CN105427286A - Gray scale and gradient segmentation-based infrared target detection method - Google Patents

Gray scale and gradient segmentation-based infrared target detection method Download PDF

Info

Publication number
CN105427286A
CN105427286A CN201510753554.3A CN201510753554A CN105427286A CN 105427286 A CN105427286 A CN 105427286A CN 201510753554 A CN201510753554 A CN 201510753554A CN 105427286 A CN105427286 A CN 105427286A
Authority
CN
China
Prior art keywords
row
pos
gradient
gray scale
detection method
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
Application number
CN201510753554.3A
Other languages
Chinese (zh)
Other versions
CN105427286B (en
Inventor
白俊奇
杜瀚宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Lesi Electronic Equipment Co., Ltd.
Original Assignee
CETC 28 Research Institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by CETC 28 Research Institute filed Critical CETC 28 Research Institute
Priority to CN201510753554.3A priority Critical patent/CN105427286B/en
Publication of CN105427286A publication Critical patent/CN105427286A/en
Application granted granted Critical
Publication of CN105427286B publication Critical patent/CN105427286B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10048—Infrared image

Landscapes

  • Image Analysis (AREA)

Abstract

The invention discloses a gray scale and gradient segmentation-based infrared target detection method. The method comprises the following steps: 1) calculating a gradient image Phi(x, y) by utilizing an original image Forg; 2) carrying out non-maximum value inhibition on the Phi(x, y) to obtain an image Phi1(x, y); 3) carrying out edge linking and edge closing judgement by utilizing the Phi1(x, y), and calculating a candidate target area Pos1; 4) selecting a local image Floc according to the Pos1 and segmenting the Floc to obtain a candidate target area Pos2; and 5) calculating an optimum target area Pos according to a formula (as shown in the specification).

Description

A kind of infrared target detection method split based on gray scale and gradient
Technical field
The present invention relates to infrared image processing field, a kind of particularly infrared target detection method split based on gray scale and gradient of applicable hardware real-time implementation
Background technology
Infrared imaging, as the important supplement means of radar system, has passive detection, is not easily found, the feature such as all weather operations, is widely used at Military and civil fields.In recent years, infrared imaging detection becomes the focus and emphasis of Chinese scholars research, proposes many valuable detection methods.The people such as Zhang Qiang propose to split based on the infrared small object of local maximum, treat detected image carry out enhancing process by Gaussian template; For the deficiency that fixed size filtering core in traditional filtering method shows Dim targets detection, Gong Junliang proposes a kind of method for detecting infrared puniness target based on Scale-space theory; The people such as Li Qiuhua adopt the Dual band IR subject fusion of D-S evidence theory to detect; For the imaging characteristic of variable resolution Small object infrared under cloudy background, Zhao Xiao proposes a kind of method for detecting infrared puniness target based on Scale-space theory; A kind of method that the people such as LiuZH propose images match realizes compensating detector motion, uses method of difference to realize moving object detection.
There is following shortcoming in existing target detection technique: the scene bad adaptability of (1) most detection algorithm, if scene changes is large, can reduce the precision of target detection; (2) under complex scene, target detection capabilities is poor; (3) But most of algorithms adopts single feature to detect, poor robustness.
Summary of the invention
Goal of the invention: technical matters to be solved by this invention is for the deficiencies in the prior art, provides a kind of infrared target detection method split based on gray scale and gradient.
In order to solve the problems of the technologies described above, the invention discloses a kind of infrared image object detection method, comprising the following steps:
(1) original image F is utilized orgcompute gradient image
(2) right carry out non-maximal value suppression, obtain image
(3) utilize carry out edge conjunction, edge closure judges, calculated candidate target area Pos 1;
(4) according to Pos 1choose topography F loc, to F locsplit, obtain candidate target region Pos 2;
(5) according to Pos=Pos 1∩ Pos 2calculate optimal objective region Pos.
The present invention is based in the infrared target detection method of gray scale and gradient segmentation, by Sobel operator compute gradient image
The present invention is based in the infrared target detection method of gray scale and gradient segmentation, edge closure judges whether comprehensive row, column direction closure is object edge point to evaluate current point.Decision criteria comprises: 1) line direction closure, column direction closure are greater than threshold value T simultaneously 1, 0≤T 1≤ 1; 2) if the i-th leftmost effective value position of row is Col mini (), rightmost effective value position is Col max(i), Col deci () > T represents the i-th behavior effectively row, otherwise the i-th behavior inactive line, wherein, Col dec(i)=Col max(i)-Col min(i), T is threshold value, 3≤T≤10, then line direction closure computing formula is as follows:
3) if jth row effective value position is topmost Row minj (), effective value position is bottom Row max(j), Row decj () > T represents that jth is classified as effective row, otherwise jth is classified as invalid row, wherein, and Row dec(j)=Row max(j)-Row min(j), T is threshold value, 3≤T≤10, then column direction closure computing formula is as follows:
The present invention is based in the infrared target detection method of gray scale and gradient segmentation, determined the segmentation threshold Q of gray level image by maximum variance between clusters (" microcomputer information ", the 26th volume 12-2 phase in 2010);
The present invention is based in the infrared target inspection method of gray scale and gradient segmentation, pass through Pos 1and Pos 2optimal objective region Pos is obtained with computing.
The present invention compared with prior art, has following remarkable advantage: 1) target detection is carried out in employing bicharacteristic (gray scale, gradient), Algorithm robustness is strong; 2) split local gray level image, Iamge Segmentation degree of accuracy is high; 3) gray threshold is calculated by scene adaptive, algorithm scene strong adaptability; 4) there is not high exponent arithmetic(al) and labyrinth, algorithm operation quantity is little, is easy to hardware real-time implementation.
Accompanying drawing explanation
To do the present invention below in conjunction with the drawings and specific embodiments and further illustrate, above-mentioned and/or otherwise advantage of the present invention will become apparent.
Fig. 1 is process flow diagram of the present invention.
Fig. 2 a is the original image of embodiment.
Fig. 2 b is the gradient segmentation result of embodiment.
Fig. 2 c is the intensity slicing result of embodiment.
Fig. 2 d is the final detection result of embodiment.
Embodiment
The invention discloses a kind of infrared target detection method split based on gray scale and gradient, comprise the following steps:
(1) original image F is utilized orgcompute gradient image
(2) right carry out non-maximal value suppression, obtain image
(3) utilize carry out edge conjunction, edge closure judges, calculated candidate target area Pos 1;
(4) according to Pos 1choose topography F loc, to F locsplit, obtain candidate target region Pos 2;
(5) according to Pos=Pos 1∩ Pos 2calculate optimal objective region Pos.
In step (1), use Sobel operator compute gradient image
In step (2), right carry out " non-maximal value suppression ": in the 3*3 neighborhood centered by pixel data, be defined as by gradient direction centered by pixel data, horizontal direction, the direction of vertical direction and positive and negative miter angle is total to the line of four direction.All directions this direction and pixel data neighborhood pixels compare, and determine local maximum.
data
If the gradient direction of the pixel data of center belongs to vertical direction, then the neighborhood pixels of pixel data and vertical direction is compared, if pixel value data is not maximal value, set to 0.
In step (3), edge closure judges whether comprehensive row, column direction closure is object edge point to evaluate current point.Decision criteria comprises: 1) line direction closure, column direction closure are greater than threshold value T simultaneously 1, 0≤T 1≤ 1; 2) if the i-th leftmost effective value position of row is Col mini (), rightmost effective value position is Col max(i), Col deci () > T represents the i-th behavior effectively row, otherwise the i-th behavior inactive line, wherein, Col dec(i)=Col max(i)-Col min(i), T is threshold value, 3≤T≤10; 3) if jth row effective value position is topmost Row minj (), effective value position is bottom Row max(j), Row decj () > T represents that jth is classified as effective row, otherwise jth is classified as invalid row, wherein, and Row dec(j)=Row max(j)-Row min(j), T is threshold value, 3≤T≤10;
In step (4), topography F loccomprise Pos 1rectangular area, utilize maximum variance between clusters to F loccarry out intensity slicing, obtain candidate target region Pos 2;
In step (5), pass through Pos 1and Pos 2optimal objective region is obtained with computing.
Embodiment 1
Present embodiment discloses a kind of infrared target detection method split based on gray scale and gradient, comprise the following steps:
(1) original image F is utilized orgcompute gradient image
(2) right carry out non-maximal value suppression, obtain image
(3) utilize carry out edge conjunction, edge closure judges, calculated candidate target area Pos 1;
(4) according to Pos 1choose topography F loc, to F locsplit, obtain candidate target region Pos 2;
(5) according to Pos=Pos 1∩ Pos 2calculate optimal objective region Pos.
In step (1), use Sobel operator compute gradient image
In step (2), carry out " non-maximal value suppression ", gradient direction is defined as AA line direction (i.e. AA-data-AA), BB line direction (i.e. BB-data-BB), CC line direction (i.e. CC-data-CC), DD line direction (i.e. DD-data-DD) four direction, all directions neighborhood pixels compares, determine local maximum, as shown in the table:
DD CC BB
AA data AA
BB CC DD
If the gradient direction of the pixel value data of center belongs to DD-data-DD, then pixel value data and pixel value DD is compared, if pixel value data is not maximal value, set to 0.
In step (3), edge closure judges whether comprehensive row, column direction closure is object edge point to evaluate current point.Decision criteria comprises: 1) row, column direction closure is greater than threshold value T simultaneously 1; 2) for line direction closure, row difference Col deci () > T represents the i-th behavior candidate target region, T is threshold value; 3) for column direction closure, the row difference Row between two row decj () > T represents that jth is classified as candidate target region, T is threshold value.
In step (4), topography F loccomprise Pos 1rectangular area, utilize maximum variance between clusters to F loccarry out intensity slicing, obtain candidate target region Pos 2;
In step (5), pass through Pos 1and Pos 2optimal objective region is obtained with computing.
Embodiment 2
Composition graphs 1, illustrates the infrared target detection method that the present invention is based on gray scale and gradient segmentation below with example.The number of pixels 640 × 512 of infrared image, frame frequency 50HZ.Digital signal passes to the special image disposable plates of FPGA+DSP framework by optical fiber, and the infrared target detection method based on gray scale and gradient segmentation realizes in dsp processor, and the processing time is less than 20ms, and meet the demand of process in real time, concrete implementation step is as follows:
(1) original image F is obtained org, utilize Sobel operator compute gradient image
F orgbe 14-bit digital picture, obtain gradient image by Sobel operator
(2) right carry out non-maximal value suppression, obtain image
(3) calculate line direction closure, hypothetical target edge at image 82nd ~ 87 row, the row minimum value Col that target is expert at min, row maximal value Col maxbe respectively Col min={ 122,118,119,118,120,121} and Col max={ 129,125,129,128,127,123}.The row difference that target is expert at is Col deccan be expressed as:
Col dec＝Col max-Col min
＝{7,7,10,10,7,2}
As threshold value T=3, can be calculated by decision criteria, line direction closure is 83.3%.
Calculated column direction closure, hypothetical target edge arranges at image 82nd ~ 87, the row minimum value Row of target column min, row maximal value Row maxbe respectively Row min={ 122,118,119,118,120,121} and Row max={ 129,125,129,128,127,123}.The row difference of target column is Row deccan be expressed as:
Row dec＝Row max-Row min
＝{7,7,10,10,7,2}
As threshold value T=3, can be calculated by decision criteria, column direction closure is 83.3%.
(4) right when carrying out edge conjunction, setting high threshold is 0.85, and Low threshold is 0.34, carries out edge conjunction, closed judgement operation, obtain candidate target region Pos by high and low threshold value 1;
(5) the intensity slicing threshold value utilizing maximum variance between clusters to obtain is 896, utilizes intensity slicing threshold value to local Image Segmentation Using, obtains candidate target region Pos 2;
(6) Pos is passed through 1and Pos 2optimal objective region is obtained with computing.
Embodiment 3
In Fig. 2, Fig. 2 a represents original image, and Fig. 2 b represents gradient segmentation result, and Fig. 2 c is intensity slicing result, and Fig. 2 d is final detection result.Image shows, and the infrared target detection method target detection based on gray scale and gradient segmentation that the present invention proposes is accurately high.
The invention provides a kind of infrared target detection method split based on gray scale and gradient; the method and access of this technical scheme of specific implementation is a lot; the above is only the preferred embodiment of the present invention; should be understood that; for those skilled in the art; under the premise without departing from the principles of the invention, can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.The all available prior art of each ingredient not clear and definite in the present embodiment is realized.

Claims (6)

1., based on the infrared target detection method that gray scale and gradient are split, it is characterized in that, comprise the following steps:
Step (1), utilizes original image F orgcompute gradient image
Step (2), to gradient image carry out non-maximal value suppression, obtain image
Step (3), utilizes image carry out edge conjunction, edge closure judges, calculated candidate target area Pos 1;
Step (4), according to candidate target region Pos 1choose topography F loc, to topography F locsplit, obtain candidate target region Pos 2;
Step (5), according to formula Pos=Pos 1∩ Pos 2calculate optimal objective region Pos.
2. a kind of infrared target detection method split based on gray scale and gradient according to claim 1, is characterized in that, in step (1), uses Sobel operator compute gradient image
3. a kind of infrared target detection method split based on gray scale and gradient according to claim 1, is characterized in that, in step (2), to gradient image carry out non-maximal value and suppress process;
Non-maximal value suppresses process to comprise: in the 3*3 neighborhood centered by pixel data, be defined as by gradient direction centered by pixel data, and the direction of horizontal direction and vertical direction and positive and negative miter angle is total to the line of four direction; All directions this direction and pixel data neighborhood pixels compare, and determine local maximum; If the gradient direction of the pixel data of center belongs to vertical direction, then the neighborhood pixels of pixel data and vertical direction is compared, if pixel value data is not maximal value, set to 0.
4. a kind of infrared target detection method split based on gray scale and gradient according to claim 1, it is characterized in that, in step (3), edge closure is used to judge that comprehensive row, column direction closure is to judge whether current point is object edge point, meets the following conditions simultaneously:
1) line direction closure, column direction closure are greater than threshold value T simultaneously 1, 0≤T 1≤ 1;
2) if the i-th leftmost effective value position of row is Col mini (), rightmost effective value position is Col max(i), Col deci () > T represents the i-th behavior effectively row, otherwise the i-th behavior inactive line, wherein, Col dec(i)=Col max(i)-Col min(i), T is threshold value, 3≤T≤10;
Wherein, line direction closure computing formula is as follows:
3) if jth row effective value position is topmost Row minj (), effective value position is bottom Row max(j), Row decj () > T represents that jth is classified as effective row, otherwise jth is classified as invalid row, wherein, and Row dec(j)=Row max(j)-Row min(j), T is threshold value, 3≤T≤10;
Wherein, column direction closure computing formula is as follows:
5. a kind of infrared target detection method split based on gray scale and gradient according to claim 1, is characterized in that, in step (4), and topography F loccomprise Pos 1rectangular area, utilize maximum variance between clusters to F loccarry out intensity slicing, obtain candidate target region Pos 2.
6. the infrared target detection method split based on gray scale and gradient according to claim 1, is characterized in that, in step (5), by topography Pos 1with topography Pos 2optimal objective region is obtained with computing.
CN201510753554.3A 2015-11-06 2015-11-06 A kind of infrared target detection method split based on gray scale and gradient Active CN105427286B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510753554.3A CN105427286B (en) 2015-11-06 2015-11-06 A kind of infrared target detection method split based on gray scale and gradient

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510753554.3A CN105427286B (en) 2015-11-06 2015-11-06 A kind of infrared target detection method split based on gray scale and gradient

Publications (2)

Publication Number Publication Date
CN105427286A true CN105427286A (en) 2016-03-23
CN105427286B CN105427286B (en) 2017-12-05

Family

ID=55505467

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510753554.3A Active CN105427286B (en) 2015-11-06 2015-11-06 A kind of infrared target detection method split based on gray scale and gradient

Country Status (1)

Country Link
CN (1) CN105427286B (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106251362A (en) * 2016-07-15 2016-12-21 中国电子科技集团公司第二十八研究所 A kind of sliding window method for tracking target based on fast correlation neighborhood characteristics point and system
CN107392885A (en) * 2017-06-08 2017-11-24 江苏科技大学 A kind of method for detecting infrared puniness target of view-based access control model contrast mechanism
CN107633505A (en) * 2017-08-24 2018-01-26 南京理工大学 A kind of undercarriage detection method based on target gray distribution character
CN108898573A (en) * 2018-04-23 2018-11-27 西安电子科技大学 Infrared small target rapid extracting method based on multi-direction annular gradient method
CN111914686A (en) * 2020-07-15 2020-11-10 云南电网有限责任公司带电作业分公司 SAR remote sensing image water area extraction method, device and system based on surrounding area association and pattern recognition
CN111931785A (en) * 2020-06-19 2020-11-13 国网山西省电力公司吕梁供电公司 Edge detection method for infrared image target of power equipment
CN113269800A (en) * 2021-05-25 2021-08-17 华能海南发电股份有限公司澄迈光伏电站 Infrared image segmentation method for solar photovoltaic module
CN116503407A (en) * 2023-06-28 2023-07-28 无锡日联科技股份有限公司 Method and device for detecting foreign object region in image and electronic equipment

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104268877A (en) * 2014-09-26 2015-01-07 中国电子科技集团公司第二十八研究所 Infrared image sea-sky-line self adaption detection method
CN104408432A (en) * 2014-11-28 2015-03-11 中国电子科技集团公司第二十八研究所 Infrared image target detection method based on histogram modification

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104268877A (en) * 2014-09-26 2015-01-07 中国电子科技集团公司第二十八研究所 Infrared image sea-sky-line self adaption detection method
CN104408432A (en) * 2014-11-28 2015-03-11 中国电子科技集团公司第二十八研究所 Infrared image target detection method based on histogram modification

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
SHUNYONG ZHOU 等: "Infrared Image Segmentation Based on Otsu and Genetic Algorithm", 《2011 CONFERENCE ON MULTIMEDIA TECHNOLOGY》 *
白俊奇 等: "基于局部梯度特征的红外微扫描成像技术研究", 《光子学报》 *
翟尚礼 等: "红外搜索跟踪系统的关键技术和解决途径", 《指挥信息系统与技术》 *
许开宇 等: "一种基于Canny算子的船舶目标检测算法", 《红外技术》 *
郭慧鑫 等: "一种基于Canny算子的红外图像边缘检测算法", 《火力与指挥控制》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106251362A (en) * 2016-07-15 2016-12-21 中国电子科技集团公司第二十八研究所 A kind of sliding window method for tracking target based on fast correlation neighborhood characteristics point and system
CN106251362B (en) * 2016-07-15 2019-02-01 南京莱斯电子设备有限公司 A kind of sliding window method for tracking target and system based on fast correlation neighborhood characteristics point
CN107392885A (en) * 2017-06-08 2017-11-24 江苏科技大学 A kind of method for detecting infrared puniness target of view-based access control model contrast mechanism
CN107633505A (en) * 2017-08-24 2018-01-26 南京理工大学 A kind of undercarriage detection method based on target gray distribution character
CN108898573A (en) * 2018-04-23 2018-11-27 西安电子科技大学 Infrared small target rapid extracting method based on multi-direction annular gradient method
CN108898573B (en) * 2018-04-23 2021-11-02 西安电子科技大学 Infrared small target rapid extraction method based on multidirectional annular gradient method
CN111931785A (en) * 2020-06-19 2020-11-13 国网山西省电力公司吕梁供电公司 Edge detection method for infrared image target of power equipment
CN111914686A (en) * 2020-07-15 2020-11-10 云南电网有限责任公司带电作业分公司 SAR remote sensing image water area extraction method, device and system based on surrounding area association and pattern recognition
CN111914686B (en) * 2020-07-15 2022-10-18 云南电网有限责任公司带电作业分公司 SAR remote sensing image water area extraction method, device and system based on surrounding area association and pattern recognition
CN113269800A (en) * 2021-05-25 2021-08-17 华能海南发电股份有限公司澄迈光伏电站 Infrared image segmentation method for solar photovoltaic module
CN116503407A (en) * 2023-06-28 2023-07-28 无锡日联科技股份有限公司 Method and device for detecting foreign object region in image and electronic equipment
CN116503407B (en) * 2023-06-28 2023-10-20 无锡日联科技股份有限公司 Method and device for detecting foreign object region in image and electronic equipment

Also Published As

Publication number Publication date
CN105427286B (en) 2017-12-05

Similar Documents

Publication Publication Date Title
CN105427286A (en) Gray scale and gradient segmentation-based infrared target detection method
WO2021012757A1 (en) Real-time target detection and tracking method based on panoramic multichannel 4k video images
CN102855622B (en) A kind of infrared remote sensing image sea ship detection method based on significance analysis
CN104268872B (en) Consistency-based edge detection method
CN105574855B (en) Infrared small target detection method under cloud background based on template convolution and false alarm rejection
CN103927520A (en) Method for detecting human face under backlighting environment
CN104301669A (en) Suspicious target detection tracking and recognition method based on dual-camera cooperation
CN107392885A (en) A kind of method for detecting infrared puniness target of view-based access control model contrast mechanism
CN101916448A (en) Moving object detecting method based on Bayesian frame and LBP (Local Binary Pattern)
CN104866843B (en) A kind of masked method for detecting human face towards monitor video
CN105404874B (en) A kind of vehicle window identifying system based on projection and hough straight-line detections
CN102663362B (en) Moving target detection method based on gray features
CN102799862A (en) System and method for pedestrian rapid positioning and event detection based on high definition video monitor image
CN104598929A (en) HOG (Histograms of Oriented Gradients) type quick feature extracting method
CN111353496B (en) Real-time detection method for infrared dim targets
CN105279772A (en) Trackability distinguishing method of infrared sequence image
CN109598706A (en) A kind of camera lens occlusion detection method and system
CN104182992A (en) Method for detecting small targets on the sea on the basis of panoramic vision
Xia et al. Automatic multi-vehicle tracking using video cameras: An improved CAMShift approach
CN103607558A (en) Video monitoring system, target matching method and apparatus thereof
CN104408432B (en) Infrared image target detection method based on histogram modification
CN101877135B (en) Moving target detecting method based on background reconstruction
CN104866844A (en) Crowd gathering detection method for monitor video
CN109919068B (en) Real-time monitoring method for adapting to crowd flow in dense scene based on video analysis
CN103578121A (en) Motion detection method based on shared Gaussian model in disturbed motion environment

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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20181024

Address after: 210007 the 5 building of Tianan Digital City, 36 Yongfeng Avenue, Qinhuai District, Nanjing, Jiangsu.

Patentee after: Nanjing Lesi Electronic Equipment Co., Ltd.

Address before: 210007 1 East Street, alfalfa garden, Nanjing, Jiangsu

Patentee before: China Electronic Technology Corporation (Group) 28 Institute