CN103810499A - Application for detecting and tracking infrared weak object under complicated background - Google Patents
Application for detecting and tracking infrared weak object under complicated background Download PDFInfo
- Publication number
- CN103810499A CN103810499A CN201410062826.0A CN201410062826A CN103810499A CN 103810499 A CN103810499 A CN 103810499A CN 201410062826 A CN201410062826 A CN 201410062826A CN 103810499 A CN103810499 A CN 103810499A
- Authority
- CN
- China
- Prior art keywords
- target
- weak
- entropy
- tracking
- weak object
- 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
Images
Landscapes
- Image Analysis (AREA)
- Aiming, Guidance, Guns With A Light Source, Armor, Camouflage, And Targets (AREA)
Abstract
The invention discloses an application for detecting and tracking an infrared weak object under a complicated background. The application is characterized by comprising the following steps of: 1. suppressing clutters and keeping the topological structure of an image, and constructing a bionic vision weighted entropy model with an adjacent airspace and a preferred direction to realize conversion for the image from a grey mode to an entropy model; 2. analysing the movement state of the weak object with burst and stationary characteristics, and constructing a self-adaptive entropy flow target movement estimation model meeting the maneuvering features of the weak object by virtue of the nonlinear diffusion smoothing and self-adaptive local restriction criterion of an entropy flow to realize the approximation of an estimation speed to the real movement state of the weak object; 3. searching a weak object tracking method adopting generic multi-feature and measurement, and constructing a multi-feature fused sequential filter model to realize accurate, robust and real-time identification for the weak object. The invention discloses a self-adaptive entropy flow detection and tracking algorithm for the infrared weak object, and enriches a detection and tracking technology for the weak object.
Description
Technical field
The present invention relates to the analysis of Infrared images pre-processing technology, weak target state and the weak target tracking domain of many Fusion Features and tolerance, be specifically related to the application of infrared weak object detecting and tracking under complex background.
Background technology
Background technology of the present invention relates to three committed steps and method: the method for tracking target that keeps Preprocessing Technique, analysis burst or the target state of smooth performance, similar many Fusion Features and the tolerance of topological structure.
Keep the Infrared images pre-processing technology of topological structure:
The existence of clutter and noise, they have improved the difficulty of weak target detection.Adopt spatial filtering method Background suppression and noise, the topological structure of its image can change.The inhibition method of research clutter and noise, the signal to noise ratio of target but also keep the topological structure of image a little less than not only having improved.
Analyze the target state of burst or smooth performance:
The motion of weak target has burst or smooth performance, is used for characterizing uncertainty of objective and maneuverability.According to image entropy pattern caused apparent be exactly to portray target travel, the feature of or acute variation mild for area pixel gray scale, the level and smooth target travel estimation model with constraint criterion of research, realization approaching target state.
The method for tracking target of similar many Fusion Features and tolerance:
The existence of pseudo-target, and the phenomenon of target occlusion, intersection, separation, appearance, disappearance, their improve the difficulty of target following.Target generally has the generic character of identical or convergence, and the method for tracking target of the research many Fusion Features of generic and tolerance can be rejected pseudo-target and be realized the tracking of weak target.
Summary of the invention
The object of the present invention is to provide the application of infrared weak object detecting and tracking under complex background, keep the preconditioning technique of infrared weak target sequence image topological structure, clutter reduction and raising target signal to noise ratio; Build the estimation model of the self-adaptation entropy flow target travel that meets weak target maneuver feature according to entropy model, approach the motion state of target; And the weak target of the method recognition and tracking that adopts the many Fusion Features of generic and tolerance, thereby realize infrared weak object detecting and tracking.
By setting up the goal in research of the target travel estimation model of weighted entropy model, self-adaptation entropy flow and the Sequential filter model of many Fusion Features, the weak target of identification and weak target trajectory.The key scientific problems that the present invention need to solve is as follows:
(1) build the spatial domain mask that bionical thing vision significance is measured.The feature of mask outstanding spatial domain vicinity and orientation preferentially, the Weighted information entropy matrix description gradation of image information matrix of employing spatial domain mask, realizes the conversion of image entropy pattern.
(2) the adaptive variation model of structure nonlinear diffusion.Adopt the level and smooth degree of nonlinear diffusion factor control, adapt to adjust the data item of entropy flow Variation Model and the scale factor of level and smooth, realize and treat that estimating speed approaches the motion state of target.
(3) establishing target correlation function.In set spatial domain and time domain, the homogeny of the continuity of target travel and consistance and target generic, determines that it will appear in adjacent domain with very big probability.Correlation function comprises spatial domain and time domain, motion and non-motion generic character, adopts many characteristic distances evaluation function realize target to detect.
The technical solution adopted for the present invention to solve the technical problems is:
Method of the present invention comprises following key step:
1, suppress the topological structure of noise, clutter and maintenance image, build the bionical visual weight entropy model of spatial domain vicinity and orientation preferentially, realize image and be transformed to entropy pattern from grayscale mode.
2, portray exactly weak target travel by the caused apparent motion of image entropy pattern, analyze the target state of burst or smooth performance.According to nonlinear smoothing and the local restriction criterion of entropy flow, adopt the level and smooth degree of nonlinear diffusion factor control, and entropy flow constraint is treated estimating speed with level and smooth set of constraints contract bundle, obtain fine and close entropy flow field, structure meets the self-adaptation entropy flow target travel estimation model of target maneuver feature, convergence target state.
3, the method for tracking target of the research many Fusion Features of generic and tolerance, builds the Sequential filter model of many Fusion Features, realize weak target accurately, robust and tracking in real time.
Its infrared weak object detecting and tracking process flow diagram as shown in Figure 1.
Advantage of the present invention is:
(1) the weighted entropy model of spatial domain vicinity with orientation preferentially proposed
Conventionally adopt spatial filtering Background suppression and noise, it is the information and the topological structure that changes image of loss target easily.For the problem of Background suppression of the present invention and noise, watch mechanism attentively according to biology, the weighting spatial domain mask of structure spatial domain vicinity and orientation preferentially.Adopt the information entropy dimensioned plan of spatial domain mask weighting as pixel grey scale, realize image and be transformed to entropy pattern by grayscale mode.It is for clutter reduction and keep the topological structure of image that a kind of new approaches are provided.
(2) the target travel estimation model of the nonlinear diffusion of proposition self-adaptation entropy flow
Weak uncertainty of objective and maneuverability and the interference of clutter to estimation, they can reduce the degree of accuracy that target travel is estimated.Carry out the research contents of estimating target motion state for the present invention, in the face of Entropy Changesization feature gently or sharply, self-adaptation is adjusted the scale factor of entropy model data item and level and smooth item, adopt the level and smooth degree of nonlinear diffusion factor control, realize entropy flow constraint and treat estimating speed with level and smooth set of constraints contract bundle, with convergence target state.The degree of accuracy that it is estimated for target travel a little less than improving provides a kind of new method.
Accompanying drawing explanation
Fig. 1 is infrared weak object detecting and tracking process flow diagram of the present invention;
Fig. 2 is weak object detecting and tracking technical scheme of the present invention;
Fig. 3 be the present invention by grayscale mode the technology path to entropy mode conversion;
Fig. 4 is the weak target travel estimation technique route in the present invention.
Embodiment
The present invention adopts infrared weak object detecting and tracking technical scheme under complex background as shown in Figure 2, and its concrete implementation step is as follows:
(1) vision is watched attentively and is embodied the position at the violent place of base sketch map Strength Changes and distribute and institutional framework for how much, and primitive normal direction and each point are from the point of discontinuity in observer's the degree of depth, the degree of depth, the point of discontinuity in surface normal direction.Neuron receptive field adopts Gauss-exponential model, by deformation district overlapping, that vary in size with one heart, the center, perimeter region of gangliocyte tradition receptive field, the district of disinthibiting are on a large scale described, wherein Gauss model is explained center and perimeter region successively, and exponential model is described the district of disinthibiting on a large scale.Gauss-index mathematical model is referring to formula (1), by two Gauss models and an exponential model stack, obtain spatial filtering mask, thereby solve the design of the spatial domain mask of the bionical vision significance tolerance of key issue, make it submit to target spatial domain vicinity and select feature with orientation preferentially, realize the object of clutter reduction and noise.
wherein
,
,
represent respectively the peak factor at central authorities, surrounding and edge,
,
represent respectively the scale coefficient of central authorities, surrounding, and
,
represent according to this gradient of vertical direction, horizontal direction.
When this model is used for processing brightness contrast edge, edge contrast can be strengthened well, regional luminance contrast and the half tone information of being experienced the filtering of Yezhong heart institute by tradition can be effectively promoted again.
(2), in information theory, image is considered the carrier of information aggregate.Shannon entropy has descriptor amount or probabilistic characteristic, adopts the quantity of information of the weighting Shannon entropy Description Image sub-block of spatial domain mask.Use successively Weighted information entropy dimensioned plan as pixel grey scale from left to right, from top to bottom, thereby build a complete entropy diagram.It will solve the conversion of image model, and its weighted entropy mathematical model is referring to formula (2):
wherein
presentation video exists
gray scale,
represent probability density function,
representation space filtering mask,
the local window size of presentation video,
presentation video size.In the time of computing information entropy,
in definition, be multiplied by the factor
, in order to guarantee Shannon entropy
in interval
monotone increasing.
Adopt technology path (1) and (2), build the weighted entropy model of spatial domain vicinity and orientation preferentially, realize image and be transformed to entropy pattern from grayscale mode.Its technology path as shown in Figure 3.
(3) direction that Entropy Changesization is strong should not add smoothness constraint, and smoothness constraint should be added in the direction perpendicular to gradient.The item of data constraint simultaneously is only set up at entropy diagram gradient larger part.In the larger some place usage data constraint of gradient, and only use Smoothing Constraint at the less some place of gradient.Definition weight function, when gradient is greater than a certain threshold value, weight function is 1, usage data constraint condition; When being less than a certain threshold value, weight function is 0, closes data constraint condition; Other situation, is used partial weighting data constraint condition.It will capture self-adaptation in key scientific problems adaptive variation model and adjust the design of data item and the scale factor of level and smooth, and its mathematical model is referring to formula (3):
wherein
represent entropy diagram picture,
represent entropy flow field,
represent weight function,
represent the gradient-norm function of entropy diagram picture,
represent data item,
represent level and smooth.
(4) for level and smooth of the estimation model of self-adaptation entropy flow, in more smooth region, entropy flow field, the conduction factor can increase automatically, make the less random fluctuation in flat region smoothed, and near the sudden change of entropy flow field, the conduction factor can be automatically less, and edge is influenced hardly so.The conduction factor, as the nonlinear function successively decreasing, adopts nonlinear diffusion factor control level and smooth degree, and its mathematical model is referring to formula (4):
wherein
represent entropy flow field,
depend on the data item in formula (3),
represent diffusion term, depend in formula (3) level and smooth with the combination of the nonlinear diffusion factor.It is by the structure of the nonlinear diffusion factor in the adaptive variation model of solution nonlinear diffusion.Adopt technology path (3) and (4), build the target travel estimation model of the nonlinear diffusion of self-adaptation entropy flow.By the optimization solution of method of steepest descent, realize entropy flow and approach weak dbjective state.Its technology path as shown in Figure 4.
(5) from the target travel estimation model of self-adaptation entropy flow, identify candidate target and motion state thereof, adopt motion and non-movable information to carry out Expressive Features collection.Meeting the adjacent domains prerequisite of spatial feature, will move and merge with non-motion feature by synthesis strategy, the Multiple feature association Distance evaluation function of establishing target.Meeting appreciation condition may be target, then identifies weak target and target following through the energy accumulation Sequential filter method of multiple image target trajectory.
Claims (4)
1. the infrared weak detection of target and the application of tracker under complex background, is characterized in that method step is as follows:
(1) keep the Preprocessing Technique of topological structure, build the contiguous bionical visual weight entropy model with orientation preferentially in spatial domain, make image be transformed to entropy pattern from grayscale mode, not only improved the signal to noise ratio of weak target but also kept the topological structure of image;
(2) burst or the easy motion state of the weak target of analysis, according to nonlinear smoothing and the local restriction criterion of entropy flow, adopt the level and smooth degree of nonlinear diffusion factor control, self-adaptation is adjusted the data item of entropy flow Variation Model and the scale factor of level and smooth, the nonlinear diffusion motion model of structure self-adaptation entropy flow, describes weak target state;
(3) study the weak method for tracking target of similar many Fusion Features and tolerance, target generally has the generic character of identical or convergence, adopt motion and the non-motion feature collection of weak target, analyze many features synthesis strategy, build the Distance evaluation function of Multiple feature association, the weak target of identification and target trajectory.
2. the application of a kind of infrared weak object detecting and tracking according to claim 1, it is characterized in that: the Infrared images pre-processing technology of described maintenance topological structure, build the bionical visual weight entropy model of spatial domain vicinity and orientation preferentially, changing image pattern, realizes the object that suppresses noise, clutter and maintenance image topology structure.
3. the application of a kind of infrared weak object detecting and tracking according to claim 1, is characterized in that: target travel is portrayed in the caused apparent motion of described image entropy pattern, analyzes the weak target state of burst or smooth performance; According to nonlinear smoothing and the local restriction criterion of entropy flow, adopt the nonlinear diffusion factor to control level and smooth degree, and the estimating speed of the weak target of self-adaptation combination constraint of entropy flow constraint and smoothness constraint, obtain fine and close entropy flow field, realize the weak target state of estimating speed approaching to reality.
4. the application of a kind of infrared weak object detecting and tracking according to claim 1, it is characterized in that: the tracking of described single features is difficult to be adapted to complicated infrared tracking applied environment, utilize similar target generally to there is the generic character of identical or convergence, build motion and the many Fusion Features of generic of non-motion and the method for tracking target of tolerance, reject pseudo-target, the weak target of identification and target trajectory.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410062826.0A CN103810499B (en) | 2014-02-25 | 2014-02-25 | Application for detecting and tracking infrared weak object under complicated background |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410062826.0A CN103810499B (en) | 2014-02-25 | 2014-02-25 | Application for detecting and tracking infrared weak object under complicated background |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103810499A true CN103810499A (en) | 2014-05-21 |
CN103810499B CN103810499B (en) | 2017-04-12 |
Family
ID=50707240
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410062826.0A Expired - Fee Related CN103810499B (en) | 2014-02-25 | 2014-02-25 | Application for detecting and tracking infrared weak object under complicated background |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103810499B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104268844A (en) * | 2014-10-17 | 2015-01-07 | 中国科学院武汉物理与数学研究所 | Small target infrared image processing method based on weighing local image entropy |
CN104992452A (en) * | 2015-06-25 | 2015-10-21 | 中国计量学院 | Flight object automatic tracking method based on thermal imaging video |
CN107280673A (en) * | 2017-06-02 | 2017-10-24 | 南京理工大学 | A kind of infrared imaging breath signal detection method based on key-frame extraction technique |
CN107657626A (en) * | 2016-07-25 | 2018-02-02 | 浙江宇视科技有限公司 | The detection method and device of a kind of moving target |
CN108364277A (en) * | 2017-12-20 | 2018-08-03 | 南昌航空大学 | A kind of infrared small target detection method of two-hand infrared image fusion |
CN110706231A (en) * | 2019-11-12 | 2020-01-17 | 安徽师范大学 | Image entropy-based three-dimensional culture human myocardial cell pulsation characteristic detection method |
CN111259942A (en) * | 2020-01-10 | 2020-06-09 | 西北工业大学 | Method for detecting weak target in water |
CN111311191A (en) * | 2020-02-25 | 2020-06-19 | 武汉轻工大学 | Method and device for acquiring raw material quality range based on wine product quality range |
CN115249254A (en) * | 2022-09-21 | 2022-10-28 | 江西财经大学 | Target tracking method and system based on AR technology |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030202701A1 (en) * | 2002-03-29 | 2003-10-30 | Jonathon Schuler | Method and apparatus for tie-point registration of disparate imaging sensors by matching optical flow |
CN102289681A (en) * | 2011-08-05 | 2011-12-21 | 上海邮政科学研究院 | Method for matching envelope images |
CN102722890A (en) * | 2012-06-07 | 2012-10-10 | 内蒙古科技大学 | Non-rigid heart image grading and registering method based on optical flow field model |
-
2014
- 2014-02-25 CN CN201410062826.0A patent/CN103810499B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030202701A1 (en) * | 2002-03-29 | 2003-10-30 | Jonathon Schuler | Method and apparatus for tie-point registration of disparate imaging sensors by matching optical flow |
CN102289681A (en) * | 2011-08-05 | 2011-12-21 | 上海邮政科学研究院 | Method for matching envelope images |
CN102722890A (en) * | 2012-06-07 | 2012-10-10 | 内蒙古科技大学 | Non-rigid heart image grading and registering method based on optical flow field model |
Non-Patent Citations (1)
Title |
---|
王忠华等: "《联合熵流与边缘算子的红外运动目标检测》", 《华中科技大学学报(自然科学版)》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104268844B (en) * | 2014-10-17 | 2017-01-25 | 中国科学院武汉物理与数学研究所 | Small target infrared image processing method based on weighing local image entropy |
CN104268844A (en) * | 2014-10-17 | 2015-01-07 | 中国科学院武汉物理与数学研究所 | Small target infrared image processing method based on weighing local image entropy |
CN104992452A (en) * | 2015-06-25 | 2015-10-21 | 中国计量学院 | Flight object automatic tracking method based on thermal imaging video |
CN107657626A (en) * | 2016-07-25 | 2018-02-02 | 浙江宇视科技有限公司 | The detection method and device of a kind of moving target |
CN107280673B (en) * | 2017-06-02 | 2019-11-15 | 南京理工大学 | A kind of infrared imaging breath signal detection method based on key-frame extraction technique |
CN107280673A (en) * | 2017-06-02 | 2017-10-24 | 南京理工大学 | A kind of infrared imaging breath signal detection method based on key-frame extraction technique |
CN108364277A (en) * | 2017-12-20 | 2018-08-03 | 南昌航空大学 | A kind of infrared small target detection method of two-hand infrared image fusion |
CN110706231A (en) * | 2019-11-12 | 2020-01-17 | 安徽师范大学 | Image entropy-based three-dimensional culture human myocardial cell pulsation characteristic detection method |
CN110706231B (en) * | 2019-11-12 | 2022-04-12 | 安徽师范大学 | Image entropy-based three-dimensional culture human myocardial cell pulsation characteristic detection method |
CN111259942A (en) * | 2020-01-10 | 2020-06-09 | 西北工业大学 | Method for detecting weak target in water |
CN111259942B (en) * | 2020-01-10 | 2022-04-26 | 西北工业大学 | Method for detecting weak target in water |
CN111311191A (en) * | 2020-02-25 | 2020-06-19 | 武汉轻工大学 | Method and device for acquiring raw material quality range based on wine product quality range |
CN115249254A (en) * | 2022-09-21 | 2022-10-28 | 江西财经大学 | Target tracking method and system based on AR technology |
CN115249254B (en) * | 2022-09-21 | 2022-12-30 | 江西财经大学 | Target tracking method and system based on AR technology |
Also Published As
Publication number | Publication date |
---|---|
CN103810499B (en) | 2017-04-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103810499A (en) | Application for detecting and tracking infrared weak object under complicated background | |
CN107748873B (en) | A kind of multimodal method for tracking target merging background information | |
CN101404086B (en) | Target tracking method and device based on video | |
CN104318258A (en) | Time domain fuzzy and kalman filter-based lane detection method | |
CN106127812B (en) | A kind of passenger flow statistical method of the non-gate area in passenger station based on video monitoring | |
CN104091348A (en) | Multi-target tracking method integrating obvious characteristics and block division templates | |
Sun et al. | A novel supervised level set method for non-rigid object tracking | |
CN102289948A (en) | Multi-characteristic fusion multi-vehicle video tracking method under highway scene | |
Shi et al. | Moving target detection algorithm in image sequences based on edge detection and frame difference | |
CN109242019A (en) | A kind of water surface optics Small object quickly detects and tracking | |
CN104036526A (en) | Gray target tracking method based on self-adaptive window | |
CN110147748A (en) | A kind of mobile robot obstacle recognition method based on road-edge detection | |
CN107622507A (en) | A kind of air target tracking method based on deep learning | |
Li et al. | Infrared small target detection using tensor based least mean square | |
CN106056078A (en) | Crowd density estimation method based on multi-feature regression ensemble learning | |
CN105427348A (en) | Video object tracking method based on bat algorithm | |
CN103985139A (en) | Particle filter target tracking method based on color model and prediction vector cluster model information fusion | |
SuganyaDevi et al. | Efficient foreground extraction based on optical flow and smed for road traffic analysis | |
CN107038710B (en) | It is a kind of using paper as the Vision Tracking of target | |
CN105243661A (en) | Corner detection method based on SUSAN operator | |
CN104077608B (en) | Activity recognition method based on the slow characteristic function of sparse coding | |
Shi et al. | A novel approach of edge detection based on gray weighted absolute correlation degree and Prewitt operator | |
CN107292282B (en) | A kind of vehicle checking method updated based on semantic reasoning and space-time memory | |
Wang et al. | Construction research on multi-threshold segmentation based on improved Otsu threshold method | |
Deng et al. | Research on pedestrian detection algorithms based on video |
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: 20170412 Termination date: 20200225 |