CN115270613A - Full-link comprehensive inversion method for target infrared intrinsic characteristics - Google Patents
Full-link comprehensive inversion method for target infrared intrinsic characteristics Download PDFInfo
- Publication number
- CN115270613A CN115270613A CN202210834914.2A CN202210834914A CN115270613A CN 115270613 A CN115270613 A CN 115270613A CN 202210834914 A CN202210834914 A CN 202210834914A CN 115270613 A CN115270613 A CN 115270613A
- Authority
- CN
- China
- Prior art keywords
- target
- infrared
- full
- link
- attitude
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Aiming, Guidance, Guns With A Light Source, Armor, Camouflage, And Targets (AREA)
Abstract
The invention discloses a full-link comprehensive inversion method for target infrared intrinsic characteristics, which starts from full-link radiation coupling imaging characteristics, constructs a target imaging link radiation coupling model, realizes the comprehensive inversion of the target intrinsic infrared characteristics based on the coupling model and related infrared BRDF parameters, environmental radiation, atmospheric transmittance, equipment noise and imaging effects, improves the accuracy of test data, provides accurate target, environmental infrared radiation data and coupling relation data for model checking and simulation verification, and improves the reliability of the whole virtual simulation test.
Description
Technical Field
The invention belongs to the technical field of infrared imaging data inversion, and particularly relates to a full-link comprehensive inversion method for target infrared intrinsic characteristics.
Background
The target external field test is generally in a high-dynamic state with a long measurement distance, and is influenced by motion blur and optical dispersion, effective data obtained by sampling is less, data is blurred, and particularly under the conditions of point targets and small targets, the dispersion phenomenon is serious. Because the energy of the target is dispersed, the infrared imaging cannot reflect the real size of the target, the edge is close to the background radiation, and the specific position is difficult to extract, while the traditional bright Wen Fanyan model inverts the brightness of the target through the gray level output by infrared equipment, and then calculates the temperature by using the emissivity of the target according to the radiation theory, so that the gray level pixel is required to truly reflect the radiation characteristic of each part of the target, namely, the gray level pixel reflects the geometric characteristic of the target and reflects the radiation characteristic of the target, correct inversion can be carried out when the gray level pixel meets the corresponding relation of geometric projection, when the infrared imaging cannot reflect the real projection relation of the target, the gray level pixel after the target is extracted cannot correctly correspond to the target part, and when dispersion and motion blur occur, the inversion value is greatly different from the real value.
The target intrinsic infrared characteristic is a final measured value required to be obtained by an outfield infrared characteristic test of weapon equipment, an accurate value of model checking can be given only by obtaining the accurate target intrinsic infrared characteristic, however, a value sampled in the test process contains target intrinsic radiation, reflection caused by environmental radiation, an atmospheric path Cheng Fushe, optical dispersion and the influence of atmospheric absorption, coupling characteristics are complex, and meanwhile, the measured values of different conditions are inconsistent due to posture change and tail jet radiation change caused by target working condition change in the sampling process, so that the target infrared characteristic inversion needs to consider the influence of a full link on target radiation and can obtain an accurate target infrared characteristic value by decoupling according to a transmission process.
However, the current target feature inversion only focuses on atmospheric absorption and path radiation, and is not pertinently developed from a full link sampling mechanism, especially the characteristics of reflection, optical dispersion and the like of a target on environmental radiation are lack of detailed research, the modeling accuracy of a reflection model and a dispersion model is low, the inversion error is large, corresponding research is also lacked for the inversion of a non-cooperative target, foreign military equipment is used for testing the non-cooperative target, the model is a critical model for a virtual system modeling simulation test, and the accuracy directly influences the reliability of the simulation test.
Therefore, the target infrared characteristic data acquisition capability can be comprehensively improved only by improving the target infrared characteristic test equipment, infrared physical parameter calibration during external field test, environment factor real-time measurement precision and target data high-precision extraction and inversion technology, so that the problems of low infrared characteristic data precision, model check and difficult simulation verification application of the current weaponry are solved, the accuracy of a typical target of a main operational object and a complex actual combat environment model is improved, the infrared scene virtual simulation test reliability verification level is enhanced, and technical support is provided for weapon system virtual test environment construction and integrated comprehensive test verification technology.
Disclosure of Invention
Aiming at the existing problems, the invention provides a target infrared intrinsic characteristic full-link comprehensive inversion method, which is characterized in that a target imaging link radiation coupling model is constructed from the full-link radiation coupling imaging characteristic, and the comprehensive inversion of the target intrinsic infrared characteristic is realized based on the coupling model and relevant infrared BRDF parameters, environmental radiation, atmospheric transmittance, equipment noise and imaging effect.
The core thought of the invention is as follows:
decoupling layer by layer from the full link infrared radiation transmission characteristic to form an analytic equation and provide the intrinsic infrared characteristics of the target. The method of the invention carries out estimation and correction based on a theoretical model, gives accurate infrared characteristics of each factor of a scene as inversion characteristics, but not isolated radiation characteristic values, thereby giving the relation among a test value, target infrared characteristics and environmental radiation from a system level, providing a complete check test value for high-accuracy modeling, and meeting the requirements of target and scene infrared modeling simulation.
The technical solution for realizing the purpose of the invention is as follows:
a full-link comprehensive inversion method for target infrared intrinsic characteristics is characterized by comprising the following steps:
step 1: imaging through an optical system to obtain a target infrared measurement image;
and 2, step: obtaining a target radiation area by using an improved target infrared imaging data extraction method under optical dispersion;
and step 3: estimating the working condition of the target to be measured in the radiation area;
and 4, step 4: and (3) inputting the target radiation area obtained in the step (2) and the working condition data obtained in the step (3) into a target full-link infrared imaging model, and performing target infrared characteristic comprehensive inversion to finally obtain target infrared measurement image characteristics.
Further, the specific operation steps of the target infrared imaging data extraction method in step 2 are as follows:
step 21: calculating the brightness average value of the surrounding background area of the target infrared measurement image, and taking the average value as a threshold value;
step 22: and judging whether the brightness of the target infrared measurement image is greater than a threshold value or not, and if so, obtaining a diffuse spot pixel area as a target radiation area.
Further, the specific operation steps of step 3 are:
step 31: estimation of a target of a game
Acquiring the working condition of the airplane in real time through flight parameters to obtain the pitch angle, the roll angle, the attitude value, the position and the speed of a target;
step 32: estimation of non-cooperative targets
And establishing a three-dimensional attitude information identification frame, and estimating the non-cooperative target based on the identification frame to obtain the pitch angle, the roll angle, the attitude value, the position and the speed of the target.
Further, the specific operation steps of step 32 are:
step 321: converting target infrared measurement images under different posture conditions into binary target images through image segmentation;
step 322: extracting a target contour from the obtained binary target image, and obtaining an azimuth angle phi based on an image shape analysis method and an angle distance curve amplitude phase inference method by using an attitude image database as a reference sample;
step 323: extracting RTS characteristic invariant from a binarization target image;
step 324: normalizing the obtained RTS characteristic invariant;
step 325: constructing a training data set by using the feature invariants obtained under different posture conditions, and training a neural network;
step 326: inputting the target infrared measurement image to be measured into the trained neural network to obtain a pitch angle theta and a roll angle gamma;
step 327: obtaining the attitude probability value of the current frame of the target infrared measurement image to be measured by utilizing a neural network, sequencing the probability values, and taking the first four as possible attitude values;
step 328: detecting any two attitude values based on an improved Hausdorff distance measure contour matching method, and carrying out filtering detection to obtain an attitude estimation value of the current frame;
step 329: and constructing a sequence image position information filtering model based on the attitude estimation value, estimating the position and the speed of the current frame by using the model, and finally obtaining position and speed information.
Further, the specific operation steps of step 329 are:
step 3291: when the position signal is marked as X (t), the following conditions are satisfied:
wherein, α and β represent a starting point and an end point in a period of time, W (t) is a white normal process, and the mean and covariance of W (t) are:
step 3292: let the mean of X (t) satisfy:
meanwhile, let the variance of X (t) satisfy:
then, the available velocity averages are:
further, the specific operation steps of step 4 include:
step 41: inputting the obtained pitch angle, roll angle, attitude value, position and speed into a target full-link infrared imaging model, wherein the imaging model is a transmission link model of a target → background → interference → atmosphere → detection system;
step 42: establishing a full link imaging equation set based on the target full link infrared imaging model and solving the equation set;
step 43: and obtaining the infrared radiation brightness value of the corresponding target part of each sampling data.
Preferably, the neural network in step 326 is any one of a BP neural network, a GRNN network classifier, and a v-SVM regression machine.
Preferably, the feature invariants obtained in step 323 include Hu moment, fourier descriptor, and DCT descriptor.
Compared with the prior art, the method has the following beneficial effects:
the invention aims at typical target infrared characteristic test equipment, infrared physical parameter calibration during outfield test, real-time environmental factor measurement and target data high-precision extraction and inversion technology to form a standard outfield infrared characteristic test method of weaponry, which is used for comprehensively improving the target infrared characteristic data acquisition capability, solving the problems of low precision, model check and difficult simulation verification application of the current weaponry infrared characteristic data, improving the accuracy of typical targets and complex actual combat environment models of main combat objects, enhancing the reliability verification level of infrared scene virtual simulation tests, and providing technical support for virtual test environment construction of weapon systems and integrated comprehensive test verification technology.
The method is based on the coupling model and combines related infrared BRDF parameters, environmental radiation, atmospheric transmittance, equipment noise and imaging effect to realize comprehensive inversion of intrinsic infrared characteristics of the target, can effectively improve the accuracy of test data, provides accurate target and environmental infrared radiation data and coupling relation data for verification and simulation verification of the traditional bright Wen Fanyan model, and improves the reliability of the whole virtual simulation test.
Drawings
FIG. 1 is a flow chart of attitude estimation for a non-cooperative aerial target;
FIG. 2 is a schematic diagram of a database of gesture images;
fig. 3 is a schematic diagram of an infrared imaging link model of an airborne target.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the following further describes the technical solution of the present invention with reference to the drawings and the embodiments.
1. Target infrared imaging data extraction under optical dispersion
In the actual external field test, when the aerial target is far away from the optical system, the image formed by the aerial target through the optical system is generally a point target, namely the area of the aerial target on a detector pixel is about several pixel areas, and at the moment, due to the diffraction effect of the optical system, the energy of the aerial target is diffused, and aiming at the situation, the application of the traditional radiation characteristic extraction method can influence the precision of radiation characteristic measurement, so that the invention provides a new target infrared imaging data extraction method under optical diffusion based on the energy, the geometric projection relation and the multiband radiation characteristic correlation.
According to the energy conservation law, the radiation flux passing through the clear aperture is the same as the radiation flux forming the diffuse speckles, the diffuse speckles are formed by projecting the diffuse speckles on a detector around a target, and the size of the diffuse speckles is related to the target energy, and the larger the energy is, the larger the formed diffuse speckles are, so the invention considers the size of the diffuse speckles in different wave bands, the lateral and head-on long waves are strongest due to the large skin radiation energy of the long wave, and the short wave energy is strong in the tail nozzle, so the union of the maximum areas (the diffuse speckle pixel areas in the target infrared imaging image) is selected as the target radiation area, namely the brightness average value of the background area around the maximum area is selected as a threshold value, and the diffuse speckle pixel area is selected if the brightness average value is larger than the threshold value, thereby obtaining the pixel diffuse speckle area, namely the target area.
2. Behavior estimation for non-cooperative targets
The target working condition and the target infrared characteristic are closely related, and are important parameters of target infrared modeling simulation, for a traditional shooting range test, the working condition of an airplane can be acquired in real time through flight parameters, but for non-cooperative targets such as foreign airplanes and the like, flight data cannot be obtained publicly, and for obtaining the related parameters such as postures and positions, the related parameters must be obtained through reverse deduction of test data, so that a three-dimensional posture information identification frame is provided and used for identifying the postures and the positions of the non-cooperative targets.
The target posture recognition belongs to the categories of computer vision and mode recognition, and the problem of monocular vision is solved by acquiring images by an infrared imaging detector and deducing the target posture of the airplane, so that the method has higher difficulty. According to computer vision, if the target of the 3D structure is known, projected images under different attitude angles can be conveniently obtained through affine transformation or projective transformation. However, images obtained by affine or projective transformation projection have strong nonlinear relation with the shot-eye distance and the visual angle, the images obtained from different distances and visual angles are different, and many geometric properties have no invariance, for example, the projection of two parallel straight lines on the images is generally not parallel, the length of a line segment is also related to the shot-eye distance and the direction, so that the problem of sample data conflict exists. This makes it very difficult to solve its three-dimensional pose from the 2D image. For the infrared image, the gray level reflects the radiation energy of each part of the airplane, does not reflect the depth information of the structural component, and can utilize the radiation characteristics of only shape features and special areas such as engine nozzles and the like, and the identified solution has the property of multiple solutions.
Therefore, the three-dimensional attitude information identification framework is designed based on the machine learning method, and the problems of sample data conflict and multi-solution are solved. And converting the multi-class identification problem of angle identification into a nonlinear function approximation problem by analyzing the distribution characteristics of the attitude image characteristic quantities. And constructing a feature vector by using translation, rotation and scaling invariants, and constructing a posture estimator by using three networks for verification research. The specific identification frame structure is shown in fig. 1, and comprises the following steps:
(1) The method comprises the steps of obtaining infrared imaging under different posture conditions through measurement of an optical system, converting real-time infrared images obtained based on the infrared imaging into binary target images through image segmentation, neglecting tail flame influence at the moment, and concentrating on target feature detection and posture estimation of utilization regions and edges.
(2) In the binarization target image, a target contour is extracted, an attitude image database shown in fig. 2 is used as a reference sample, and an included angle between the projection of a crankshaft in an imaging plane and the x axis of the imaging plane, namely an azimuth angle phi, is obtained based on an image shape analysis method and an angle distance curve amplitude phase inference method which are disclosed in the prior art.
(3) And extracting RTS characteristic invariants including Hu moment, fourier descriptor and DCT descriptor in the binarization target image.
(4) Normalization (normalization) of the data. Before training and testing of the network, normalization processing is carried out on the invariant, and the neuron output saturation phenomenon caused by the overlarge net input absolute value is prevented.
(5) And constructing a training data set by using the feature invariants obtained under different posture conditions, and training a BP neural network or a GRNN classifier or a v-SVM regression machine.
(6) And inputting the normalized feature vector obtained from the test image into a classifier/regression machine, and fitting by utilizing network approximation capability to obtain an included angle between the aircraft space attitude and an imaging plane, namely a pitch angle theta and a roll angle gamma.
(7) Estimating the obtained attitude, obtaining a probability value through a classifier, sequencing the probability value, taking the first four attitude values as corresponding four attitude values, checking two possible attitude values by using a profile matching method based on the disclosed improved Hausdorff distance measure, and performing filtering check by using the correlation of the attitude change of the sequence image to obtain the attitude estimation value of the current frame.
A sequence image position information filtering model is constructed on the basis of the attitude estimation value, and the solution of information filtering is based on the fact that the position and speed change continuity. Namely, the position and the speed of the current frame are estimated by utilizing the sequences of the former frames through the continuity of the position and the speed of the sequence image, and the construction step of the sequence image position information filtering model comprises the following steps:
step 1: assuming that the change of the attitude signal along with the time is a continuous normal-Markov process, if the position signal is recorded as X (t), the following conditions are satisfied:
wherein W (t) is a white normal process with mean and covariance:
step 2: let the mean of X (t) satisfy:
let the variance of X (t)
Equations (1) - (2) are the established sequence image position information filtering model, and the state equation set of the attitude and position parameters can be established by using the sequence image position information filtering model, and the position and speed information can be estimated by using the Kalman filtering model.
The most obvious performance of the working condition of the engine in the infrared target characteristic is the medium wave and short wave characteristics, and the estimation of the working condition of the engine is realized by combining the change of distance and attitude estimation analysis medium wave imaging data.
3. Target infrared characteristic comprehensive inversion method based on target full-link infrared imaging model
In the process of testing the infrared characteristics of the aerial target, the signal transmission and conversion process of the optoelectronic system is a complex dynamic process under the coupling effect of the full link influence factors, and the link model is a transmission link model of "target → background → interference → atmosphere → detection system", as shown in fig. 3.
In general, intrinsic radiation of a target includes nonlinear and linear physical effect couplings between an environment and each unit module in an infrared imaging process, so that corresponding quantity is obtained by measuring a target physical parameter and a background environment radiation parameter in a test process, and then from the perspective of truly reflecting a real test process of a system, based on a ray tracing principle, an infrared full-link imaging model is combined with a test value to construct a full-link imaging equation set, and an infrared radiation brightness value of a corresponding target part of each sampling data is solved.
Considering that the equation set may be an contradictory equation set in the presence of the test error, the approximate radiance value can be obtained by solving the contradictory equation set through the least square method.
Those not described in detail in this specification are within the skill of the art. Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes in the embodiments and modifications of the invention can be made, and equivalents of some features of the invention can be substituted, and any changes, equivalents, improvements and the like, which fall within the spirit and principle of the invention, are intended to be included within the scope of the invention.
Claims (8)
1. A full-link comprehensive inversion method for target infrared intrinsic characteristics is characterized by comprising the following steps:
step 1: imaging through an optical system to obtain a target infrared measurement image;
step 2: obtaining a target radiation area by using an improved target infrared imaging data extraction method under optical dispersion;
and step 3: estimating the working condition of the target to be measured in the radiation area;
and 4, step 4: and (3) inputting the target radiation area obtained in the step (2) and the working condition data obtained in the step (3) into a target full-link infrared imaging model, and performing target infrared characteristic comprehensive inversion to finally obtain target infrared measurement image characteristics.
2. The method for full-link comprehensive inversion of the infrared intrinsic characteristics of the target according to claim 1, wherein the method for extracting the infrared imaging data of the target in the step 2 comprises the following specific operation steps:
step 21: calculating the brightness average value of the surrounding background area of the target infrared measurement image, and taking the average value as a threshold value;
step 22: and judging whether the brightness of the target infrared measurement image is greater than a threshold value or not, and if so, obtaining a diffuse spot pixel area as a target radiation area.
3. The method for full-link comprehensive inversion of the infrared intrinsic characteristics of the target according to claim 1, wherein the specific operation steps in step 3 are as follows:
step 31: estimation of a target of a game
Acquiring the working condition of the airplane in real time through flight parameters to obtain the pitch angle, the roll angle, the attitude value, the position and the speed of a target;
step 32: estimation of non-cooperative targets
And establishing a three-dimensional attitude information identification frame, and estimating the non-cooperative target based on the identification frame to obtain the pitch angle, the roll angle, the attitude value, the position and the speed of the target.
4. The method of claim 3, wherein the step 32 comprises the following specific steps:
step 321: converting target infrared measurement images under different posture conditions into binary target images through image segmentation;
step 322: extracting a target contour from the obtained binary target image, and obtaining an azimuth angle phi based on an image shape analysis method and an angle distance curve amplitude phase inference method by using an attitude image database as a reference sample;
step 323: extracting RTS characteristic invariant from a binarization target image;
step 324: normalizing the obtained RTS characteristic invariant;
step 325: constructing a training data set by using the feature invariants obtained under different posture conditions, and training a neural network;
step 326: inputting the target infrared measurement image to be measured into the trained neural network to obtain a pitch angle theta and a roll angle gamma;
step 327: obtaining the attitude probability value of the current frame of the target infrared measurement image to be measured by utilizing a neural network, sequencing the probability values, and taking the first four as possible attitude values;
step 328: detecting any two attitude values based on an improved Hausdorff distance measure contour matching method, and carrying out filtering detection to obtain an attitude estimation value of the current frame;
step 329: and constructing a sequence image position information filtering model based on the attitude estimation value, estimating the position and the speed of the current frame by using the model, and finally obtaining position and speed information.
5. The method for full-link comprehensive inversion of infrared intrinsic characteristics of targets according to claim 4, wherein the specific operation steps of step 329 are as follows:
step 3291: when the position signal is marked as X (t), the following conditions are satisfied:
wherein, α and β represent a starting point and an end point in a period of time, W (t) is a white normal process, and the mean and covariance of W (t) are:
mW(t)＝E(W(t))
step 3292: let the mean of X (t) satisfy:
meanwhile, let the variance of X (t) satisfy:
then, the average of the available velocities is:
6. the method for full-link comprehensive inversion of the infrared intrinsic characteristics of the target according to claim 4, wherein the specific operation steps of the step 4 comprise:
step 41: inputting the obtained pitch angle, roll angle, attitude value, position and speed into a target full-link infrared imaging model, wherein the imaging model is a transmission link model of a target → background → interference → atmosphere → detection system;
step 42: establishing a full link imaging equation set based on the target full link infrared imaging model and solving the equation set;
step 43: and obtaining the infrared radiation brightness value of the corresponding target part of each sampling data.
7. The method of claim 4, wherein the neural network of step 326 is any one of a BP neural network, a GRNN network classifier, or a v-SVM regression.
8. The method for full-link comprehensive inversion of target infrared intrinsic characteristics as claimed in claim 4, wherein the feature invariants obtained in step 323 include Hu moment, fourier descriptor, DCT descriptor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210834914.2A CN115270613A (en) | 2022-07-16 | 2022-07-16 | Full-link comprehensive inversion method for target infrared intrinsic characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210834914.2A CN115270613A (en) | 2022-07-16 | 2022-07-16 | Full-link comprehensive inversion method for target infrared intrinsic characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115270613A true CN115270613A (en) | 2022-11-01 |
Family
ID=83765046
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210834914.2A Pending CN115270613A (en) | 2022-07-16 | 2022-07-16 | Full-link comprehensive inversion method for target infrared intrinsic characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115270613A (en) |
-
2022
- 2022-07-16 CN CN202210834914.2A patent/CN115270613A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11428803B2 (en) | Method and apparatus for SAR image data enhancement, and storage medium | |
| CN110533069B (en) | Two-dimensional foil strip distribution characteristic identification method based on support vector machine algorithm | |
| CN106949907B (en) | A kind of quick system detection method of side of a ship window star | |
| CN103226196B (en) | Radar target recognition method based on sparse feature | |
| CN103353988A (en) | Method for evaluating performance of heterogeneous SAR (synthetic aperture radar) image feature matching algorithm | |
| CN112949380B (en) | Intelligent underwater target identification system based on laser radar point cloud data | |
| CN104268880A (en) | Depth information obtaining method based on combination of features and region matching | |
| CN101146233A (en) | A computing and image correction method for light source color | |
| CN111241667A (en) | Method for recognizing plasma configuration based on image processing and probe data processing | |
| CN106296708B (en) | Car tracing method and apparatus | |
| CN111401168A (en) | Multi-layer radar feature extraction and selection method for unmanned aerial vehicle | |
| CN114529615B (en) | Radar calibration method, device and storage medium | |
| CN113902663A (en) | Air small target dynamic infrared simulation method and device capable of automatically adapting to weather | |
| Yang et al. | An RGB channel operation for removal of the difference of atmospheric scattering and its application on total sky cloud detection | |
| Chen et al. | SAR raw data simulation for fluctuant terrain: A new shadow judgment method and simulation result evaluation framework | |
| CN116740332B (en) | Method for positioning center and measuring angle of space target component on satellite based on region detection | |
| CN115270613A (en) | Full-link comprehensive inversion method for target infrared intrinsic characteristics | |
| CN116630937A (en) | Multimode fusion 3D target detection method | |
| Noor et al. | A hybrid deep learning model for UAVs detection in day and night dual visions | |
| Li et al. | Spatio-context-based target tracking with adaptive multi-feature fusion for real-world hazy scenes | |
| CN114429593A (en) | Infrared small target detection method based on rapid guided filtering and application thereof | |
| CN109781259B (en) | Method for accurately measuring infrared spectrum of small aerial moving target through spectrum correlation | |
| Tang et al. | Target recognition method of laser imaging fuze based on deep transfer learning | |
| CN110991354A (en) | ISAR image identification method based on inverse space transformation network | |
| CN117523428B (en) | Ground target detection method and device based on aircraft platform |
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 |