z3)-(Z7+2Z8+Z9)
Vertical gradient amplitude Sy=(Z1+2Z4+Z7)-(Z3+2Z6+Z9)
Wherein, Z1,Z2,Z3,Z4,Z5,Z6,Z7,Z8,Z9It is respectively 8 neighborhood territory pixel gray values of pixel f (i, j).
As shown in figure 3, in the inventive method, a kind of adaptive template based on Hausdorff distance coupling infrared with
Track method detects and sets up contour mould using what the Sobel operator filtering template in positive and negative 45 degree of directions carried out objective contour, public
Formula is as follows：
Positive 45 degree of direction gradient amplitudes S45°=(Z2+2Z3+Z6)-(Z4+2Z7+Z8)
Minus 45 degree of direction gradient amplitudes S-45°=(Z2+2Z1+Z4)-(Z6+2Z8+Z9)
In a word, in the present invention：The infrared track method of the adaptive template based on Hausdorff distance coupling employs and changes
The Sobel edge edge detection template entering, enhances the ability of infrared target rim detection；Employ improved Hausdorff distance
Matching process, greatly improves the operating rate of algorithm；Adaptive template-updating strategy makes the robustness of target following increase
By force, improve the precision of tracking.
Claims (8)
1. a kind of infrared track method is it is characterised in that include：
Step 1：Input infrared video, selected target region in the initial two field picture of described infrared video, and to described initial
Two field picture carries out binary conversion treatment, obtains the edge contour information of target and sets up To Template；
Step 2：Read next two field picture, binary conversion treatment is carried out to next two field picture described, obtains the edge in region to be detected
Profile information；
Step 3：Described To Template and described region to be detected are mated, finds out the position of best match；
Step 4：Whether meet update condition using adaptive template conditional judgment To Template, if meeting update condition, more
Fresh target template simultaneously reads next two field picture execution step 1 to step 4, does not otherwise update To Template and reads next two field picture
Execution step 1 is to step 4, until all images in described infrared video are all disposed.
2. the method for claim 1 is it is characterised in that described step 1 includes：Using improved Sobel operator to institute
State initial two field picture and carry out binary conversion treatment, the edge pixel gray value of described target area is set to 1, rest of pixels value sets
For 0, obtain the edge contour information of target and set up To Template.
3. method as claimed in claim 2 is it is characterised in that described step 1 includes：
For each pixel in described initial two field picture, calculate respectively：
Sx=(Z1+2Z2+Z3)-(Z7+2Z8+Z9)
Sy=(Z1+2Z4+Z7)-(Z3+2Z6+Z9)
S45°=(Z2+2Z3+Z6)-(Z4+2Z7+Z8)
S-45°=(Z2+2Z1+Z4)-(Z6+2Z8+Z9)
Wherein, Z1,Z2,Z3,Z4,Z5,Z6,Z7,Z8,Z9It is respectively 8 neighborhood territory pixel gray values of current pixel point；
S when current pixel pointx,Sy,S45°,S-45°In any one be more than default threshold value when, current pixel point be edge picture
Element.
4. the method for claim 1 is it is characterised in that described step 2 includes：Using improved Sobel operator to institute
State next two field picture and carry out binary conversion treatment, the edge pixel gray value in next two field picture described is set to 1, rest of pixels
Value is set to 0, obtains the edge contour information in region to be detected.
5. method as claimed in claim 4 is it is characterised in that described step 2 includes：
For each pixel in next two field picture described, calculate respectively：
Sx=(Z1+2Z2+Z3)-(Z7+2Z8+Z9)
Sy=(Z1+2Z4+Z7)-(Z3+2Z6+Z9)
S45°=(Z2+2Z3+Z6)-(Z4+2Z7+Z8)
S-45°=(Z2+2Z1+Z4)-(Z6+2Z8+Z9)
Wherein, Z1,Z2,Z3,Z4,Z5,Z6,Z7,Z8,Z9It is respectively 8 neighborhood territory pixel gray values of current pixel point；
S when current pixel pointx,Sy,S45°,S-45°In any one be more than default threshold value when, current pixel point be edge picture
Element.
6. the method for claim 1 it is characterised in that：In step 3, to described To Template and described area to be detected
The two-way Hausdorff distance coupling that domain improves.
7. method as claimed in claim 6 is it is characterised in that described step 3 includes：
To the marginal information set A={ a in To Template1,a2,...,amAnd region to be measured marginal information set B={ b1,
b2,...,bnCarrying out two-way Hausdorff distance coupling, formula is as follows：
H (A, B)=max (h (A, B), h (B, A))
Wherein, a1,a2,...,amFor the coordinate information of To Template edge aggregation, b1,b2,...,bmFor candidate template to be measured
Edge coordinate,
H (A, B)=max (a ∈ A) min (b ∈ B) | | a-b | |
H (B, A)=max (b ∈ B) min (a ∈ A) | | b-a | |
In formula (6), h (A, B) represents the point a from set AiDistance set to the point of the set B nearest apart from this point is ranked up,
Take maximum therein, in formula (7), h (B, A) represents the point b from set BiDistance to the point putting nearest set A apart from this
Set is ranked up, and takes maximum therein, | | | | it is apart from normal form；
When the Hausdorff improving is apart from matching primitives, to current point ai8 contiguous range in look for another set coordinate
In whether have the point of this scope, if aiThe corresponding set B of four neighborhood position up and down in have at least one point corresponding, then count
Calculate thisiPoint is 1 to set B point Hausdorff distance, otherwise continually looks for aiHave accordingly at least in the corresponding set B of eight neighborhood point
One point, then calculate thisiPoint is 2 to set B point Hausdorff distance, and otherwise distance is just designated as 10.
8. the method for claim 1 is it is characterised in that step 4 includes：
By calculating the edge point set sum S in To TemplateAAnd the edge point set of best match position region template is total
Number SB, calculate Adaptive template-updating conditional parameter
P1=SA/SB, and by the gray average h of To Template region original imageAWith best match position region template
Original image gray average hB, calculate Adaptive template-updating conditional parameter P2=hA/hB, and if only if meets condition α ＜ P1＜ β
With condition γ ＜ P2During ＜ λ, renewal current region template is To Template, does not otherwise update, wherein α, beta, gamma, and λ is constant.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610803370.8A CN106447685B (en) | 2016-09-06 | 2016-09-06 | A kind of infrared track method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610803370.8A CN106447685B (en) | 2016-09-06 | 2016-09-06 | A kind of infrared track method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106447685A true CN106447685A (en) | 2017-02-22 |
| CN106447685B CN106447685B (en) | 2019-04-02 |
Family
ID=58163947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610803370.8A Expired - Fee Related CN106447685B (en) | 2016-09-06 | 2016-09-06 | A kind of infrared track method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106447685B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741305A (en) * | 2018-12-26 | 2019-05-10 | 中国科学院宁波工业技术研究院慈溪生物医学工程研究所 | A kind of detection method of capsule endoscope imaging damaged image |
| CN114566056A (en) * | 2022-02-28 | 2022-05-31 | 招商局重庆公路工程检测中心有限公司 | Highway tunnel driving safety risk identification, prevention and control method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1738426A (en) * | 2005-09-09 | 2006-02-22 | 南京大学 | Video motion goal division and track method |
| CN101673403A (en) * | 2009-10-10 | 2010-03-17 | 安防制造（中国）有限公司 | Target following method in complex interference scene |
| CN101739686A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Moving object tracking method and system thereof |
| CN102005054A (en) * | 2010-11-24 | 2011-04-06 | 中国电子科技集团公司第二十八研究所 | Real-time infrared image target tracking method |
| CN103020956A (en) * | 2012-11-20 | 2013-04-03 | 华中科技大学 | Image matching method for judging Hausdorff distance based on decision |
| CN104036524A (en) * | 2014-06-18 | 2014-09-10 | 哈尔滨工程大学 | Fast target tracking method with improved SIFT algorithm |
| CN105184822A (en) * | 2015-09-29 | 2015-12-23 | 中国兵器工业计算机应用技术研究所 | Target tracking template updating method |
-
2016
- 2016-09-06 CN CN201610803370.8A patent/CN106447685B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1738426A (en) * | 2005-09-09 | 2006-02-22 | 南京大学 | Video motion goal division and track method |
| CN101739686A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Moving object tracking method and system thereof |
| CN101673403A (en) * | 2009-10-10 | 2010-03-17 | 安防制造（中国）有限公司 | Target following method in complex interference scene |
| CN102005054A (en) * | 2010-11-24 | 2011-04-06 | 中国电子科技集团公司第二十八研究所 | Real-time infrared image target tracking method |
| CN103020956A (en) * | 2012-11-20 | 2013-04-03 | 华中科技大学 | Image matching method for judging Hausdorff distance based on decision |
| CN104036524A (en) * | 2014-06-18 | 2014-09-10 | 哈尔滨工程大学 | Fast target tracking method with improved SIFT algorithm |
| CN105184822A (en) * | 2015-09-29 | 2015-12-23 | 中国兵器工业计算机应用技术研究所 | Target tracking template updating method |
Non-Patent Citations (4)
| Title |
|---|
| 尚俊: "基于Hausdorff距离的目标跟踪算法与实现", 《湖北第二师范学院学报》 * |
| 张凯 等: "红外导引头边缘模板匹配跟踪算法研究", 《西北工业大学学报》 * |
| 李英 等: "结合SURF与聚类分析方法实现运动目标的快速跟踪", 《液晶与显示》 * |
| 黄李: "基于机器视觉的印刷线路板缺陷检测技术研究", 《中国优秀硕士论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741305A (en) * | 2018-12-26 | 2019-05-10 | 中国科学院宁波工业技术研究院慈溪生物医学工程研究所 | A kind of detection method of capsule endoscope imaging damaged image |
| CN114566056A (en) * | 2022-02-28 | 2022-05-31 | 招商局重庆公路工程检测中心有限公司 | Highway tunnel driving safety risk identification, prevention and control method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106447685B (en) | 2019-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110517288B (en) | Real-time target detection tracking method based on panoramic multi-path 4k video images | |
| CN105469113B (en) | A kind of skeleton point tracking method and system in two-dimensional video stream | |
| CN103116896B (en) | Automatic detection tracking method based on visual saliency model | |
| CN104217428B (en) | A kind of fusion feature matching and the video monitoring multi-object tracking method of data correlation | |
| CN110688987A (en) | Pedestrian position detection and tracking method and system | |
| CN109086724B (en) | Accelerated human face detection method and storage medium | |
| CN104298968B (en) | A kind of method for tracking target under complex scene based on super-pixel | |
| CN105654508B (en) | Monitor video method for tracking moving target and system based on adaptive background segmentation | |
| CN108388885A (en) | The identification in real time of more people's features towards large-scale live scene and automatic screenshot method | |
| CN103955949B (en) | Moving target detecting method based on Mean-shift algorithm | |
| CN110782477A (en) | Moving target rapid detection method based on sequence image and computer vision system | |
| CN103268480A (en) | System and method for visual tracking | |
| CN110070557A (en) | A kind of target identification and localization method based on edge feature detection | |
| CN103218605A (en) | Quick eye locating method based on integral projection and edge detection | |
| CN103067662A (en) | Self-adapting sightline tracking system | |
| CN104036523A (en) | Improved mean shift target tracking method based on surf features | |
| CN109685827B (en) | Target detection and tracking method based on DSP | |
| CN108921880A (en) | A kind of vision multi-object tracking method based on multiple single trackers | |
| CN113608663B (en) | Fingertip tracking method based on deep learning and K-curvature method | |
| CN106067031A (en) | Cooperate with the degree of depth learning network Machine Vision Recognition system based on artificial mechanism for correcting errors | |
| CN108805902A (en) | A kind of space-time contextual target tracking of adaptive scale | |
| Shi et al. | Image edge detection based on the Canny edge and the ant colony optimization algorithm | |
| CN103578121B (en) | Method for testing motion based on shared Gauss model under disturbed motion environment | |
| CN109658441B (en) | Foreground detection method and device based on depth information | |
| CN106447685A (en) | Infrared tracking method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190402 |