CN104766079A - Remote infrared weak object detecting method - Google Patents
Remote infrared weak object detecting method Download PDFInfo
- Publication number
- CN104766079A CN104766079A CN201510220899.2A CN201510220899A CN104766079A CN 104766079 A CN104766079 A CN 104766079A CN 201510220899 A CN201510220899 A CN 201510220899A CN 104766079 A CN104766079 A CN 104766079A
- Authority
- CN
- China
- Prior art keywords
- target
- pixel
- background
- value
- image
- 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
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a remote infrared weak object detecting method. The method comprises the nine steps of median filtering, local maximum value searching, predicting of the local background of a candidate target region, predicated value correcting, residual obtaining through background suppression, residual mean value solving, self-adaptive threshold dividing, target clustering and false target rejecting. According to the target detecting algorithm based on the local background, by extracting the local maximum value point, the technologies of local background predicting, false target rejecting and the like are adopted, the computational burden is effectively reduced, the detecting probability of weak targets is increased, and the false alarm rate is reduced.
Description
Technical field
The present invention relates to a kind of remote method for detecting infrared puniness target, particularly relate to a kind of remote method for detecting infrared puniness target being applicable to photoelectric technology image processing field.
Background technology
Remote method for detecting infrared puniness target belongs to photoelectric technology image processing field.Small IR targets detection is a gordian technique in the electro-optical systems such as precise guidance, Search/Track, supervision and early warning.Long Range Moving Target signal to noise ratio (S/N ratio) is low, and target only accounts for the area of one or several pixel in the picture, and lacking shape and structural information, is the difficulties of photodetection field target detection.
At present for remote small IR targets detection mostly based on the detection algorithm of overall signal to noise ratio (S/N ratio), the method exists that operand is large, detection probability is low and the shortcoming such as false alarm rate is higher, is difficult to meet the requirement of Photodetection system to detection probability and false alarm rate.
Summary of the invention
It is little that the technical problem to be solved in the present invention is to provide a kind of operand, and target detection false alarm rate is low, the remote method for detecting infrared puniness target high to remote small IR targets detection probability and system.
For remote infrared small object, the feature possibility that target shows in entire image is also not obvious, but in topography, always show the feature of singular point.
The technical solution used in the present invention is as follows: a kind of remote method for detecting infrared puniness target, and concrete grammar step is:
Step one, collection target image;
Step 2, medium filtering is carried out to the image of target gathered;
Step 3, choose be of a size of 5*5 to 9*9 pixel wicket as search window, the position of window center is current pixel position, search for entire image successively, when window center pixel is the maximal value of window area, the alternatively target area, region that the ratio search window centered by search window center is young;
Step 4, the prediction of candidate target region local background;
Step 5, deduct the background forecast value in this region with the pixel value of original image candidate target region, obtain residual error;
Step 6, ask for residual error average, represent the projection degree of target relative to ambient background by the residual error average of candidate target region;
Step 7, self-adaptive filtering method, carry out self-adaptive filtering method to residual error average and obtain potential target;
Step 8, labeled clusters is carried out to potential target;
Step 9, carry out false target rejecting;
Step 10, target information export.
As preferably, in described step 2,5 medium filterings are adopted to eliminate salt-pepper noise and detector blind element.
As preferably, in described step 4, the concrete grammar of candidate target region local background prediction is: replace the intermediate value of its background of pixel of each pixel of candidate target region.
As preferably, the computing method of described background intermediate value are: the background intermediate value calculating this candidate target region pixel by each pixel neighboring background pixel pixel of candidate target region.
As preferably, method step between described step 4 and step 5 also comprises correction predicted value, and concrete grammar is: with the intermediate value of all predicted values in step 4 for thresholding, judges whether predicted value is less than threshold value, be replace predicted value by threshold value, otherwise keep predicted value constant.
As preferably, the concrete grammar of described correction predicted value also comprises: the pixel maximal value of all predicted values in center, target area is replaced.
As preferably, in described step 7, segmentation threshold is sued for peace by the intermediate value of all residual error averages and the mean square deviation of all residual error averages and is obtained.
As preferably, potential target pixel is obtained after labeled clusters in described step 8, after potential target pixel carries out linear nonlinear filtering to the image information after medium filtering, image before and after linear nonlinear filtering do difference obtain after background suppress image, then utilize the image after background suppress to reject false target, finally export target information.
Compared with prior art, the invention has the beneficial effects as follows: by extracting Local modulus maxima, adopting the technology such as local background's prediction and false target rejecting, effectively reducing operand, improve the detection probability to Weak target, reduce false alarm rate.
Accompanying drawing explanation
Fig. 1 is the remote small IR targets detection process flow diagram of the present invention's wherein embodiment.
Fig. 2 is the candidate target region schematic diagram of the present invention's wherein embodiment.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
Arbitrary feature disclosed in this instructions (comprising any accessory claim, summary and accompanying drawing), unless specifically stated otherwise, all can be replaced by other equivalences or the alternative features with similar object.That is, unless specifically stated otherwise, each feature is an example in a series of equivalence or similar characteristics.
As shown in Figure 1, a kind of remote method for detecting infrared puniness target, concrete grammar step is:
Step one, collection target image; Directly target image can be gathered, also directly target analysis detection can be carried out to the image stored.
Step 2, medium filtering is carried out to the image of target gathered; In this specific embodiment, 5 medium filterings are adopted to eliminate salt-pepper noise and detector blind element etc.
Step 3, choose be of a size of 5*5 to 9*9 pixel wicket as search window, the position of window center is current pixel position, search for entire image successively, when window center pixel is the maximal value of window area, the alternatively target area, region that the ratio search window centered by search window center is young; This step will search for the local maximum of image after medium filtering, and follow-up process is all have image information based on this local maximum and local cell thereof, no longer processes entire image, effectively reduces the operand of subsequent treatment.For the background around target, Small object gray scale comparatively ambient background wants high, and uncorrelated with background, is the isolated speck in image, energy distribution shows as the envelope of a protuberance, the envelope most likely target that therefore volume is larger.And the summit of envelope must be the local maximum of this target region.Because the pixel of Small object is few, in this specific embodiment, we choose be of a size of 5*5 wicket as search window, as shown in Figure 2, the position of window center is current pixel position, searches for entire image successively.When window center pixel is the maximal value in 5*5 region, the 3*3 region centered by search window center is as a candidate target region.
Step 4, the prediction of candidate target region local background; In classical background prediction methods, background forecast value is normally converted by combination of pixels all in its surrounding neighbors and obtains, such background pixel and object pixel have mixed, and they in fact and uncorrelated, so the background estimated like this is difficult to accurately, consider that the non-correlation of background and target carries out background forecast and can dope background more accurately.In this specific embodiment, the concrete grammar of candidate target region local background prediction is: replace the intermediate value of its background of pixel of each pixel of candidate target region.And the computing method of described background intermediate value are: the background intermediate value calculating this candidate target region pixel by each pixel neighboring background pixel pixel of candidate target region.Select intermediate value to make this algorithm have certain noise removal capability, and only estimated to obtain by its neighboring background pixel due to the pixel in candidate target region, make predicted value comparatively accurate.
In this specific embodiment, also comprise correction predicted value, concrete grammar is: with the intermediate value of all predicted values in step 4 for thresholding, judges whether predicted value is less than threshold value, is, replaces predicted value by threshold value, otherwise keeps predicted value constant.Due to the complicacy of infrared imagery background, candidate target region probably drops on the intersection of bright dark two kinds of background areas, thus it is comparatively far away to make the background pixel value estimated depart from actual value, based on above-mentioned consideration, after estimating background pixel value, also to do further correction to it.The concrete grammar of described correction predicted value also comprises: the pixel maximal value of all predicted values in center, target area is replaced.This algorithm has well taken into account the situation of background acute variation.
Step 5, deduct the background forecast value in this region with the pixel value of original image candidate target region, obtain residual error.
Step 6, ask for residual error average, represent the projection degree of target relative to ambient background by the residual error average of candidate target region.
Step 7, self-adaptive filtering method, carry out self-adaptive filtering method to residual error average and obtain potential target; In this specific embodiment, segmentation threshold is sued for peace by the intermediate value of all residual error averages and the mean square deviation of all residual error averages and is obtained.
Step 8, labeled clusters is carried out to potential target; Potential target pixel under normal circumstances after thresholding segmentation is less, and this step operation amount is little.
Step 9, carry out false target rejecting;
Step 10, target information export.
In this specific embodiment, potential target pixel is obtained after labeled clusters in described step 8, after potential target pixel carries out linear nonlinear filtering (being 7*7 pixel filter in this specific embodiment) to the image information after medium filtering, image before and after linear nonlinear filtering do difference obtain after background suppress image, play the effect strengthening target, then utilize the image after background suppress to reject false target, finally export target information.Remote Weak target meets point spread function, and local gray level distributional class is similar to the distribution of Gauss's circle, if do not meet the distribution of Gauss's circle in the image after the target gray after cluster is distributed in background suppress be namely judged to false target, is rejected.
Claims (8)
1. a remote method for detecting infrared puniness target, concrete grammar step is:
Step one, collection target image;
Step 2, medium filtering is carried out to the image of target gathered;
Step 3, choose be of a size of 5*5 to 9*9 pixel wicket as search window, the position of window center is current pixel position, search for entire image successively, when window center pixel is the maximal value of window area, the alternatively target area, region that the ratio search window centered by search window center is young;
Step 4, the prediction of candidate target region local background;
Step 5, deduct the background forecast value in this region with the pixel value of original image candidate target region, obtain residual error;
Step 6, ask for residual error average, represent the projection degree of target relative to ambient background by the residual error average of candidate target region;
Step 7, self-adaptive filtering method, carry out self-adaptive filtering method to residual error average and obtain potential target;
Step 8, labeled clusters is carried out to potential target;
Step 9, carry out false target rejecting;
Step 10, target information export.
2. remote method for detecting infrared puniness target according to claim 1, in described step 2, adopts 5 medium filterings to eliminate salt-pepper noise and detector blind element.
3. remote method for detecting infrared puniness target according to claim 2, in described step 4, the concrete grammar of candidate target region local background prediction is: replace the intermediate value of its background of pixel of each pixel of candidate target region.
4. remote method for detecting infrared puniness target according to claim 3, the computing method of described background intermediate value are: the background intermediate value calculating this candidate target region pixel by each pixel neighboring background pixel pixel of candidate target region.
5. according to the remote method for detecting infrared puniness target one of claim 1 to 4 Suo Shu, method step between described step 4 and step 5 also comprises correction predicted value, concrete grammar is: with the intermediate value of all predicted values in step 4 for thresholding, judge whether predicted value is less than threshold value, be replace predicted value by threshold value, otherwise keep predicted value constant.
6. remote method for detecting infrared puniness target according to claim 5, the concrete grammar of described correction predicted value also comprises: the pixel maximal value of all predicted values in center, target area is replaced.
7. remote method for detecting infrared puniness target according to claim 1, in described step 7, segmentation threshold is sued for peace by the intermediate value of all residual error averages and the mean square deviation of all residual error averages and is obtained.
8. the remote method for detecting infrared puniness target according to claim 1,2,3,4,6 or 7, potential target pixel is obtained after labeled clusters in described step 8, after potential target pixel carries out linear nonlinear filtering to the image information after medium filtering, image before and after linear nonlinear filtering do difference obtain after background suppress image, then utilize the image after background suppress to reject false target, finally export target information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510220899.2A CN104766079B (en) | 2015-05-05 | 2015-05-05 | A kind of remote method for detecting infrared puniness target |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510220899.2A CN104766079B (en) | 2015-05-05 | 2015-05-05 | A kind of remote method for detecting infrared puniness target |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104766079A true CN104766079A (en) | 2015-07-08 |
| CN104766079B CN104766079B (en) | 2018-12-07 |
Family
ID=53647893
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510220899.2A Active CN104766079B (en) | 2015-05-05 | 2015-05-05 | A kind of remote method for detecting infrared puniness target |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104766079B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105976403A (en) * | 2016-07-25 | 2016-09-28 | 中国电子科技集团公司第二十八研究所 | Infrared imaging target tracking method based on kernel function centroid drifting |
| CN106709497A (en) * | 2016-11-16 | 2017-05-24 | 北京理工大学 | PCNN-based infrared motion weak target detection method |
| CN108492320A (en) * | 2018-03-14 | 2018-09-04 | 四川长九光电科技有限责任公司 | A kind of method for detecting infrared puniness target based on parallel processing |
| CN109948623A (en) * | 2019-03-04 | 2019-06-28 | 北京空间飞行器总体设计部 | A kind of space Weak target Relative Navigation image binaryzation image generating method |
| CN111353496A (en) * | 2018-12-20 | 2020-06-30 | 中国科学院沈阳自动化研究所 | Real-time detection method for infrared small and weak target |
| CN112541486A (en) * | 2020-12-31 | 2021-03-23 | 洛阳伟信电子科技有限公司 | Infrared weak and small target detection algorithm based on improved Pixel segmentation |
| CN116736256A (en) * | 2023-08-11 | 2023-09-12 | 南京隼眼电子科技有限公司 | Radar identification method and device and electronic equipment |
| CN117115128A (en) * | 2023-09-11 | 2023-11-24 | 杭州深度视觉科技有限公司 | Image pixel value calculating method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005069197A1 (en) * | 2003-12-31 | 2005-07-28 | Lockheed Martin Missiles & Fire Control | A method and system for adaptive target detection |
| CN102496016A (en) * | 2011-11-22 | 2012-06-13 | 武汉大学 | Infrared target detection method based on space-time cooperation framework |
| CN102819740A (en) * | 2012-07-18 | 2012-12-12 | 西北工业大学 | Method for detecting and positioning dim targets of single-frame infrared image |
-
2015
- 2015-05-05 CN CN201510220899.2A patent/CN104766079B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005069197A1 (en) * | 2003-12-31 | 2005-07-28 | Lockheed Martin Missiles & Fire Control | A method and system for adaptive target detection |
| CN102496016A (en) * | 2011-11-22 | 2012-06-13 | 武汉大学 | Infrared target detection method based on space-time cooperation framework |
| CN102819740A (en) * | 2012-07-18 | 2012-12-12 | 西北工业大学 | Method for detecting and positioning dim targets of single-frame infrared image |
Non-Patent Citations (2)
| Title |
|---|
| 侯旺 等: "红外弱小目标检测技术研究现状与发展趋势", 《红外技术》 * |
| 俞志刚: ""红外弱小目标检测算法研究"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105976403B (en) * | 2016-07-25 | 2018-09-21 | 中国电子科技集团公司第二十八研究所 | A kind of IR imaging target tracking method based on the drift of kernel function barycenter |
| CN105976403A (en) * | 2016-07-25 | 2016-09-28 | 中国电子科技集团公司第二十八研究所 | Infrared imaging target tracking method based on kernel function centroid drifting |
| CN106709497A (en) * | 2016-11-16 | 2017-05-24 | 北京理工大学 | PCNN-based infrared motion weak target detection method |
| CN108492320B (en) * | 2018-03-14 | 2022-04-12 | 四川长九光电科技有限责任公司 | Infrared dim target detection method based on parallel processing |
| CN108492320A (en) * | 2018-03-14 | 2018-09-04 | 四川长九光电科技有限责任公司 | A kind of method for detecting infrared puniness target based on parallel processing |
| CN111353496B (en) * | 2018-12-20 | 2023-07-28 | 中国科学院沈阳自动化研究所 | Real-time detection method for infrared dim targets |
| CN111353496A (en) * | 2018-12-20 | 2020-06-30 | 中国科学院沈阳自动化研究所 | Real-time detection method for infrared small and weak target |
| CN109948623A (en) * | 2019-03-04 | 2019-06-28 | 北京空间飞行器总体设计部 | A kind of space Weak target Relative Navigation image binaryzation image generating method |
| CN112541486A (en) * | 2020-12-31 | 2021-03-23 | 洛阳伟信电子科技有限公司 | Infrared weak and small target detection algorithm based on improved Pixel segmentation |
| CN112541486B (en) * | 2020-12-31 | 2022-11-08 | 洛阳伟信电子科技有限公司 | Infrared weak and small target detection algorithm based on improved Pixel segmentation |
| CN116736256A (en) * | 2023-08-11 | 2023-09-12 | 南京隼眼电子科技有限公司 | Radar identification method and device and electronic equipment |
| CN116736256B (en) * | 2023-08-11 | 2023-10-24 | 南京隼眼电子科技有限公司 | Radar identification method and device and electronic equipment |
| CN117115128A (en) * | 2023-09-11 | 2023-11-24 | 杭州深度视觉科技有限公司 | Image pixel value calculating method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104766079B (en) | 2018-12-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104766079A (en) | Remote infrared weak object detecting method | |
| CN102063614B (en) | Method and device for detecting lost articles in security monitoring | |
| US10127448B2 (en) | Method and system for dismount detection in low-resolution UAV imagery | |
| Sengar et al. | Moving object area detection using normalized self adaptive optical flow | |
| CN103246896B (en) | A kind of real-time detection and tracking method of robustness vehicle | |
| CN103077539A (en) | Moving object tracking method under complicated background and sheltering condition | |
| CN102663362B (en) | Moving target detection method based on gray features | |
| CN103971386A (en) | Method for foreground detection in dynamic background scenario | |
| CN103826102A (en) | Recognition method and device of moving object | |
| CN103456009B (en) | Object detection method and device, supervisory system | |
| CN102332167A (en) | Target detection method for vehicles and pedestrians in intelligent traffic monitoring | |
| CN106778570B (en) | A kind of pedestrian detection and tracking in real time | |
| CN108804987B (en) | Door opening and closing state detection method and device and people flow detection system | |
| CN102799862A (en) | System and method for pedestrian rapid positioning and event detection based on high definition video monitor image | |
| CN109711256B (en) | Low-altitude complex background unmanned aerial vehicle target detection method | |
| Li et al. | Fast pedestrian detection and dynamic tracking for intelligent vehicles within V2V cooperative environment | |
| CN111191535B (en) | Pedestrian detection model construction method based on deep learning and pedestrian detection method | |
| Lian et al. | A novel method on moving-objects detection based on background subtraction and three frames differencing | |
| CN111353496B (en) | Real-time detection method for infrared dim targets | |
| CN105427286A (en) | Gray scale and gradient segmentation-based infrared target detection method | |
| CN103793715A (en) | Underground worker target tracing method based on scene information mining | |
| CN103646242B (en) | Extended target tracking based on maximum stable extremal region feature | |
| Ying-hong et al. | An improved Gaussian mixture background model with real-time adjustment of learning rate | |
| CN105469054A (en) | Model construction method of normal behaviors and detection method of abnormal behaviors | |
| CN107729811B (en) | Night flame detection method based on scene modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |