CN103530891A - Extended target stable trace point extracting method based on minimum inscribed circle of head - Google Patents
Extended target stable trace point extracting method based on minimum inscribed circle of head Download PDFInfo
- Publication number
- CN103530891A CN103530891A CN201310475804.2A CN201310475804A CN103530891A CN 103530891 A CN103530891 A CN 103530891A CN 201310475804 A CN201310475804 A CN 201310475804A CN 103530891 A CN103530891 A CN 103530891A
- Authority
- CN
- China
- Prior art keywords
- target
- point
- head
- theta
- trace point
- 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)
Abstract
The invention discloses an extended target stable trace point extracting method based on a minimum inscribed circle of a head. The extended target stable trace point extracting method based on the minimum inscribed circle of the head comprises the following steps of S1: computing the contour of an extended target after being subjected to target contour filtering and sharp segmental arc eliminating in a direction of a target head to obtain shift centroid of the target head; S2: referring to the shift centroid, obtaining the minimum inscribed circle of an extended target contour head; S3: utilizing the intersection point, in the leading edge direction of the target contour head, of two crossing points of the minimum inscribed circle and a target principal axis straight line as a final trace point so as to reduce the dithering of the trace point. According to the extended target stable trace point extracting method based on the minimum inscribed circle of the head, which is disclosed by the invention, the method for obtaining the trace point starts from the interior of the target, so that the characteristic that the grayscale of the target is comparatively uniform is fully utilized, the influence of the unstable airflow and the other disturbance outside the target is excluded, the target trace point finally obtained by the method disclosed by the invention has higher stability compared with the target trace point obtained by the traditional method, the dithering in a tracking process is reduced, and the risk of trace failure is also reduced.
Description
Technical field
The invention belongs to computer vision and art of image analysis, relate to a kind of target following point extracting method, particularly a kind of expansion target tenacious tracking point extracting method based on the minimum incircle of head.
Background technology
Be accompanied by optical imagery and follow the tracks of the development in field, for target with high precision tenacious tracking, the tenacious tracking demand particularly having under the expansion target conditions of large wake flame or tail cigarette is more and more urgent.When target exists in strong luminescence or heating situation, target area occupied in the TV visual field of imaging tracing system sharply increases, and conventionally will account for pick the more than 1/3 of visual field, when serious, even may exceed visual field.Now, it cannot be guaranteed target in global feature point as the tracking mode of the centre of form, barycenter and not overflow visual field, only adopts the tracking getting rid of based on local feature point and just has higher tracking robustness.
Yet, the modes such as traditional angle point, edge are difficult to adapt to the fuzzy behaviour of object edge and the variation of targeted attitude in tracing process, cause these trace point extracting modes cannot adapt to the clear target of such blur margin, easily cause unique point to be difficult to correspondence in interframe, extract the situation appearance that drift occurs point
[1-2].The trace point extracting method in based target forward position is applicable to the tracking of this class expansion target very much, but, because target body is luminous or it is inhomogeneous to generate heat, cause the trace point of extraction affected by unsettled Fujian value segmentation result, and then cause the violent shake of trace point.Document [3] adopts the method for contour smoothing filtering to eliminate unsettled sharp-pointed segmental arc on target two-value profile, reduced the impact that target body extracts for trace point, reduced to a certain extent the shake of trace point, but, because the string arc of its use is more limited than the adaptability of Fujian value, the extraction shake of final trace point is in most cases still greater than a pixel.
List of references
[1] Fan Zhihua, Wang Chunhong ，Wu Ming army. " the expansion method for tracking target based on geometric properties point ",
Photoelectric project, 2009,36 (10), 12-15,23.
[2] Zhang Kai, Liu Zhanai, Zhong Doudou, Yan Jie. " research of infrared seeker head edge template matches track algorithm ",
Northwestern Polytechnical University's journal, 2008,26 (3), 331-334.
[3] Lei Tao, Jiang Ping, week enters, Wu Yinzhang, " a kind of expansion target tenacious tracking point extracting method ",
Patent first please number: 201210268512.7.
Summary of the invention
The present invention is in order to overcome the deficiencies in the prior art, a kind of tenacious tracking point extracting method based on the minimum incircle of head is provided, overcome the impact that the unstable air-flow of target external and target body change, make full use of the characteristic that target wake flame interior intensity is evenly distributed, further reduced expansion near edge of the objective tracking in document [3] and obtained the shake of extracting point.
In order to achieve the above object, the calculation procedure of a kind of extracting method of expanding target tenacious tracking point provided by the invention is as follows:
Step S1: to through objective contour filtering and eliminate the expansion objective contour after sharp-pointed segmental arc, adopt target cephalad direction to calculate, obtain target head bias barycenter;
Step S2: take this skew barycenter is benchmark, gets the minimum incircle of this expansion objective contour head;
Step S3: be positioned at the intersection point of forward position direction of objective contour head in this minimum incircle and target main shaft straight line are intersected 2 as final trace point, thereby reduce the shake of trace point.
The tracking position tool that the present invention extracts has the following advantages:
(1) trace point of choosing is positioned at the head of target main outline, has avoided target to be full of the failure of cutting apart that visual field causes.
(2) point of the forefront based on extracting after contour smoothing, and take that this puts the incenter of being asked for as benchmark, more stable than the center of circle of directly asking for without smoothing processing.
(3) because this method of asking for trace point is from target internal, the trace point obtaining by asking for the intersection point of the minimum incircle of target head and main shaft, take full advantage of the uniform characteristic of target internal intensity profile, the impact of having avoided the outside unstable air-flow of objective contour and target body to disturb, the tracking position finally obtaining compares to classic method and has higher stability, significantly reduced the shake of trace point, reduced and followed the tracks of the risk losing efficacy, improved the precision of target measurement and the stability that output tracking number is gathered around.
Trace point and the difference between actual value that the inventive method is extracted are less, and rock littlely, extract the impact that point changes by target body less than document [3] method, and the trace point obtaining is more stable.
Accompanying drawing explanation
Fig. 1 is the process flow diagram that tenacious tracking point of the present invention extracts;
Fig. 2 a is the schematic diagram that in embodiment, N frame tracking position extracts result;
Fig. 2 b is the schematic diagram that in embodiment, N ten l frame tracking positions extract result;
Fig. 3 is the shake of extraction result of the present invention and the comparison that document [3] method is extracted jitter conditions in result orientation;
Fig. 4 is the shake of extraction result of the present invention and the comparison that document [3] method is extracted jitter conditions in result pitching.
Embodiment
For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, to this grey bright further description.
The extraction of expansion target tenacious tracking point of the present invention can realize by hardware mode, also can realize by software mode.For example, at personal computer, industrial computer and server. upper with the form of software, install and carry out, also can make embedded chip and embody with the form of hardware.With reference to the accompanying drawings the specific embodiment of the present invention is described.
The present invention is applicable to imaging tracking equipment the expansion target in visual field is followed the tracks of, and especially, when target range imaging device is nearer, the imaging in visual field of target body is comparatively obvious, and target has the tracking in stronger wake flame or tail cigarette situation simultaneously.Concrete, expansion target tenacious tracking point extraction step comprises:
Step S1: to through objective contour filtering and eliminate the expansion objective contour after sharp-pointed segmental arc, adopt target cephalad direction to calculate, obtain target head bias barycenter;
Step S2: take this skew barycenter is benchmark, gets the minimum incircle of this expansion objective contour head;
Step S3: be positioned at the intersection point of forward position direction of objective contour head in this minimum incircle and target main shaft straight line are intersected 2 as final trace point, thereby reduce the shake of trace point.
In wood inventive method, expand the flow process of target tenacious tracking point extraction as shown in Figure 1, the expansion target two-value profile of having eliminated sharp-pointed segmental arc after objective contour smothing filtering of take is basis, first completing expansion target cephalad direction and forefront point calculates, then centered by this forward position point, adopt isotropic border circular areas to calculate, try to achieve than the better target head bias of this initial forward position point stability barycenter, and then use this barycenter for calculating the minimum incircle of trying to achieve expansion target contouring head in the center of circle, finally ask for the intersection point of this minimum incircle and target main shaft straight line, getting the point near target head forward position direction in intersection point is final trace point:
Concrete, the trace point that the present invention asks for based on the minimum incircle of expansion objective contour head comprises the following steps:
1. first according to the target two-value profile after level and smooth, calculate target cephalad direction θ
n.
θ wherein
α xisfor the minimum principal axis of inertia direction of objective contour, θ
vfor the angle of target at current time target aggregate velocity projection and X-axis forward in picture plane, wherein α xisv, n are corresponding footnote.
2. in the following manner along target cephalad direction θ
ndirection is asked for the forefront point P of objective contour cephalad direction
α, wherein α is corresponding footnote, as shown in Figure 2 a:
Objective contour is pressed to objective contour cephalad direction θ
nbe rotated, based target main shaft and direction of motion calculate the forefront point P along objective contour cephalad direction
α, i point be cephalad direction θ according to target
nabscissa value xv α lue[i after being rotated] be expressed as follows:
Xv α lue[7] two P
ix * cos θ
n+ Pi
,y * sin θ
n(2)
P wherein
ix and P
iy is respectively holistic symbol and represents, has represented the horizontal ordinate of i point and the coordinate figure of ordinate, cos θ
nwith as sin θ
nbe respectively angle θ
ncosine value and sine value, xv α lue[i] point while getting maximal value is the forefront point p α of objective contour.Wherein, i ∈ [O, nCount], nCount is objective contour point number.
3. with the forefront point P of objective contour cephalad direction
αfor the center of circle, radius of a circle is R
α(R herein
αvalue is 50), calculate the barycenter p of this circle regional area image
b, the barycenter P calculating
bpoint as shown in Figure 2 a.
4. as Fig. 2 a, with center of mass point P
bfor the center of circle, the minimum incircle of asking for expansion objective contour head is designated as C
ins, wherein adding s, b is corresponding footnote.
5. cross center of mass point P
b, ask for objective contour cephalad direction θ
nthe straight line L of direction
p, at this straight line L
pwith incircle C
instwo points that intersect are P
t1with P
t2in, get the final trace point P of conduct near objective contour cephalad direction
t, wherein, t1.t2.t, p are corresponding footnote.
Fig. 2 a shown and calculates the schematic diagram of following the tracks of and shown that in certain sequence image, N frame trace point extracts result, and Fig. 2 b has shown that the trace point of N ten l frames in this sequence image extracts result, and N is the sequence number of a certain frame in tracking image.In figure, crosshair has represented the trace point that adopts the inventive method finally to calculate.From Fig. 2 a and Fig. 2 b, can find out, although all carried out profile filtering between two continuous frames, because original filtering algorithm adaptability is not enough, cause target body still to have larger saltus step in interframe.Even used in this case the method for document [3], still can there is certain shake in the trace point of extraction.And adopt in the result of the inventive method extraction, although cut apart the unstable shake that can cause incenter of rear target body,, so only can cause the variation of the center of circle and inscribed circle radius, and incircle and straight line L
pintersection point near the intersection point of objective contour cephalad direction, but can keep stable.
Fig. 3 and Fig. 4 shown on the profile filtering basis based on document [3], relatively add that the present invention proposes may be before and after the minimum incircle algorithm of head the stability of the trace point that extracts.By Fig. 3, can be found out with circle 4, added and compiled after the minimum incircle algorithm of head, the stability of the trace point extracting has further had significantly raising.Wherein, in used test image, the corresponding Space Angle resolution of every pixel is 15.23 rads, and in figure, curve is that actual catching got a composite value and it carried out to the poor of result after 7 least square fittings, weighs the shake of tracking by this difference.Fig. 3 has shown the track and extract point shake in orientation, and Fig. 4 has shown the track and extract point shake in pitching.By result, can find out, extraction result of the present invention is obviously better than the extraction result of document [3].Data demonstration, after processing by the inventive method, the shake RMS that extracts point in orientation is 1.17 "; being converted to pixel is 0.077 pixel; and extract shake RMS in pitching, be 3.87 ", being converted to pixel is 0.25 pixel, trace point extraction stability is further largely increased.
As can be seen from the above results, the trace point that the inventive method is extracted has further reduced shake, extracts point and is subject to the impact of target body variation less than former method, and the stability that trace point extracts is obviously better than original method.The above; be only the embodiment in the present invention, but protection scope of the present invention is not limited to this, any people who is familiar with this technology is in the disclosed technical scope of the present invention; can understand conversion or the replacement expected, all should be encompassed in wood invention comprise scope within.
Claims (6)
1. the expansion target tenacious tracking point extracting method based on the minimum incircle of head, is characterized in that, expansion target tenacious tracking point extraction step comprises:
Step S1: to through objective contour filtering and eliminate the expansion objective contour after sharp-pointed segmental arc, adopt target cephalad direction to calculate, obtain target head bias barycenter;
Step S2: take this skew barycenter is benchmark, gets the minimum incircle of this expansion objective contour head;
Step S3: be positioned at the intersection point of forward position direction of objective contour head in this minimum incircle and target main shaft straight line are intersected 2 as final trace point, thereby reduce the shake of trace point.
2. expand as claimed in claim 1 the extracting method of tracking position, it is characterized in that, the step of asking for the minimum incircle of expansion objective contour head is as follows:
Step 21: get in the following way target cephalad direction θ
n
θ wherein
α xisfor the minimum principal axis of inertia direction of target, θ
vfor the angle of target at its aggregate velocity of current time projection and X-axis forward in picture plane;
Step 22: along target cephalad direction θ
nask for the forefront point P of objective contour cephalad direction
α;
Step 23: with objective contour forefront point P
αfor the center of circle, radius is R
α, calculate the barycenter P of isotropic circular regional area
b;
Step 24: with barycenter P
bpoint is the center of circle, asks for the minimum incircle C of expansion objective contour
ins.
3. expand as claimed in claim 1 the extracting method of tracking position, it is characterized in that, ask for final trace point step as follows:
Step 31: ask for target cephalad direction θ
nfor deflection, cross center of mass point P
bstraight line L
p
Step 32: at straight line L
pwith incircle C
instwo intersection point P
t1with P
t2in, get the intersection point of close target cephalad direction as final trace point P
t.
4. expand as claimed in claim 2 the extracting method of tracking position, it is characterized in that, adopt based target main shaft and direction of motion to try to achieve the forefront point of objective contour cephalad direction.
5. expand as claimed in claim 4 the extracting method of tracking position, it is characterized in that, centered by this forefront point, adopt isotropic border circular areas to try to achieve than this forefront point and stablize quiet better target head bias barycenter.
6. expand as claimed in claim 1 the extracting method of tracking position, it is characterized in that, take this target head bias barycenter as the center of circle, to calculate and get the minimum incircle of expanding target contouring head after filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310475804.2A CN103530891B (en) | 2013-10-12 | 2013-10-12 | A kind of Extended target tenacious tracking point extracting method based on the minimum incircle of head |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310475804.2A CN103530891B (en) | 2013-10-12 | 2013-10-12 | A kind of Extended target tenacious tracking point extracting method based on the minimum incircle of head |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103530891A true CN103530891A (en) | 2014-01-22 |
| CN103530891B CN103530891B (en) | 2016-03-30 |
Family
ID=49932869
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310475804.2A Active CN103530891B (en) | 2013-10-12 | 2013-10-12 | A kind of Extended target tenacious tracking point extracting method based on the minimum incircle of head |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103530891B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112116632A (en) * | 2020-09-21 | 2020-12-22 | 中国科学院长春光学精密机械与物理研究所 | Method, device and medium for tracking target along with target tail smoke |
| CN113746451A (en) * | 2021-08-10 | 2021-12-03 | 深圳市乐创享科技有限公司 | Filtering operation method of circular filter |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831619A (en) * | 2012-07-31 | 2012-12-19 | 中国科学院光电技术研究所 | Expansion target stable trace point extraction method |
| CN103136525A (en) * | 2013-02-28 | 2013-06-05 | 中国科学院光电技术研究所 | Hetero-type expanded goal high-accuracy positioning method with generalized Hough transposition |
| CN103247032A (en) * | 2013-04-26 | 2013-08-14 | 中国科学院光电技术研究所 | Method for positioning slight expanded target based on gesture compensation |
-
2013
- 2013-10-12 CN CN201310475804.2A patent/CN103530891B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831619A (en) * | 2012-07-31 | 2012-12-19 | 中国科学院光电技术研究所 | Expansion target stable trace point extraction method |
| CN103136525A (en) * | 2013-02-28 | 2013-06-05 | 中国科学院光电技术研究所 | Hetero-type expanded goal high-accuracy positioning method with generalized Hough transposition |
| CN103247032A (en) * | 2013-04-26 | 2013-08-14 | 中国科学院光电技术研究所 | Method for positioning slight expanded target based on gesture compensation |
Non-Patent Citations (2)
| Title |
|---|
| YAN WEN等: "《A Robust Method of Detecting Hand Gestures Using Depth Sensors》", 《HAPTIC AUDIO VISUAL ENVIRONMENTS AND GAMES (HAVE), 2012 IEEE INTERNATIONAL WORKSHOP ON》 * |
| 江速勇: "《基于最大内接的圆曲线拟合算法编程》", 《内燃机》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112116632A (en) * | 2020-09-21 | 2020-12-22 | 中国科学院长春光学精密机械与物理研究所 | Method, device and medium for tracking target along with target tail smoke |
| CN112116632B (en) * | 2020-09-21 | 2023-12-05 | 中国科学院长春光学精密机械与物理研究所 | Method, device and medium for tracking target along tail smoke of target |
| CN113746451A (en) * | 2021-08-10 | 2021-12-03 | 深圳市乐创享科技有限公司 | Filtering operation method of circular filter |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103530891B (en) | 2016-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105023278B (en) | A kind of motion target tracking method and system based on optical flow method | |
| US9619709B2 (en) | Visual localization of unmanned aerial vehicles based on marker detection and processing | |
| WO2016015547A1 (en) | Machine vision-based method and system for aircraft docking guidance and aircraft type identification | |
| WO2017041396A1 (en) | Driving lane data processing method, device, storage medium and apparatus | |
| CN104484868B (en) | The moving target of a kind of combination template matches and image outline is taken photo by plane tracking | |
| WO2019129255A1 (en) | Target tracking method and device | |
| CN103116896A (en) | Visual saliency model based automatic detecting and tracking method | |
| CN107305632A (en) | Destination object distance measurement method and system based on monocular computer vision technique | |
| CN103325099A (en) | Image correcting method and device | |
| Nassu et al. | A vision-based approach for rail extraction and its application in a camera pan–tilt control system | |
| CN103886325A (en) | Cyclic matrix video tracking method with partition | |
| CN105741325A (en) | Moving target tracking method and moving target tracking equipment | |
| US20160132734A1 (en) | Apparatus and method for detecting object for vehicle | |
| CN105809718A (en) | Object tracking method with minimum trajectory entropy | |
| CN105654518B (en) | A kind of trace template adaptive approach | |
| CN103247032A (en) | Method for positioning slight expanded target based on gesture compensation | |
| CN104168444A (en) | Target tracking method of tracking ball machine and tracking ball machine | |
| CN103914829A (en) | Method for detecting edge of noisy image | |
| CN111861968B (en) | Infrared dim target detection method and detection system | |
| CN103617631A (en) | Tracking method based on center detection | |
| CN103530891A (en) | Extended target stable trace point extracting method based on minimum inscribed circle of head | |
| Wang et al. | Hand posture recognition from disparity cost map | |
| CN105160341A (en) | Triangle detection method based on linear segment detection and inner natures | |
| Hao et al. | LMVI-SLAM: Robust low-light monocular visual-inertial simultaneous localization and mapping | |
| CN103488801A (en) | Geographical information space database-based airport target detection 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |