CN114067188A - Infrared polarization image fusion method for camouflage target - Google Patents
Infrared polarization image fusion method for camouflage target Download PDFInfo
- Publication number
- CN114067188A CN114067188A CN202111404888.1A CN202111404888A CN114067188A CN 114067188 A CN114067188 A CN 114067188A CN 202111404888 A CN202111404888 A CN 202111404888A CN 114067188 A CN114067188 A CN 114067188A
- Authority
- CN
- China
- Prior art keywords
- polarization
- image
- infrared
- target
- stokes
- 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
- 230000010287 polarization Effects 0.000 title claims abstract description 110
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 16
- 230000004927 fusion Effects 0.000 claims abstract description 26
- 238000013528 artificial neural network Methods 0.000 claims abstract description 15
- 238000004364 calculation method Methods 0.000 claims description 15
- 230000003042 antagnostic effect Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 7
- 238000000034 method Methods 0.000 claims description 5
- 239000000463 material Substances 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000003702 image correction Methods 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 description 5
- 238000003384 imaging method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000007499 fusion processing Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 239000013077 target material Substances 0.000 description 1
- 238000001931 thermography Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an infrared polarization image fusion method of a camouflage target, which performs characteristic fusion on an infrared polarization image of the camouflage target through pseudo-color fusion, and then resists a neural network through a multi-characteristic discriminator to ensure that the image generated by a generator performs high fusion on the polarization characteristic image, the shape characteristic and the like of the target. The secondary fusion method enables the polarization characteristics of the camouflage target to be fully extracted and fused, retains and enhances the polarization information of the target, and enables the target identification and detection to be more accurate.
Description
Technical Field
The invention relates to the field of image processing, in particular to an infrared polarization image fusion method for a camouflage target.
Background
The infrared imaging detection is an important means for obtaining target image information, and has wide application in various fields such as military target identification, geological exploration, map drawing, monitoring and the like. The traditional imaging technology mainly utilizes the spectrum and intensity information of target radiation, and when a target is in a hidden or disguised state, the traditional imaging technology is difficult to effectively detect and identify.
The infrared polarization imaging detection technology is a novel photoelectric imaging detection technology for realizing target detection by utilizing polarization characteristic information, and can reveal target detail characteristics, improve the contrast of a target and increase the detection distance. The polarization degree and the polarization angle image can be obtained by utilizing the polarization characteristics, and further, the details of the outline, the structure and the like of the object can be obtained. At present, the research aiming at infrared polarization image fusion mainly takes image information of polarization degree and polarization angle as main bodies and is fused with original image information to achieve the purpose of identifying a target or enhancing the target information under the conventional background. However, the research on the fusion of the infrared polarization images under the camouflage background is less, and how to extract the polarization information from the infrared polarization images of the camouflage target and perform the fusion of the images is the key of the application of the polarization technology.
Disclosure of Invention
The purpose of the invention is as follows: the invention provides an infrared polarization image fusion method of a disguised target, aiming at solving the technical problems of infrared polarization feature extraction and infrared polarization image fusion aiming at the disguised target.
The technical scheme is as follows: the invention relates to an infrared polarization image fusion method of a camouflage target, which comprises the following steps:
s1: collecting an infrared polarization image and an infrared image of the disguised target, and performing filtering and noise reduction pretreatment;
s2: performing Stokes polarization state calculation on the infrared polarization image preprocessed in the step S1 to obtain 4 Stokes parameters I, Q, U and V of the infrared polarization image, wherein I represents the total intensity of light; q represents the difference in intensity between horizontal and vertical polarization; u represents the intensity difference between 45 ° and 135 ° for the polarization direction; v represents the intensity difference of the left-hand and right-hand circular polarization components of the light;
s3: calculating the polarization degree P and the polarization angle according to the Stokes polarization state calculation result
S4: using degree of polarization P and angle of polarization
Carrying out complete polarized light decomposition calculation and extracting a polarization characteristic image;
s5: respectively selecting 3 polarization component gray-scale images with different characteristics from the polarization component images formed by the Stokes parameters Q, U and V as 3 different components of the RGB images, and performing RGB pseudo-color fusion to generate pseudo-color fusion images;
s6: the method comprises the steps of constructing a plurality of discriminators in an original antagonistic neural network, inputting different target essential characteristics in the discriminators respectively to generate a multi-characteristic discriminator antagonistic neural network, entering images generated by a generator into the different discriminators respectively to discriminate whether the images meet the different essential characteristics of a target, and continuously adjusting training until the images generated by the generator are almost not different from real images.
Preferably, the infrared polarization image and the infrared image of the disguised target can be acquired through a long-wave thermal infrared imager, a long-wave infrared polarizing film and an infrared camera, or acquired through Stokes polarization state calculation.
Preferably, the filtering and denoising preprocessing in S1 includes: image correction, filtering denoising, registration and clipping.
Preferably, the Stokes polarization state calculation performed on the preprocessed infrared polarization image in S2 specifically includes: the polarization state and the intensity of the light wave are described by 4 Stokes parameters of a Stokes vector, the Stokes parameters are all time average values of light intensity, have the dimension of intensity, can be directly detected by a detector, and the 4 Stokes parameters of the light wave are as follows:
in the formula I1、I2、I3And I4Representing light intensities with polarization directions of 0 °, 45 °, 90 ° and 135 °, respectively.
Preferably, the polarization degree P and the polarization angle in S3
The calculation formula of (a) is as follows:
preferably, the multi-feature discriminator in S6 needs to fuse and stitch the pseudo-color fusion map and the infrared image against the neural network.
Preferably, the different target essential features in S6 include a pseudo color fusion map, a polarization feature image, a shape feature and a material feature.
Has the advantages that: the invention adopts pseudo-color fusion to perform feature fusion on the infrared polarization image of the disguised target, and then resists the neural network through the multi-feature discriminator, so as to ensure that the image generated by the generator performs high fusion on the polarization feature image, the shape feature and the like of the target. The secondary fusion method enables the polarization characteristics of the camouflage target to be fully extracted and fused, retains and enhances the polarization information of the target, and enables the target identification and detection to be more accurate.
Drawings
FIG. 1 is a pseudo color fusion process diagram of a polarization component diagram of a decoy target in the present invention;
fig. 2 is a diagram of secondary fusion processing performed by the countermeasure neural network on the pseudo-color fused image and the infrared image in the present invention.
Detailed Description
The invention is further illustrated by the following figures and examples. The examples are provided for the purpose of illustration only and are not intended to limit the scope of the invention.
Example 1: as shown in fig. 1 and 2, an infrared polarization image fusion method for a camouflaged target includes the following steps:
s1: collecting an infrared polarization image and an infrared image of the disguised target, and carrying out filtering and noise reduction pretreatment, wherein the infrared polarization image and the infrared image of the disguised target can be obtained through a long-wave thermal infrared imager, a long-wave infrared polarizing film and an infrared camera, or obtained through Stokes polarization state resolving;
s2: performing Stokes polarization state calculation on the infrared polarization image preprocessed in the step S1 to obtain 4 Stokes parameters I, Q, U and V of the infrared polarization image, wherein I represents the total intensity of light; q represents the difference in intensity between horizontal and vertical polarization; u represents the intensity difference between 45 ° and 135 ° for the polarization direction; v represents the intensity difference of the left-hand and right-hand circular polarization components of the light;
specifically, the polarization state and the intensity of the light wave are described by 4 stokes parameters of a stokes vector, the stokes parameters are all time average values of light intensity, have intensity dimensions, can be directly detected by a detector, and the 4 stokes parameters of the light wave are as follows:
in the formula I1、I2、I3And I4Respectively representing light intensities with polarization directions of 0 °, 45 °, 90 ° and 135 °
S3: calculating the polarization degree P and the polarization angle according to the Stokes polarization state calculation result
Wherein:
s4: using degree of polarization P and angle of polarization
Carrying out complete polarized light decomposition calculation and extracting a polarization characteristic image;
s5: respectively selecting 3 polarization component gray-scale images with different characteristics from the polarization component images formed by the Stokes parameters Q, U and V as 3 different components of the RGB images, and performing RGB pseudo-color fusion to generate pseudo-color fusion images;
s6: the method comprises the steps of constructing a plurality of discriminators in an original antagonistic neural network, inputting different target essential characteristics in the discriminators respectively to generate a multi-characteristic discriminator antagonistic neural network, fusing and splicing a pseudo-color fusion image and an infrared image through the multi-characteristic discriminator antagonistic neural network, then enabling images generated by a generator to enter different discriminators respectively to discriminate whether the images conform to the different essential characteristics of a target or not, and continuously adjusting and training until the images generated by the generator are almost not different from real images, wherein the different target essential characteristics comprise a pseudo-color fusion image, a polarization characteristic image, shape characteristics and material characteristics.
Example 2: taking a disguised vehicle target as an example, the testing distance is 500m, and the testing equipment adopts a medium wave/long wave infrared polarization imaging device, a traditional thermal imager and an infrared long wave polaroid. Firstly, acquiring and analyzing the obtained image data, performing Stokes calculation and image fusion processing on the polarization image, and giving a traditional thermal imaging graph, a polarization degree P graph and a polarization angle
The image, the pseudo color fusion image and the analysis evaluation table.
The method comprises the steps of constructing a plurality of discriminators in an original antagonistic neural network, inputting different target intrinsic characteristics such as a pseudo-color fusion image, a polarization characteristic image, a shape characteristic and a material characteristic into the discriminators respectively to generate the antagonistic neural network of the multi-characteristic discriminators, fusing and splicing the pseudo-color fusion image and an infrared image through the antagonistic neural network of the multi-characteristic discriminators, then respectively entering images generated by a generator into the discriminators to judge whether the images conform to the different intrinsic characteristics of a target or not, and continuously adjusting training until the images generated by the generator are almost not different from real images. And finally, a polarization fusion image containing target material, texture information, shape characteristics, polarization degree and polarization angle information is obtained, so that the subsequent target detection and identification are more accurate.
While there have been shown and described what are at present considered to be the fundamental principles of the invention, its essential features and advantages, it will be understood by those skilled in the art that the invention is not limited by the embodiments described above, which are included to illustrate the principles of the invention and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (7)
1. An infrared polarization image fusion method of a camouflage target is characterized in that: the method comprises the following steps:
s1: collecting an infrared polarization image and an infrared image of the disguised target, and performing filtering and noise reduction pretreatment;
s2: performing Stokes polarization state calculation on the infrared polarization image preprocessed in the step S1 to obtain 4 Stokes parameters I, Q, U and V of the infrared polarization image, wherein I represents the total intensity of light; q represents the difference in intensity between horizontal and vertical polarization; u represents the intensity difference between 45 ° and 135 ° for the polarization direction; v represents the intensity difference of the left-hand and right-hand circular polarization components of the light;
s3: calculating the polarization degree P and the polarization angle according to the Stokes polarization state calculation result
S4: using degree of polarization P and angle of polarization
Carrying out complete polarized light decomposition calculation and extracting a polarization characteristic image;
s5: respectively selecting 3 polarization component gray-scale images with different characteristics from the polarization component images formed by the Stokes parameters Q, U and V as 3 different components of the RGB images, and performing RGB pseudo-color fusion to generate pseudo-color fusion images;
s6: the method comprises the steps of constructing a plurality of discriminators in an original antagonistic neural network, inputting different target essential characteristics in the discriminators respectively to generate a multi-characteristic discriminator antagonistic neural network, entering images generated by a generator into the different discriminators respectively to discriminate whether the images meet the different essential characteristics of a target, and continuously adjusting training until the images generated by the generator are almost not different from real images.
2. The infrared polarization image fusion method of the camouflaged target according to claim 1, characterized in that: the infrared polarization image and the infrared image of the disguised target can be acquired through a long-wave thermal infrared imager, a long-wave infrared polarizing film and an infrared camera or acquired through Stokes polarization state resolving.
3. The infrared polarization image fusion method of the camouflaged target according to claim 1, characterized in that: the filtering and denoising preprocessing in the S1 includes: image correction, filtering denoising, registration and clipping.
4. The infrared polarization image fusion method of the camouflaged target according to claim 1, characterized in that: in the step S2, the Stokes polarization state calculation of the preprocessed infrared polarization image specifically includes: the polarization state and the intensity of the light wave are described by 4 Stokes parameters of a Stokes vector, the Stokes parameters are all time average values of light intensity, have the dimension of intensity, can be directly detected by a detector, and the 4 Stokes parameters of the light wave are as follows:
in the formula I1、I2、I3And I4Representing light intensities with polarization directions of 0 °, 45 °, 90 ° and 135 °, respectively.
5. The infrared polarization image fusion method of the camouflaged target according to claim 1, characterized in that: degree of polarization P and angle of polarization in the S3
The calculation formula of (a) is as follows:
6. the infrared polarization image fusion method of the camouflaged target according to claim 1, characterized in that: the S6 specifically includes the following steps: the multi-feature discriminator needs to fuse and splice the pseudo-color fusion image and the infrared image against the neural network.
7. The infrared polarization image fusion method of the camouflaged target according to claim 1, characterized in that: the following steps: the different target essential characteristics in the S6 comprise a pseudo color fusion image, a polarization characteristic image, a shape characteristic and a material characteristic.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111404888.1A CN114067188B (en) | 2021-11-24 | 2021-11-24 | Infrared polarized image fusion method for camouflage target |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111404888.1A CN114067188B (en) | 2021-11-24 | 2021-11-24 | Infrared polarized image fusion method for camouflage target |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114067188A true CN114067188A (en) | 2022-02-18 |
| CN114067188B CN114067188B (en) | 2024-07-05 |
Family
ID=80275818
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111404888.1A Active CN114067188B (en) | 2021-11-24 | 2021-11-24 | Infrared polarized image fusion method for camouflage target |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114067188B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115219026A (en) * | 2022-07-14 | 2022-10-21 | 中国科学院长春光学精密机械与物理研究所 | Polarization intelligent sensing system and sensing method |
| CN115346094A (en) * | 2022-08-25 | 2022-11-15 | 杭州电子科技大学 | Camouflage target detection method based on main body area guidance |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060164643A1 (en) * | 2003-06-25 | 2006-07-27 | Giakos George C | Multispectral, multifusion, laser-polarimetric optical imaging system |
| CN104978724A (en) * | 2015-04-02 | 2015-10-14 | 中国人民解放军63655部队 | Infrared polarization fusion method based on multi-scale transformation and pulse coupled neural network |
| CN108492274A (en) * | 2018-04-03 | 2018-09-04 | 中国人民解放军国防科技大学 | Long-wave infrared polarization feature extraction and fusion image enhancement method |
-
2021
- 2021-11-24 CN CN202111404888.1A patent/CN114067188B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060164643A1 (en) * | 2003-06-25 | 2006-07-27 | Giakos George C | Multispectral, multifusion, laser-polarimetric optical imaging system |
| CN104978724A (en) * | 2015-04-02 | 2015-10-14 | 中国人民解放军63655部队 | Infrared polarization fusion method based on multi-scale transformation and pulse coupled neural network |
| CN108492274A (en) * | 2018-04-03 | 2018-09-04 | 中国人民解放军国防科技大学 | Long-wave infrared polarization feature extraction and fusion image enhancement method |
Non-Patent Citations (2)
| Title |
|---|
| QINGQING XU, ET AL.: "Effective face detector based on YOLOv5 and superresolution reconstruction", RESEARCH ARTICLE, 16 November 2021 (2021-11-16), pages 1 - 9 * |
| 李世维;黄丹飞;刘辉;王惠敏;: "基于多尺度边缘表示的偏振图像二次融合算法", 激光与红外, no. 01, 20 January 2018 (2018-01-20) * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115219026A (en) * | 2022-07-14 | 2022-10-21 | 中国科学院长春光学精密机械与物理研究所 | Polarization intelligent sensing system and sensing method |
| WO2024012117A1 (en) * | 2022-07-14 | 2024-01-18 | 中国科学院长春光学精密机械与物理研究所 | Polarization intelligent sensing system and sensing method |
| CN115346094A (en) * | 2022-08-25 | 2022-11-15 | 杭州电子科技大学 | Camouflage target detection method based on main body area guidance |
| CN115346094B (en) * | 2022-08-25 | 2023-08-22 | 杭州电子科技大学 | Camouflage target detection method based on main body region guidance |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114067188B (en) | 2024-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhan et al. | Log-based transformation feature learning for change detection in heterogeneous images | |
| CN108492274B (en) | Long-wave infrared polarization feature extraction and fusion image enhancement method | |
| Acito et al. | Hyperspectral airborne “Viareggio 2013 Trial” data collection for detection algorithm assessment | |
| CN114067188B (en) | Infrared polarized image fusion method for camouflage target | |
| CN106372601A (en) | In vivo detection method based on infrared visible binocular image and device | |
| CN107560592B (en) | Precise distance measurement method for photoelectric tracker linkage target | |
| CN109754384B (en) | Infrared polarization image fusion method of uncooled infrared focal plane array | |
| CN104700381A (en) | Infrared and visible light image fusion method based on salient objects | |
| Fan et al. | SAR image registration using multiscale image patch features with sparse representation | |
| Qiao et al. | Source camera device identification based on raw images | |
| CN106991448A (en) | A kind of portrait compares processing method | |
| CN111582118A (en) | Face recognition method and device | |
| Ayhan et al. | A new approach to change detection using heterogeneous images | |
| Ye et al. | Robust optical and SAR image matching using attention-enhanced structural features | |
| US8649607B1 (en) | Spectral ratio contrast for edge detection in spectral images | |
| Palenichka et al. | Multiscale isotropic matched filtering for individual tree detection in LiDAR images | |
| CN113344987A (en) | Infrared and visible light image registration method and system for power equipment under complex background | |
| CN107145820B (en) | Binocular positioning method based on HOG characteristics and FAST algorithm | |
| US10614559B2 (en) | Method for decamouflaging an object | |
| You et al. | Tampering detection and localization base on sample guidance and individual camera device convolutional neural network features | |
| Bammey | Analysis and experimentation on the ManTraNet image forgery detector | |
| Xiao et al. | MOFA: A novel dataset for Multi-modal Image Fusion Applications | |
| CN114972140B (en) | Sea surface infrared and visible light image fusion method based on target segmentation | |
| Liu et al. | SAR image matching based on speeded up robust feature | |
| Jafarzadeh et al. | Probability estimation of change maps using spectral similarity |
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 |