CN107316318A - Aerial target automatic testing method based on multiple subarea domain Background fitting - Google Patents
Aerial target automatic testing method based on multiple subarea domain Background fitting Download PDFInfo
- Publication number
- CN107316318A CN107316318A CN201710384738.6A CN201710384738A CN107316318A CN 107316318 A CN107316318 A CN 107316318A CN 201710384738 A CN201710384738 A CN 201710384738A CN 107316318 A CN107316318 A CN 107316318A
- Authority
- CN
- China
- Prior art keywords
- target
- gray value
- average gray
- segmentation
- subregion
- 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
- 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
- G06T7/248—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based 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
- G06T7/254—Analysis of motion involving subtraction of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
Abstract
The invention discloses a kind of aerial target automatic testing method based on multiple subarea domain Background fitting, image to be detected is divided into P sub-regions by this method, asks for average gray value A (1)~A (P) of all subregion；Then average gray value A (1)~A (P) of all subregions is ranked up in the way of from big to small, obtains new subregion average gray value sequence A ' (1)~A ' (P)；Choose that second is black or the second white subregion average gray value is suitably zoomed in or out as segmentation threshold, and according to picture contrast according to A ' (1)~A ' (P), then carry out image binaryzation, obtain segmentation figure picture；Target positioning is carried out to segmentation figure picture again.The present invention can reduce the generation of false target, also inherit existing algorithm and be succinctly easy to hard-wired advantage.
Description
Technical field
The present invention relates to the aerial target Automatic Measurement Technique field of visual TV image or infrared image, more particularly to
A kind of aerial target automatic testing method based on multiple subarea domain Background fitting.
Background technology
With the development of information technology, carry out the Intelligent Measurement of target with the mode of Computer Vision and identification is obtained
Greatly development, especially in military field, when automatic detection and tracking to target can greatly shorten the reaction of armament systems
Between, this is most important to the performance indications for improving whole system.
Traditional real-time detection method to aerial target mainly has the target detection based on background subtraction, based on image line phase
The methods such as the aerial target detection of pass.But, these methods all have some limitations.
Object detection method basic thought based on background subtraction is to draw mesh using current frame image and background image subtraction
Logo image, still, this method can only be just effective when video camera remains static and Sky background is also at inactive state.So
And, in most cases, system needs automatic search, and video camera is kept in motion, and therefore, this method is not applied to simultaneously.
Based on the related aerial target detection method main thought of image line be based on the correlation between image adjacent lines,
Image is subjected to gray inversion first, then by the use of the average gray of certain a line as benchmark, remaining all row subtracts this
One average gray value, so as to remove image background, obtains real target.But this method practical function is preferable not to the utmost, mainly
Reason is the interference by the angle and cloud layer of illumination etc., it is seen that the sky that light video or infrared video are shot as a rule
Middle background is simultaneously uneven, when we randomly select the average value of certain a line as benchmark, for background rejecting not
Ideal, can cause more " falseness " target.Therefore, this method is not applied to yet.
The content of the invention
In view of this, the invention provides a kind of aerial target automatic testing method based on multiple subarea domain Background fitting,
On the one hand inheriting existing algorithm is succinctly easy to hard-wired advantage, simultaneously, it is contemplated that the motion of video camera and cloud layer etc.
Interference feature, targetedly obtained a more practical novel air target automatic testing method, reduced false mesh
Target is produced.
In order to solve the above-mentioned technical problem, the present invention is realized in：
A kind of aerial target automatic testing method based on multiple subarea domain Background fitting, including：
Step 1: image to be detected is divided into P sub-regions, P is the integer of setting；
Step 2: asking for the average gray value A (m), m=1,2 ... of all subregion, P；
Step 3: being ranked up to average gray value A (1)~A (P) of all subregions in the way of from big to small, obtain
To new subregion average gray value sequence A ' (m), m=1,2 ..., P；
Step 4: when target is " black " relative to background, then Th=A ' (P-1) × σ is chosen as gray threshold Th, its
In, picture contrast is bigger, and σ value is smaller, σ<1；Binarization segmentation is carried out to image using gray threshold Th, pixel value is small
Then it is set as 255 in or equal to Th, obtains segmentation figure picture；
When target is " white " relative to background, then Th=A ' (2) × σ is chosen as gray threshold Th, wherein, image pair
Bigger than degree, σ value is bigger, σ>1；Binarization segmentation is carried out to image using gray threshold Th, pixel value is more than or equal to
Th is then set as 255, obtains segmentation figure picture；
Step 5: the segmentation figure picture obtained using step 4 carries out target positioning.
Preferably, the step 5 obtains the position of target by the way of successive ignition asks for image centroid.
Preferably, the target obtained using known target minimum dimension to step 5 differentiates that target highlight number is small
In target minimum dimension, then it is assumed that be false target.
Preferably, the step 2 is：To every sub-regions, any one position of the subregion is positioned over square template
Put, the average gray value of each pixel in modulus plate is used as the average gray value A (m), m=1,2 ... of the subregion, P.
Beneficial effect：
It is black and second by second according to the average gray value of subregion it is considered herein that most black and most white part is target
The average gray value of white subregion carries out appropriate scaling as segmentation threshold and according to picture contrast, so for difference
Image, the more targeted accurate segmentation threshold of acquisition divided with the average value that randomly selects certain a line as benchmark
Cut and compare, result in more preferable segmentation effect, reduce the generation of false target.Moreover, the algorithm of this programme is very simple
It is clean to be easy to hard-wired advantage.
Brief description of the drawings
Fig. 1 is flow chart of the invention.
Embodiment
The present invention will now be described in detail with reference to the accompanying drawings and examples.
The general principle of the present invention is there is certain poor contrast, namely gray scale difference value using target and background, is passed through
Certain method accurately removes background in real time, obtains target.
Step 1: entire image is divided into P sub-regions, P is bigger under normal circumstances, and it is more accurate that background is obtained, and effect is got over
It is good.P=16 is chosen in the present embodiment.
Step 2: to every sub-regions, being positioned on any one position of the subregion, being taken with Q × Q square template
The average gray value of each pixel in template, is used as the average gray value A (m), m=1,2 ... of the subregion, P.Q value requirement
When Q × Q need more than 50% less than subregion size, preferably subregion area, take Q=10 in this preferred embodiment.
Wherein, S (m) is the region that size is Q × Q in m-th of subregion, i.e. region on template placement location, f (i,
J) it is the gray value of pixel (i, j).
Step 3: being ranked up to the subregion gray value asked in the way of from big to small, obtain new subregion and put down
Equal gray value sequence A ' (m), m=1,2 ..., P.
Step 4: when target is " black " relative to background, being in general the situation of visible images, then choosing reciprocal
2nd small average gray value is simultaneously suitably reduced as gray threshold Th, now Th=A ' (P-1) × σ, σ<1.σ is used for compensating pair
Than degree, σ value principle is：When target differs more obvious with the gray value of background, i.e., when picture contrast is larger, σ phases
It is smaller to choosing, conversely, then choosing larger.Under normal circumstances, it is 0.9 to choose σ.Then, entire image is utilized above-mentioned
Gray threshold Th carry out binary conversion treatment, obtain new segmentation figure picture：
Wherein, T (i, j) represents the pixel value of pixel (i, j) in segmentation figure picture.
However, when target relative to background be " white " when, be in general the situation of infrared image, then choose the 2nd greatly
Average gray value simultaneously suitably amplifies as gray threshold, now Th=A ' (2) × σ, σ>1.σ is used for compensating contrast, σ value
Principle is：When target differs more obvious with the gray value of background, i.e., when picture contrast is larger, σ is relative to choose larger,
Conversely, then choosing smaller.Under normal circumstances, it is 1.1 to choose σ.Then, entire image is carried out using above-mentioned gray threshold
Binary conversion treatment, obtains new segmentation figure picture：
Step 5: carrying out target positioning to the image after segmentation.The general side that image centroid is asked for by successive ignition
Formula can obtain the position of target.Concrete mode is as follows：
It is assumed that the size of image is W × H, then M in above-mentioned formula, N initial value can be set to M=W, N=H.It is upper when utilizing
State formula (1) and (2) are calculated after the first width barycenter (X1, Y1), the point centered on (X1, Y1), M and N reduce, for example, reduced
20%, barycenter (X2, Y2) is asked for again.By that analogy.Repeatedly ask for after barycenter, you can the accurate actual position for obtaining target,
Typically ask for three times.
Step 6: after asking for the position of target, it is necessary to which the true and false to target differentiates.The present embodiment utilizes target
Size is differentiated.It is assumed that the minimum dimension of target is Ws × Hs, then the point centered on target position location, with B × B region
As region to be measured, bright spot number is calculated, if bright spot number is more than or equal to Ws × Hs, it is real goal to show target, conversely,
It is then false target.Wherein, B × B regions are at least greater than Ws × Hs, can preferably select 2 times of Ws × Hs.
So far, this flow terminates.
In summary, presently preferred embodiments of the present invention is these are only, is not intended to limit the scope of the present invention.
Within the spirit and principles of the invention, any modification, equivalent substitution and improvements made etc., should be included in the present invention's
Within protection domain.
Claims (4)
1. a kind of aerial target automatic testing method based on multiple subarea domain Background fitting, it is characterised in that including：
Step 1: image to be detected is divided into P sub-regions, P is the integer of setting；
Step 2: asking for the average gray value A (m), m=1,2 ... of all subregion, P；
Step 3: being ranked up to average gray value A (1)~A (P) of all subregions in the way of from big to small, obtain new
Subregion average gray value sequence A ' (m), m=1,2 ..., P；
Step 4: when target is " black " relative to background, then Th=A ' (P-1) × σ is chosen as gray threshold Th, wherein,
Picture contrast is bigger, and σ value is smaller, σ<1；Binarization segmentation is carried out to image using gray threshold Th, pixel value is less than
Or then it is set as 255 equal to Th, obtain segmentation figure picture；
When target is " white " relative to background, then Th=A ' (2) × σ is chosen as gray threshold Th, wherein, picture contrast
Bigger, σ value is bigger, σ>1；Binarization segmentation is carried out to image using gray threshold Th, pixel value is more than or equal to Th then
It is set as 255, obtains segmentation figure picture；
Step 5: the segmentation figure picture obtained using step 4 carries out target positioning.
2. the aerial target automatic testing method as claimed in claim 1 based on multiple subarea domain Background fitting, it is characterised in that
The step 5 obtains the position of target by the way of successive ignition asks for image centroid.
3. the aerial target automatic testing method as claimed in claim 1 or 2 based on multiple subarea domain Background fitting, its feature exists
In the target obtained using known target minimum dimension to step 5 is differentiated, target highlight number is less than the minimum chi of target
It is very little, then it is assumed that to be false target.
4. the aerial target automatic testing method as claimed in claim 1 based on multiple subarea domain Background fitting, it is characterised in that
The step 2 is：To every sub-regions, it is positioned over square template on any one position of the subregion, it is each in modulus plate
The average gray value of pixel, is used as the average gray value A (m), m=1,2 ... of the subregion, P.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710384738.6A CN107316318B (en) | 2017-05-26 | 2017-05-26 | Air target automatic detection method based on multi-subregion background fitting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710384738.6A CN107316318B (en) | 2017-05-26 | 2017-05-26 | Air target automatic detection method based on multi-subregion background fitting |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107316318A true CN107316318A (en) | 2017-11-03 |
| CN107316318B CN107316318B (en) | 2020-06-02 |
Family
ID=60182137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710384738.6A Active CN107316318B (en) | 2017-05-26 | 2017-05-26 | Air target automatic detection method based on multi-subregion background fitting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107316318B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110887563A (en) * | 2019-11-18 | 2020-03-17 | 中国科学院上海技术物理研究所 | Hyperspectral area array detector bad element detection method |
| CN111862131A (en) * | 2020-07-31 | 2020-10-30 | 易思维（杭州）科技有限公司 | Adhesive tape edge detection method and application thereof |
| CN112581374A (en) * | 2019-09-29 | 2021-03-30 | 深圳市光鉴科技有限公司 | Speckle sub-pixel center extraction method, system, device and medium |
| CN116703741A (en) * | 2022-09-27 | 2023-09-05 | 荣耀终端有限公司 | Image contrast generation method and device and electronic equipment |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5859928A (en) * | 1996-06-21 | 1999-01-12 | Hewlett-Packard Company | Jitter-form background control for minimizing spurious gray cast in scanned images |
| CN102855634B (en) * | 2011-06-28 | 2017-03-22 | 中兴通讯股份有限公司 | Image detection method and image detection device |
| US9123133B1 (en) * | 2014-03-26 | 2015-09-01 | National Taipei University Of Technology | Method and apparatus for moving object detection based on cerebellar model articulation controller network |
| CN103955940B (en) * | 2014-05-16 | 2018-01-16 | 天津重方科技有限公司 | A kind of detection method of the human body cache based on X ray backscatter images |
| CN105761238B (en) * | 2015-12-30 | 2018-11-06 | 河南科技大学 | A method of passing through gray-scale statistical data depth information extraction well-marked target |
| CN106557549A (en) * | 2016-10-24 | 2017-04-05 | 珠海格力电器股份有限公司 | The method and apparatus of identification destination object |
-
2017
- 2017-05-26 CN CN201710384738.6A patent/CN107316318B/en active Active
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112581374A (en) * | 2019-09-29 | 2021-03-30 | 深圳市光鉴科技有限公司 | Speckle sub-pixel center extraction method, system, device and medium |
| CN110887563A (en) * | 2019-11-18 | 2020-03-17 | 中国科学院上海技术物理研究所 | Hyperspectral area array detector bad element detection method |
| CN110887563B (en) * | 2019-11-18 | 2021-10-01 | 中国科学院上海技术物理研究所 | Hyperspectral area array detector bad element detection method |
| CN111862131A (en) * | 2020-07-31 | 2020-10-30 | 易思维（杭州）科技有限公司 | Adhesive tape edge detection method and application thereof |
| CN111862131B (en) * | 2020-07-31 | 2021-03-19 | 易思维（杭州）科技有限公司 | Adhesive tape edge detection method and application thereof |
| CN116703741A (en) * | 2022-09-27 | 2023-09-05 | 荣耀终端有限公司 | Image contrast generation method and device and electronic equipment |
| CN116703741B (en) * | 2022-09-27 | 2024-03-15 | 荣耀终端有限公司 | Image contrast generation method and device and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107316318B (en) | 2020-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Deng et al. | Infrared moving point target detection based on spatial–temporal local contrast filter | |
| CN106650665B (en) | Face tracking method and device | |
| Bai et al. | Enhancement of dim small target through modified top-hat transformation under the condition of heavy clutter | |
| CN107316318A (en) | Aerial target automatic testing method based on multiple subarea domain Background fitting | |
| CN101976436B (en) | Pixel-level multi-focus image fusion method based on correction of differential image | |
| Park | Shape-resolving local thresholding for object detection | |
| CN105469090B (en) | Small target detecting method and device in infrared image based on frequency-domain residual | |
| US20150092051A1 (en) | Moving object detector | |
| Wang et al. | Clutter-adaptive infrared small target detection in infrared maritime scenarios | |
| CN101326549A (en) | Method for detecting streaks in digital images | |
| JP2007156655A (en) | Variable region detection apparatus and its method | |
| CN111191535B (en) | Pedestrian detection model construction method based on deep learning and pedestrian detection method | |
| Zhao et al. | Principal curvature for infrared small target detection | |
| US20200302155A1 (en) | Face detection and recognition method using light field camera system | |
| US9014426B2 (en) | Method and device for the detection of moving objects in a video image sequence | |
| JP5367244B2 (en) | Target detection apparatus and target detection method | |
| CN103607558A (en) | Video monitoring system, target matching method and apparatus thereof | |
| JP4818285B2 (en) | Congestion retention detection system | |
| JP2014048131A (en) | Image processing device, method, and program | |
| CN113409334B (en) | Centroid-based structured light angle point detection method | |
| US10331977B2 (en) | Method for the three-dimensional detection of objects | |
| CN111325073A (en) | Monitoring video abnormal behavior detection method based on motion information clustering | |
| CN116703755A (en) | Omission risk monitoring system for medical waste refrigeration house | |
| Duncan et al. | Relational entropy-based saliency detection in images and videos | |
| JP2009116686A (en) | Imaging target detection apparatus and method |
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 |