CN109712158A - A kind of infrared small target catching method based on target background pixel statistical restraint - Google Patents
A kind of infrared small target catching method based on target background pixel statistical restraint Download PDFInfo
- Publication number
- CN109712158A CN109712158A CN201811408910.8A CN201811408910A CN109712158A CN 109712158 A CN109712158 A CN 109712158A CN 201811408910 A CN201811408910 A CN 201811408910A CN 109712158 A CN109712158 A CN 109712158A
- Authority
- CN
- China
- Prior art keywords
- image
- target
- infrared
- background
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 230000000877 morphologic effect Effects 0.000 claims abstract description 10
- 238000005260 corrosion Methods 0.000 claims abstract description 5
- 230000007797 corrosion Effects 0.000 claims abstract description 5
- 230000003628 erosive effect Effects 0.000 claims abstract description 5
- 230000010339 dilation Effects 0.000 claims abstract description 4
- 238000001514 detection method Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000005764 inhibitory process Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 230000003044 adaptive effect Effects 0.000 abstract description 9
- 238000012216 screening Methods 0.000 abstract description 5
- 230000008859 change Effects 0.000 abstract description 2
- 230000011218 segmentation Effects 0.000 description 9
- 239000000284 extract Substances 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000003331 infrared imaging Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002902 bimodal effect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of infrared small target catching methods based on target background pixel statistical restraint, collected infrared image is subjected to morphological erosion and morphological dilations first, eliminate the point target and isolated noise point in image, the image with original image image subtraction after corrosion and expansion can obtain the point target of high brightness in this sub-picture, isolated noise point and the similar point target of structure size and isolated noise point and a small number of high gradients corner again, to achieve the purpose that inhibit image background；Candidate target and some isolated noises are contained in the image inhibited by background, target information is extracted by the method that adaptive threshold fuzziness is combined with connected domain screening；Adaptive threshold calculates gray average, variance, the maximum statistical informations such as gray scale and minimal gray to every width figure, it calculates different threshold values to be split, therefore with the variation of image, threshold value also corresponding change therewith, target information accurately is extracted, the accuracy of capture can be improved.
Description
Technical field
The invention belongs to infrared early warning technical fields, and in particular to a kind of based on the infrared of target background pixel statistical restraint
Small object catching method.
Background technique
The performance of infrared small target acquisition algorithm determines the detectivity and detection range of infrared imaging system of defense,
Detectivity and detection range are the reflection vital core index of Observable infrared small target recognition capability.Due to
Small object pixel ratio shared in infrared image is very low, and Small object is usually flooded by background and noise, it is difficult to it detects,
It is therefore desirable to pre-process to such image then further extracting target by way of Threshold segmentation.Common threshold value
Dividing method has artificial experience back-and-forth method, utilizes histogram analysis method, maximum variance between clusters etc..Artificial experience back-and-forth method needs
The priori knowledge of the image handled as needed analyzes the target in image with background, and efficiency is lower and can not achieve
Automatic threshold value is chosen, the situation less suitable for picture number.Histogram analysis method is according to the pass between peaks and troughs
System choose a threshold value appropriate, but only for there is a situation where one target and a background and the two compare it is significant ability it is effective
Fruit；Maximum variance between clusters are to divide the image into foreground and background two parts according to image grayscale characteristic, when threshold value selection is proper
When, two-part difference should be the largest.This method is the unimodal preferable segmentation effect of image generation to inter-class variance, but
When the size great disparity of target and background (such as being illuminated by the light the factors such as uneven, reflective or background is complicated influence), side between class
Bimodal or multimodal may be presented in poor criterion function, and effect is bad at this time.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of infrared small targets based on target background pixel statistical restraint to catch
Method is obtained, can identify the simultaneously farther away non-cooperation aircraft of tracking range, reserves the enough reaction time as far as possible for air defence system.
A kind of infrared small target catching method based on target background pixel statistical restraint, includes the following steps:
Step 1: for every piece image of input, carrying out the operation of morphological erosion first, then carry out morphology again
The operation of expansion；Image of the original image image subtraction after corrosion and expansion is finally used again, is achieved in infrared image background inhibition；
Step 2: the every piece image completed for step 1 processing, by ti=Mean (f)+λiStd (f) is used as threshold value,
Binary conversion treatment is carried out to image；Wherein,Max (f) is image maximum gradation value,
Mean (f) is image grayscale mean value, and Std (f) is image grayscale standard deviation；
Step 3 carries out connected region detection to the binary image handled by step 2, to extract real time
Select target.
Preferably,
The invention has the following beneficial effects:
Infrared small target is usually expressed as dotted or light spot-like in the picture, lacks the features such as textural shape, so being not easy
It is detected.Therefore the present invention is a kind of based on the adaptive of image statistical information itself for proposing for this kind of Small object
The method of Threshold segmentation passes through average gray, maximum gradation value and ash using the grey-level statistics of image itself as foundation
Weight is calculated based on degree standard deviation to obtain adaptivenon-uniform sampling threshold value, to improve the detectability of infrared small target.
Collected infrared image is subjected to morphological erosion and morphological dilations first, eliminates point target and isolated noise in image
Point, then the point target of high brightness, orphan in this sub-picture can be obtained with image of the original image image subtraction after corrosion and expansion
Vertical noise spot and the similar point target of structure size and isolated noise point and a small number of high gradients corner, carry on the back image to reach
The purpose that scape is inhibited；Candidate target and some isolated noises are contained in the image inhibited by background, by certainly
It adapts to the method that Threshold segmentation is combined with connected domain screening and extracts target information.It is equal that adaptive threshold calculates gray scale to every width figure
Value, variance, the maximum statistical informations such as gray scale and minimal gray, calculate different threshold values and are split, therefore with image
Variation, threshold value also corresponding change therewith, accurately extracts target information, the accuracy of capture can be improved.
Detailed description of the invention
Fig. 1 is a kind of infrared small target catching method flow chart based on target background pixel statistical restraint of the invention；
Fig. 2 is the Top-Hat top cap algorithm flow chart that the present invention uses；
Fig. 3 is the adaptive threshold fuzziness flow chart used in the present invention；
Fig. 4 is that the candidate target used in the present invention extracts flow chart.
Specific embodiment
The present invention will now be described in detail with reference to the accompanying drawings and examples.
The basis of Method of Target Detection in Infrared is analyzed target, background and the noise in infrared image.This hair
Bright studied object is infrared small target, and infrared small target usually shows as dotted or light spot-like in the picture, lacks shape
Texture eigenvalue；The background for the infrared small target that the present invention is studied is mainly low latitude or high-altitude background, may there is cloud
The chaff interferents such as layer；Noise refers to that system is intrinsic, by all parts of infrared imaging, such as optical system generation.
One width infrared small target image can simply be described are as follows:
F (x, y)=fT(x,y)+fB(x,y)+n(x,y)
Wherein, f (x, y) is original infrared image, fT(x, y) is target, fB(x, y) is background, and n (x, y) is noise.
In order to detect infrared small target, ambient noise bring interference in image is removed, needs to carry out infrared small target image pre-
Processing operation.Due to transform domain filter and method based on mankind's attention mechanism that there are calculation amounts is excessive, hardware realization is difficult
The shortcomings that, therefore the preprocess method that can be used has calculation amount smaller, realize simple Butterworth high-pass filtering and
Morphologic filtering algorithm.
Overall flow figure carries out background inhibition as shown in Figure 1, passing through Top-Hat algorithm first to acquired image, so
Binary conversion treatment is carried out by adaptive threshold fuzziness again afterwards, filters out real candidate target finally by connected domain area.
Step 1: infrared image background inhibits
Under normal circumstances, original image formed by imaging system is due to by external factor such as sensor itself and environment
Influence will lead to it and cannot directly be used in vision system, especially infrared small target, it is easy to by ambient noise
It influences, therefore, needs to carry out original image certain pretreatment in preliminary stage, such as inhibit background to inhibit, enhancing contrast
Deng, the pretreatment of infrared small target image is mainly inhibited with background based on.Infrared image background suppressing method in the present invention
Main that the Top-Hat top cap algorithm based on mathematical morphology is used to realize, flow chart is as shown in Figure 1.Come in for input
Image, first carry out morphological erosion operation, then carry out the operation of morphological dilations again, thus eliminate will be in image
Point target and isolated noise point, finally the image again with original image image subtraction after corrosion and expansion can obtain this sub-picture
Point target, isolated noise point and the similar point target of structure size and isolated noise point of middle high brightness and a small number of high gradients side
Angle, to achieve the purpose that inhibit image background.
Step 2: candidate target extracts
Candidate target and some isolated noises are contained in the image inhibited by background, so needing to real
Candidate target extracts, and the present invention extracts candidate mesh using the method that adaptive threshold fuzziness is combined with connected domain screening
Mark.Adaptive threshold method generally first calculates gray average, variance, the maximum statistics such as the gray scale and minimal gray letter of image
Breath, then uses the combination of certain several statistical information therein as threshold value, and each statistical information has a weight, general each weight
The sum of be equal to 1.Since target background is fairly simple, in the image after pretreatment, target distribution is and low in high-frequency range
Frequency range is mainly noise.Therefore, the present invention believes using image grayscale mean value, maximum gradation value and gray standard deviation as statistics
Breath, by certain operation between three so that it is determined that weight λi, and then carry out adaptive threshold.Whole process is as follows:
If Max (f) is the maximum gradation value in f, Mean (f) is f gray average, and Std (f) is f gray standard deviation.Take ti
=Mean (f)+λiStd (f), whereinIt is obtained by test of many times, whenWhen, target can be detected well, and without iterative calculation, therefore it can be concluded that threshold value:
The ti of every piece image is different, and then realizes the binaryzation to image using following formula.
Its whole flow chart is as shown in Figure 3.The gray value of the image of input is counted first, first is that in order to pick out
Maximum gradation value in image, second is that power can be calculated by then substituting into empirical equation in order to calculate the gray average of image
Weight, and then obtain the threshold value of adaptivenon-uniform sampling.
In order to from the image after Threshold segmentation, obtain the coordinate position of target, then carry out connected region detection.For two
It is worth for image, connected region generally refers to the region of the point composition in adjacent position with same pixel value.It will be in image
Each connected region is found out and referred to as connected component analysis is marked.It may include multiple to the picture that Threshold segmentation obtains
Region carries out connected domain screening, filters out the connected domain of area coinciding requirement, seeks to the target found.Check each pixel with
The connectivity of its adjacent pixel is the simple and effective method in each region in image after marking progress Threshold segmentation.In binary map
As in, the value of background area pixel is 0, and the pixel value of target area is 1.To Threshold segmentation image, connected domain detection is carried out, then is needed
Image is begun stepping through from the upper left corner of image, and mark the connectivity between current pixel and traversed pixel.Its
Overall flow figure is as shown in figure 4, carry out connected domain screening to the image after Threshold segmentation, to extract real candidate
Target.
By this algorithm in embedded platform --- it is verified on C6678 multi-core DSP hardware, host computer will be schemed by network interface
It is handled as data are loaded to DSP, by carrying out test verifying to image data under a variety of scenes, obtained test result is such as
Shown in table 1, the target acquistion accuracy of this algorithm can analyze out up to 92% or more.
1 experimental result of table
In conclusion the above is merely preferred embodiments of the present invention, being not intended to limit the scope of the present invention.
All within the spirits and principles of the present invention, any modification, equivalent replacement, improvement and so on should be included in of the invention
Within protection scope.
Claims (2)
1. a kind of infrared small target catching method based on target background pixel statistical restraint, which is characterized in that including walking as follows
It is rapid:
Step 1: for every piece image of input, carrying out the operation of morphological erosion first, then carry out morphological dilations again
Operation；Image of the original image image subtraction after corrosion and expansion is finally used again, is achieved in infrared image background inhibition；
Step 2: the every piece image completed for step 1 processing, by ti=Mean (f)+λiStd (f) is used as threshold value, to figure
As carrying out binary conversion treatment；Wherein,Max (f) is image maximum gradation value, Mean (f)
For image grayscale mean value, Std (f) is image grayscale standard deviation；
Step 3 carries out connected region detection to the binary image handled by step 2, to extract really candidate mesh
Mark.
2. a kind of infrared small target catching method based on target background pixel statistical restraint as described in claim 1, special
Sign is,
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811408910.8A CN109712158A (en) | 2018-11-23 | 2018-11-23 | A kind of infrared small target catching method based on target background pixel statistical restraint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811408910.8A CN109712158A (en) | 2018-11-23 | 2018-11-23 | A kind of infrared small target catching method based on target background pixel statistical restraint |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109712158A true CN109712158A (en) | 2019-05-03 |
Family
ID=66254336
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811408910.8A Pending CN109712158A (en) | 2018-11-23 | 2018-11-23 | A kind of infrared small target catching method based on target background pixel statistical restraint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109712158A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276788A (en) * | 2019-06-12 | 2019-09-24 | 北京轩宇空间科技有限公司 | Method and apparatus for infrared imaging formula target seeker target following |
| CN110660065A (en) * | 2019-09-29 | 2020-01-07 | 云南电网有限责任公司电力科学研究院 | Infrared fault detection and identification algorithm |
| CN110738106A (en) * | 2019-09-05 | 2020-01-31 | 天津大学 | optical remote sensing image ship detection method based on FPGA |
| CN110765912A (en) * | 2019-10-15 | 2020-02-07 | 武汉大学 | SAR image ship target detection method based on statistical constraint and Mask R-CNN |
| CN110765631A (en) * | 2019-10-31 | 2020-02-07 | 中国人民解放军95859部队 | Effective imaging pixel-based small target judgment method for infrared radiation characteristic measurement |
| CN111242062A (en) * | 2020-01-17 | 2020-06-05 | 于兴虎 | Image processing method and system for detecting skin position of zebra fish juvenile fish |
| CN111833319A (en) * | 2020-07-02 | 2020-10-27 | 南京工程学院 | Automatic detection method for hot spots of retired power lithium battery based on infrared video |
| CN113160076A (en) * | 2021-04-06 | 2021-07-23 | 中航航空电子有限公司 | Ground object infrared target acquisition method based on target edge neighborhood information |
| CN113362319A (en) * | 2021-06-30 | 2021-09-07 | 深圳市创想三维科技股份有限公司 | Laser printing method and device based on image processing, laser printer and computer readable storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268844A (en) * | 2014-10-17 | 2015-01-07 | 中国科学院武汉物理与数学研究所 | Small target infrared image processing method based on weighing local image entropy |
| CN106327522A (en) * | 2016-08-25 | 2017-01-11 | 上海航天控制技术研究所 | Infrared small target detection method based on multi-direction morphological filtering complex cloud background |
| CN107505546A (en) * | 2017-08-25 | 2017-12-22 | 国家电网公司 | A kind of method that corona discharge is monitored using ultraviolet imager |
| CN107563370A (en) * | 2017-07-07 | 2018-01-09 | 西北工业大学 | Visual attention mechanism-based marine infrared target detection method |
-
2018
- 2018-11-23 CN CN201811408910.8A patent/CN109712158A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268844A (en) * | 2014-10-17 | 2015-01-07 | 中国科学院武汉物理与数学研究所 | Small target infrared image processing method based on weighing local image entropy |
| CN106327522A (en) * | 2016-08-25 | 2017-01-11 | 上海航天控制技术研究所 | Infrared small target detection method based on multi-direction morphological filtering complex cloud background |
| CN107563370A (en) * | 2017-07-07 | 2018-01-09 | 西北工业大学 | Visual attention mechanism-based marine infrared target detection method |
| CN107505546A (en) * | 2017-08-25 | 2017-12-22 | 国家电网公司 | A kind of method that corona discharge is monitored using ultraviolet imager |
Non-Patent Citations (1)
| Title |
|---|
| 张毅刚 等: ""基于形态学Top-hat滤波的红外小目标检测"", 《计算机测量与控制》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276788A (en) * | 2019-06-12 | 2019-09-24 | 北京轩宇空间科技有限公司 | Method and apparatus for infrared imaging formula target seeker target following |
| CN110738106A (en) * | 2019-09-05 | 2020-01-31 | 天津大学 | optical remote sensing image ship detection method based on FPGA |
| CN110660065A (en) * | 2019-09-29 | 2020-01-07 | 云南电网有限责任公司电力科学研究院 | Infrared fault detection and identification algorithm |
| CN110660065B (en) * | 2019-09-29 | 2023-10-20 | 云南电网有限责任公司电力科学研究院 | Infrared fault detection and identification algorithm |
| CN110765912B (en) * | 2019-10-15 | 2022-08-05 | 武汉大学 | SAR image ship target detection method based on statistical constraint and Mask R-CNN |
| CN110765912A (en) * | 2019-10-15 | 2020-02-07 | 武汉大学 | SAR image ship target detection method based on statistical constraint and Mask R-CNN |
| CN110765631A (en) * | 2019-10-31 | 2020-02-07 | 中国人民解放军95859部队 | Effective imaging pixel-based small target judgment method for infrared radiation characteristic measurement |
| CN110765631B (en) * | 2019-10-31 | 2023-03-14 | 中国人民解放军95859部队 | Effective imaging pixel-based small target judgment method for infrared radiation characteristic measurement |
| CN111242062A (en) * | 2020-01-17 | 2020-06-05 | 于兴虎 | Image processing method and system for detecting skin position of zebra fish juvenile fish |
| CN111242062B (en) * | 2020-01-17 | 2023-05-09 | 于兴虎 | Image processing method and system for detecting skin position of juvenile zebra fish |
| CN111833319A (en) * | 2020-07-02 | 2020-10-27 | 南京工程学院 | Automatic detection method for hot spots of retired power lithium battery based on infrared video |
| CN113160076A (en) * | 2021-04-06 | 2021-07-23 | 中航航空电子有限公司 | Ground object infrared target acquisition method based on target edge neighborhood information |
| CN113362319A (en) * | 2021-06-30 | 2021-09-07 | 深圳市创想三维科技股份有限公司 | Laser printing method and device based on image processing, laser printer and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109712158A (en) | A kind of infrared small target catching method based on target background pixel statistical restraint | |
| CN108038476B (en) | A kind of facial expression recognition feature extracting method based on edge detection and SIFT | |
| CN106856002B (en) | Unmanned aerial vehicle shooting image quality evaluation method | |
| CN102214298B (en) | Method for detecting and identifying airport target by using remote sensing image based on selective visual attention mechanism | |
| CN109948566B (en) | Double-flow face anti-fraud detection method based on weight fusion and feature selection | |
| KR101827889B1 (en) | Real-time Small Target Detection Method using Local Contrast Difference Measure | |
| WO2013088175A1 (en) | Image processing method | |
| CN106530281A (en) | Edge feature-based unmanned aerial vehicle image blur judgment method and system | |
| CN109902618A (en) | A kind of sea ship recognition methods and device | |
| CN103488992B (en) | A kind of oil spilling detection method towards complicated SAR image scene | |
| CN107808161A (en) | A kind of Underwater targets recognition based on light vision | |
| CN109685045A (en) | A kind of Moving Targets Based on Video Streams tracking and system | |
| CN106446925A (en) | Dolphin identity recognition method based on image processing | |
| CN106709426A (en) | Ship target detection method based on infrared remote sensing image | |
| CN105512622B (en) | A kind of visible remote sensing image sea land dividing method based on figure segmentation and supervised learning | |
| CN111695373B (en) | Zebra stripes positioning method, system, medium and equipment | |
| CN109190456A (en) | Pedestrian detection method is overlooked based on the multiple features fusion of converging channels feature and gray level co-occurrence matrixes | |
| CN111047624A (en) | Image dim target detection method, device, equipment and storage medium | |
| Wang et al. | Underwater object detection based on enhanced yolo | |
| CN110516565A (en) | A kind of sea ship recognition methods based on Zernike and SVM | |
| CN109815784A (en) | A kind of intelligent method for classifying based on thermal infrared imager, system and storage medium | |
| Akther et al. | Detection of Vehicle's Number Plate at Nighttime using Iterative Threshold Segmentation (ITS) Algorithm | |
| CN109784229A (en) | A kind of composite identification method of above ground structure data fusion | |
| Yang et al. | Intelligent collection of rice disease images based on convolutional neural network and feature matching | |
| CN104820818A (en) | Fast recognition method for moving object |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190503 |
|
| RJ01 | Rejection of invention patent application after publication |