CN108764338A - A kind of pedestrian tracking algorithm applied to video analysis - Google Patents
A kind of pedestrian tracking algorithm applied to video analysis Download PDFInfo
- Publication number
- CN108764338A CN108764338A CN201810527019.XA CN201810527019A CN108764338A CN 108764338 A CN108764338 A CN 108764338A CN 201810527019 A CN201810527019 A CN 201810527019A CN 108764338 A CN108764338 A CN 108764338A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- frame
- feature
- denoted
- logic
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of pedestrian tracking algorithms applied to video analysis, including：The pedestrian in video scene is detected by background subtraction method；The movement position that pedestrian's subsequent time is inferred by optical flow algorithm, as whether be same person measurement, this feature is denoted as A；Compare the similitude of pedestrian's rectangle frame size, this feature is denoted as B；The color histogram for extracting pedestrian in each rectangle frame, compares the similitude of present frame detection block and next frame detection block color histogram, this feature is denoted as C；Three of the above feature is combined, feature F is denoted as；With this special grader of feature F training logic, make this special grader of logic possess judge whether be same person ability；The association of pedestrian detection frame between this special grader is carried out per frame with trained logic.The present invention is by the combination of series of features, and using logic, this special grader completes the association of data between rectangle frame, to realize the pedestrian tracking of monitor video.
Description
Technical field
The present invention relates to the image procossing of computer vision and area of pattern recognition, more particularly to a kind of applied to video point
The pedestrian tracking algorithm of analysis.
Background technology
In recent years, image procossing based on computer vision and pattern-recognition are a popular classes of field of machine vision
Topic, is widely used general.Wherein, pedestrian detection and tracking are an important branch.Pedestrian detection and tracking refer to from video
The location information of pedestrian is detected in sequence, and assigns identical label in the position of different frame to the same pedestrian, determines it
The process of movement locus.Currently, many effectively track algorithms are constantly proposed, but due to tracking target by it is a variety of it is extraneous because
The interference of element, such as illumination variation, scale telescopic variation, object block, and be easy to cause target loss, tracking failure.It is existing
In pedestrian tracking algorithm, various features are combined by most common feature such as color characteristic, HOG features, edge feature etc.,
The direction that a kind of robustness is stronger, the higher detection algorithm of precision is a worth research is provided.
Invention content
In order to overcome deficiency in the prior art, the present invention to provide a kind of pedestrian tracking algorithm applied to video analysis,
Three kinds of features are combined, by training this special grader of logic to complete the association of data between pedestrian's rectangle frame, to real
The track algorithm of existing monitor video analysis.
In order to reach foregoing invention purpose, technical solution is as follows used by solving its technical problem：
A kind of pedestrian tracking algorithm applied to video analysis, including with lower part：
The pedestrian in video scene is detected by background subtraction method, obtains the initial square of corresponding one or more pedestrian targets
Shape region；
The fortune of pedestrian's subsequent time is inferred by the variation of pixel motion velocity information in image sequence with optical flow algorithm
Whether dynamic position, and making comparisons with next frame pedestrian detection position uses the similitude of the two as be the same pedestrian being to measure,
This feature is denoted as A；
The similitude for comparing pedestrian's rectangle frame size, as whether be the same pedestrian measurement, this feature is denoted as B；
The color histogram for extracting pedestrian in each rectangle frame, compares present frame detection block and next frame detection block color is straight
The similitude of square figure, as whether be the same pedestrian measurement, this feature is denoted as C；
Three of the above feature is combined, a new feature is obtained, is denoted as feature F；
The input of this special grader using feature F as logic, this special grader of training logic make this special grader of logic gather around
Judge whether be same person ability；
The association of pedestrian detection frame, realizes pedestrian tracking between with trained logic, this special grader is carried out per frame.
Further, the background subtraction method is mixed Gaussian background modeling, calculates model Gaussian function at pixel (x, y)
Several mean value u and variance d calculates probability P of the new frame image midpoint (x, y) in probabilistic model, passes through probability P and threshold value T
Comparison, judge foreground point and background dot.
Further, the optical flow algorithm is Farneback global optical flow algorithms, passes through the optical flow computation light stream of two interframe
, the speed of pedestrian movement is obtained, and the movement position of pedestrian's subsequent time is inferred with this, if its rectangle frame central point is (x1,
y1), with next frame pedestrian detection rectangle frame central point (x2, y2) make comparisons, calculate two frame rectangle frame central point distancesThis feature is denoted as A.
Further, the similitude of relatively pedestrian's rectangle frame size is the ratio by two rectangle frame intersection unions
It is measured, calculates the intersection area I and union area U of present frame and pedestrian's rectangle frame in next frame image, pass through I's and U
Ratio indicates the similitude of pedestrian's rectangle frame size, as whether be same pedestrian measurement, i.e.,This feature is remembered
For B.
Further, in each rectangle frame of the extraction pedestrian color histogram, color space is divided into 24 BIN, R,
G, each 8 BIN in tri- channels B calculate separately gray value and fall the pixel quantity in each BIN, obtain vector { x1, x2, x3... x24,
Next frame color histogram vector is denoted as { y1, y2, y3... y24, it calculates
This feature is denoted as C.
Further, the assemblage characteristic F is the fusion of features described above A, feature B, feature C, F=[A, B, C, A*A, B*
B, C*C, A*B, A*C, B*C, A*B*C].
Further, using assemblage characteristic F as logic, grader is trained in the input of this special grader, and the grader is made to possess
Judge whether be same person ability, complete rectangle frame between data correlation, wherein be that same person is denoted as 1, be not same
One people is denoted as 0, and same person is positive example, and different artificially negative examples, positive and negative sample proportion is 1: 1.
Further, the association of pedestrian detection frame between this special grader is carried out per frame with trained logic, to video
Pedestrian's rectangle frame label sequence that first frame image detection goes out, by logic, this special grader judges to detect in next frame image
Pedestrian whether be same person in first frame, be then to mark identical with former frame serial number, be not to assign new label,
The rest may be inferred, realizes pedestrian tracking.
The present invention due to using the technology described above, is allowed to compared with prior art, have the following advantages that and actively imitate
Fruit：
A kind of pedestrian tracking algorithm applied to video analysis of the invention, by carrying out the various features of pedestrian in video
Combination, and using logic, this special grader completes the association of data between rectangle frame, and high-precision is carried out to target to reach
With the tracking effect of high robust, it is suitable for pedestrian detection and the tracking of fixed seat in the plane monitor camera.
Description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, required use in being described below to embodiment
Attached drawing be briefly described.It is clear that drawings in the following description are only some embodiments of the invention, for ability
For field technique personnel, without creative efforts, other drawings may also be obtained based on these drawings.It is attached
In figure：
Fig. 1 is pedestrian detection and track algorithm overview flow chart in the present invention, wherein：
Feature A：The movement position for pedestrian's subsequent time that present frame is inferred to is similar to next frame pedestrian detection position
Property；
Feature B：The similitude of present frame and next frame pedestrian's rectangle frame size；
Feature C：The similitude of present frame and color histogram in next frame rectangle frame；
Fig. 2 is that mixed Gaussian background modeling detects pedestrian's flow chart in the present invention；
Fig. 3 is specific embodiment design sketch in the present invention.
Specific implementation mode
Below with reference to the attached drawing of the present invention, technical solution in the embodiment of the present invention carries out clear, complete description
And discussion, it is clear that as described herein is only a part of example of the present invention, is not whole example, based on the present invention
In embodiment, the every other implementation that those of ordinary skill in the art are obtained without making creative work
Example, belongs to protection scope of the present invention.
The present invention provides a kind of pedestrian tracting method applied to monitor video analysis, overall procedure schematic diagram such as Fig. 1 institutes
Show, including with lower part：
The pedestrian in video scene is detected by background subtraction method, obtains the initial square of corresponding one or more pedestrian targets
Shape region；
The fortune of pedestrian's subsequent time is inferred by the variation of pixel motion velocity information in image sequence with optical flow algorithm
Whether dynamic position, and making comparisons with next frame pedestrian detection position uses the similitude of the two as be the same pedestrian being to measure,
This feature is denoted as A；
The similitude for comparing pedestrian's rectangle frame size, as whether be the same pedestrian measurement, this feature is denoted as B；
The color histogram for extracting pedestrian in each rectangle frame, compares present frame detection block and next frame detection block color is straight
The similitude of square figure, as whether be the same pedestrian measurement, this feature is denoted as C；
Three of the above feature is combined, a new feature is obtained, is denoted as feature F；
The input of this special grader using feature F as logic, this special grader of training logic make this special grader of logic gather around
Judge whether be same person ability；
The association of pedestrian detection frame, realizes pedestrian tracking between with trained logic, this special grader is carried out per frame.
The pedestrian in background subtraction method detection video scene is used first, and the method is specially that mixed Gaussian background is built
Mould, as shown in Figure 2.The mean value u and variance d of model Gaussian function at pixel (x, y) are calculated, new frame image midpoint is calculated
The probability P of (x, y) in probabilistic model judges foreground point and background dot by the comparison of probability P and threshold value T.According to difference
Application scenarios, adjust threshold value T, take T=0.1 in specific embodiment here, finally obtain corresponding one or more pedestrian's mesh
Target initial rectangular region.
Further, the optical flow algorithm is Farneback global optical flow algorithms, passes through the optical flow computation light stream of two interframe
, the speed of pedestrian movement is obtained, and the movement position of pedestrian's subsequent time is inferred with this, if its rectangle frame central point is (x1,
y1), with next frame pedestrian detection rectangle frame central point (x2, y2) make comparisons, calculate two frame rectangle frame central point distancesThis feature is denoted as A；.
Secondly, the similitude of relatively pedestrian's rectangle frame size is carried out by the ratio of two rectangle frame intersection unions
Measurement calculates the intersection area I and union area U of present frame and pedestrian's rectangle frame in next frame image, passes through the ratio of I and U
Indicate the similitude of pedestrian's rectangle frame size, as whether be same pedestrian measurement, i.e.,This feature is denoted as B.
Further more, in each rectangle frame of the extraction pedestrian color histogram, color space is divided into 24 BIN, R,
G, each 8 BIN in tri- channels B calculate separately gray value and fall the pixel quantity in each BIN, obtain vector { x1, x2, x3... x24,
Next frame color histogram vector is denoted as { y1, y2, y3... y24, it calculates
This feature is denoted as C.
Then, feature A, feature B, feature C are combined, obtain a new feature F, i.e. F=[A, B, C, A*A, B*
B, C*C, A*B, A*C, B*C, A*B*C].Both single feature A, B, C had been contained in feature F, while also including three kinds of features
Nonlinear combination, such as A*A, A*B.
Further, using assemblage characteristic F as logic, grader is trained in the input of this special grader, so that the grader is possessed and is sentenced
It is disconnected whether be same person ability, complete the data correlation between rectangle frame, wherein be that same person is denoted as 1, be not same
Individual is denoted as 0, and same person is positive example, different artificially negative examples, the accuracy of this special classifier training result in order to ensure logic,
It is 1: 1 to choose positive and negative sample proportion.
Finally, the association of pedestrian detection frame between this special grader is carried out per frame with trained logic, to video first
Pedestrian's rectangle frame label sequence that frame image detection goes out, by logic, this special grader judges the row detected in next frame image
People whether be same person in first frame, be then to mark identical with former frame serial number, be not to assign new label, according to this
Analogize, realizes pedestrian tracking.It is as shown in Figure 3 that effect is embodied in algorithm.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Any one skilled in the art in the technical scope disclosed by the present invention, the change or replacement that can be readily occurred in,
It should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims
Subject to.
Claims (8)
1. a kind of pedestrian tracking algorithm applied to video analysis, which is characterized in that including with lower part：
The pedestrian in video scene is detected by background subtraction method, obtains the initial rectangular area of corresponding one or more pedestrian targets
Domain；
The motion bit of pedestrian's subsequent time is inferred by the variation of pixel motion velocity information in image sequence with optical flow algorithm
It sets, and makes comparisons with next frame pedestrian detection position, use the similitude of the two as whether be the same pedestrian being measurement, this is special
Sign is denoted as A；
The similitude for comparing pedestrian's rectangle frame size, as whether be the same pedestrian measurement, this feature is denoted as B；
The color histogram for extracting pedestrian in each rectangle frame compares present frame detection block and next frame detection block color histogram
Similitude, as whether be the same pedestrian measurement, this feature is denoted as C；
Three of the above feature is combined, a new feature is obtained, is denoted as feature F；
The input of this special grader using feature F as logic, this special grader of training logic make this special grader of logic possess and sentence
It is disconnected whether be same person ability；
The association of pedestrian detection frame, realizes pedestrian tracking between with trained logic, this special grader is carried out per frame.
2. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1, which is characterized in that the background
Relief method is mixed Gaussian background modeling, calculates the mean value u and variance d of model Gaussian function at pixel (x, y), calculates new one
Probability P of the frame image midpoint (x, y) in probabilistic model judges foreground point and background by the comparison of probability P and threshold value T
Point.
3. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1, which is characterized in that the light stream
Algorithm obtains the speed of pedestrian movement for Farneback global optical flow algorithms by the optical flow computation optical flow field of two interframe, and
The movement position of pedestrian's subsequent time is inferred with this, if its rectangle frame central point is (x1, y1), with next frame pedestrian detection rectangle
Frame central point (x2, y2) make comparisons, calculate two frame rectangle frame central point distancesThis feature is remembered
For A.
4. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1, which is characterized in that the comparison
The similitude of pedestrian's rectangle frame size is measured by the ratio of two rectangle frame intersection unions, calculate present frame with it is next
The intersection area I of pedestrian's rectangle frame and union area U, pedestrian's rectangle frame size is indicated by the ratio of I and U in frame image
Similitude, as whether be the same pedestrian measurement, i.e.,This feature is denoted as B.
5. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1, which is characterized in that the extraction
The color histogram of pedestrian, 24 BIN, each 8 BIN in tri- channels R, G, B are divided by color space in each rectangle frame, point
Not Ji Suan gray value fall the pixel quantity in each BIN, obtain vector { x1, x2, x3... x24, next frame color histogram to
Amount is denoted as { y1, y2, y3... y24, it calculatesThis feature is denoted as C.
6. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1, which is characterized in that the combination
Feature F is the fusion of features described above A, feature B, feature C, F=[A, B, C, A*A, B*B, C*C, A*B, A*C, B*C, A*B*C].
7. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1 or 6, which is characterized in that with group
Closing feature F, grader is trained in the input of this special grader as logic, so that the grader is possessed and is judged whether it is same person
Ability completes the data correlation between rectangle frame, wherein and it is that same person is denoted as 1, is not that same person is denoted as 0, same person
For positive example, different artificially negative examples, positive and negative sample proportion is 1: 1.
8. a kind of pedestrian tracking algorithm applied to video analysis according to claim 1, which is characterized in that with training
Logic this special grader carry out per frame between pedestrian detection frame association, pedestrian's rectangle that video first frame image detection is gone out
Collimation mark number sorts, by logic this special grader judge the pedestrian detected in next frame image whether be same in first frame
Individual is then to mark serial number identical with former frame, is not to assign new label, and so on, realize pedestrian tracking.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810527019.XA CN108764338B (en) | 2018-05-28 | 2018-05-28 | Pedestrian tracking method applied to video analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810527019.XA CN108764338B (en) | 2018-05-28 | 2018-05-28 | Pedestrian tracking method applied to video analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108764338A true CN108764338A (en) | 2018-11-06 |
| CN108764338B CN108764338B (en) | 2021-05-04 |
Family
ID=64003134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810527019.XA Active CN108764338B (en) | 2018-05-28 | 2018-05-28 | Pedestrian tracking method applied to video analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108764338B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991280A (en) * | 2019-11-20 | 2020-04-10 | 北京影谱科技股份有限公司 | Video tracking method and device based on template matching and SURF |
| WO2020103462A1 (en) * | 2018-11-21 | 2020-05-28 | 百度在线网络技术（北京）有限公司 | Video search method and apparatus, computer device, and storage medium |
| CN111882582A (en) * | 2020-07-24 | 2020-11-03 | 广州云从博衍智能科技有限公司 | Image tracking correlation method, system, device and medium |
| CN113379985A (en) * | 2020-02-25 | 2021-09-10 | 北京君正集成电路股份有限公司 | Nursing electronic fence alarm device |
| CN113379984A (en) * | 2020-02-25 | 2021-09-10 | 北京君正集成电路股份有限公司 | Electronic nursing fence system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521616A (en) * | 2011-12-28 | 2012-06-27 | 江苏大学 | Pedestrian detection method on basis of sparse representation |
| CN103632170A (en) * | 2012-08-20 | 2014-03-12 | 深圳市汉华安道科技有限责任公司 | Pedestrian detection method and device based on characteristic combination |
| CN104050460A (en) * | 2014-06-30 | 2014-09-17 | 南京理工大学 | Pedestrian detection method with multi-feature fusion |
| CN106778570A (en) * | 2016-12-05 | 2017-05-31 | 清华大学深圳研究生院 | A kind of pedestrian's real-time detection and tracking |
| CN106778478A (en) * | 2016-11-21 | 2017-05-31 | 中国科学院信息工程研究所 | A kind of real-time pedestrian detection with caching mechanism and tracking based on composite character |
| US20170300780A1 (en) * | 2014-09-24 | 2017-10-19 | Denso Corporation | Object detection apparatus |
| CN107492116A (en) * | 2017-09-01 | 2017-12-19 | 深圳市唯特视科技有限公司 | A kind of method that face tracking is carried out based on more display models |
-
2018
- 2018-05-28 CN CN201810527019.XA patent/CN108764338B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521616A (en) * | 2011-12-28 | 2012-06-27 | 江苏大学 | Pedestrian detection method on basis of sparse representation |
| CN103632170A (en) * | 2012-08-20 | 2014-03-12 | 深圳市汉华安道科技有限责任公司 | Pedestrian detection method and device based on characteristic combination |
| CN104050460A (en) * | 2014-06-30 | 2014-09-17 | 南京理工大学 | Pedestrian detection method with multi-feature fusion |
| US20170300780A1 (en) * | 2014-09-24 | 2017-10-19 | Denso Corporation | Object detection apparatus |
| CN106778478A (en) * | 2016-11-21 | 2017-05-31 | 中国科学院信息工程研究所 | A kind of real-time pedestrian detection with caching mechanism and tracking based on composite character |
| CN106778570A (en) * | 2016-12-05 | 2017-05-31 | 清华大学深圳研究生院 | A kind of pedestrian's real-time detection and tracking |
| CN107492116A (en) * | 2017-09-01 | 2017-12-19 | 深圳市唯特视科技有限公司 | A kind of method that face tracking is carried out based on more display models |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020103462A1 (en) * | 2018-11-21 | 2020-05-28 | 百度在线网络技术（北京）有限公司 | Video search method and apparatus, computer device, and storage medium |
| US11348254B2 (en) | 2018-11-21 | 2022-05-31 | Baidu Online Network Technology (Beijing) Co., Ltd. | Visual search method, computer device, and storage medium |
| CN110991280A (en) * | 2019-11-20 | 2020-04-10 | 北京影谱科技股份有限公司 | Video tracking method and device based on template matching and SURF |
| CN113379985A (en) * | 2020-02-25 | 2021-09-10 | 北京君正集成电路股份有限公司 | Nursing electronic fence alarm device |
| CN113379984A (en) * | 2020-02-25 | 2021-09-10 | 北京君正集成电路股份有限公司 | Electronic nursing fence system |
| CN113379984B (en) * | 2020-02-25 | 2022-09-23 | 北京君正集成电路股份有限公司 | Electronic nursing fence system |
| CN113379985B (en) * | 2020-02-25 | 2022-09-27 | 北京君正集成电路股份有限公司 | Nursing electronic fence alarm device |
| CN111882582A (en) * | 2020-07-24 | 2020-11-03 | 广州云从博衍智能科技有限公司 | Image tracking correlation method, system, device and medium |
| CN111882582B (en) * | 2020-07-24 | 2021-10-08 | 广州云从博衍智能科技有限公司 | Image tracking correlation method, system, device and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108764338B (en) | 2021-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108764338A (en) | A kind of pedestrian tracking algorithm applied to video analysis | |
| CN107545582B (en) | Video multi-target tracking method and device based on fuzzy logic | |
| CN104183127B (en) | Traffic surveillance video detection method and device | |
| CN101739551B (en) | Method and system for identifying moving objects | |
| CN109977782B (en) | Cross-store operation behavior detection method based on target position information reasoning | |
| CN108171112A (en) | Vehicle identification and tracking based on convolutional neural networks | |
| CN100565557C (en) | System for tracking infrared human body target based on corpuscle dynamic sampling model | |
| CN104978567B (en) | Vehicle checking method based on scene classification | |
| CN106127807A (en) | A kind of real-time video multiclass multi-object tracking method | |
| CN103886325B (en) | Cyclic matrix video tracking method with partition | |
| CN102867311A (en) | Target tracking method and target tracking device | |
| CN103164858A (en) | Adhered crowd segmenting and tracking methods based on superpixel and graph model | |
| CN104200494A (en) | Real-time visual target tracking method based on light streams | |
| CN110298297A (en) | Flame identification method and device | |
| CN108074234A (en) | A kind of large space flame detecting method based on target following and multiple features fusion | |
| CN108776974A (en) | A kind of real-time modeling method method suitable for public transport scene | |
| CN106709938B (en) | Based on the multi-target tracking method for improving TLD | |
| CN106204594A (en) | A kind of direction detection method of dispersivity moving object based on video image | |
| CN112241969A (en) | Target detection tracking method and device based on traffic monitoring video and storage medium | |
| CN104599291B (en) | Infrared motion target detection method based on structural similarity and significance analysis | |
| CN106570490A (en) | Pedestrian real-time tracking method based on fast clustering | |
| CN104915642A (en) | Method and apparatus for measurement of distance to vehicle ahead | |
| CN110197121A (en) | Moving target detecting method, moving object detection module and monitoring system based on DirectShow | |
| CN106056078A (en) | Crowd density estimation method based on multi-feature regression ensemble learning | |
| CN104778676A (en) | Depth ranging-based moving target detection method and system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |