Perception evaluation method for tiny signal limit caused by positive and negative low temperature difference change

Abstract

The invention discloses a method for perceiving and evaluating a tiny signal limit caused by positive and negative low temperature difference change, which comprises the following steps of: establishing a full-link time sequence simulation model based on target parameters, an atmospheric transmission path and environmental background simulation, and obtaining image simulation sequences under different radiation levels in an imaging system; the simulation sequence is corrected, a point target position is obtained through template matching according to an actual difference image sequence, and target scale calculation of a sub-pixel scale and background and target energy separation are achieved; and carrying out limit perception evaluation on the tiny signal target by adopting a mode of weighting a signal-to-noise ratio, a speed ratio and a camera energy concentration ratio ratio of a target center pixel. According to the method, the problem that a single tiny signal cannot be sensed under the low signal-to-noise ratio is solved, multiple parameters of the target are integrated, and a more stable evaluation method is established for tiny signal limit sensing.

Description

Technical Field

[0001] The present invention belongs to the field of target detection, and in particular relates to a method for evaluating the extreme perception of tiny signals caused by changes in positive and negative low temperature differences. Background Art

[0002] In target detection, space-based infrared remote sensing detection systems can effectively detect heat source information of various vehicles on the ground. Target detection and identification based on such systems has become the focus of the detection field. In detection, for aircraft targets, in the long-wave infrared detection band, only the spontaneous heat radiation of the aircraft skin can be detected. If the target is flying at a low speed, the spontaneous radiation generated by friction with the air is limited. In the case of high background and atmospheric radiation and other noise, its infrared image signal is close to the environmental noise, and the signal-to-noise ratio is very low, making it difficult to perceive such target signals.

[0003] Traditional remote sensing infrared target detection methods mainly rely on the significant difference between point high-temperature targets and background noise, and detect targets through background subtraction and threshold segmentation.

[0004] For example, a Chinese patent document with publication number CN1606033A discloses a method for detecting and tracking small targets in infrared image sequences. First, the image background is obtained through morphological filtering, and the image background is subtracted from the original image to obtain a background-removed image containing the target and noise. Then, the image is whitened for preprocessing. In the tracking stage, the projection transformation method is used to start tracking. In the tracking maintenance stage, an auxiliary particle filter is used to estimate the target state. In the detection stage, a likelihood ratio hypothesis test method is used to determine whether the target exists. Finally, the target's track and detection probability are obtained.

[0005] A Chinese patent document with publication number CN115937275A discloses a processing method for infrared image recognition, including acquiring an infrared image; preprocessing the infrared image to generate a score image; performing multi-scale segmentation processing on the score image to generate a high signal-to-noise ratio image; performing local thresholding processing on the high signal-to-noise ratio image to generate a target area; and filtering the target area to generate a target contour.

[0006] However, the above method requires the target to have an obvious high-energy response. It is difficult to detect the target when the target energy is insufficient and the target is in the form of a point image.

[0007] In general, traditional methods have limited perception capabilities for weak targets and cannot effectively meet the needs of target detection and subsequent tracking and identification when the infrared signal generated by the target is small. Summary of the invention

[0008] The present invention provides a method for evaluating the extreme perception of tiny signals caused by changes in positive and negative low temperature differences, which can improve the perception ability of weak targets and effectively meet the needs of target detection and subsequent tracking and identification when the infrared signal generated by the target is small.

[0009] A method for evaluating the extreme perception of tiny signals caused by changes in positive and negative low temperature differences comprises the following steps:

[0010] (1) Establish a full-link timing simulation model based on target parameters, atmospheric transmission path, and environmental background simulation to obtain image simulation sequences under different radiation levels in the imaging system;

[0011] (2) For the detector output image, each pixel is divided into sub-pixel scales, and a basic model of intra-pixel sensitivity is established based on the camera parameters. The point spread function of the imaging system used is obtained through testing;

[0012] (3) according to the preset working environment, under the image simulation sequence of step (1) and the point spread function convolution of step (2), the sub-pixel scale sequence of the target signal is calculated; after further differential processing, a simulated differential image template sequence is obtained;

[0013] (4) Obtaining an image sequence collected by the remote sensing observation equipment in actual operation, performing differential analysis on two images with a fixed frame difference in the image sequence, and obtaining a differential image sequence to be actually detected;

[0014] (5) For the difference image sequence obtained in step (4) and the difference image template sequence obtained in step (3), the position of the roughly located target point is obtained by template matching;

[0015] (6) Based on the point spread function value obtained by the test in step (2), target positioning and energy extraction are performed on the coarse positioning target point obtained by template matching in step (5);

[0016] (7) Based on the sub-pixel position and energy information of the target obtained in step (6), the signal-to-noise ratio, velocity ratio, and camera energy concentration ratio of the target center pixel are weighted to perform an extreme perception evaluation of the small signal target.

[0017] The specific process of step (1) is as follows:

[0018] (1-1) Simulation of the original radiation state of the target: According to the expected small signal target parameters (area, temperature, etc.), its emission characteristics are simulated to obtain the target's outgoing radiation in the observed environment;

[0019] (1-2) Simulation of environmental background radiation status: Based on the ground reflectivity model and solar radiation model, simulation is performed according to the average statistical values ​​of public meteorological data to obtain the environmental background radiation, including the uniform atmospheric thermal radiation in the transmission path and the radiation value of the surrounding environment where the target is located;

[0020] (1-3) Accumulating the original radiation state of the target and the environmental background radiation state to obtain the theoretical radiation state of the point target under different environmental background radiation states;

[0021] The relative position of the target in the environmental background is changed based on the preset target speed to obtain image simulation sequences with different target radiation levels, different environmental backgrounds and different motion states.

[0022] The specific process of step (2) is as follows:

[0023] For the detector output image, each pixel is divided into N×N sub-pixels. Through intra-pixel micro-scanning of a fixed point source in the laboratory, the point spread function of the imaging system used is calculated by fitting the convolution relationship based on the point spread function of the imaging system and the energy distribution of the fixed point source.

[0024] In step (2), the point spread function model is expressed by the following formula:

[0025]

[0026] Among them, IPS is the intra-pixel sensitivity of the detector, which is represented by a model based on the detector charge diffusion and capacitive coupling mechanism and its test parameters; PSF is the point spread function of the optical system; (Δx, Δy) is the distance between the position of the ideal point light source incident on the focal plane of the detector and the center of the target pixel.

[0027] The specific process of step (3) is as follows:

[0028] According to the actual working distance of the imaging system, the response data sequence of the preset target is calculated under the convolution of the image simulation sequence obtained in step (1) and the point spread function obtained in step (2), based on the sub-pixel scale division of the target signal; at the same time, according to the noise model of the detector used for imaging, a random noise component is added; the image data of two adjacent frames are differentially processed to suppress the background signal, retain the target signal, and obtain a new differential image template sequence.

[0029] In step (6), the coarsely positioned target point obtained by template matching in step (5) is subjected to target positioning and energy extraction, that is, an optimization solution expressed as follows is performed:

[0030]

[0031] Among them, P i′ ,j is the grayscale value of the actual pixel, B′ is the background estimation value, I′ is the target energy intensity to be determined, (x c ′ ,y c ′ ) is the centroid position of the target to be determined; ePSF is the point spread function model established in step (2).

[0032] The specific process of step (7) is as follows:

[0033] For the target point with a small signal, the signal-to-noise ratio, speed, and energy concentration of the target are calculated based on the sub-pixel position and energy information of the target. After comparing with the actual differential image sequence to be detected obtained in step (4), the signal-to-noise ratio, speed ratio, and energy concentration ratio are obtained, and the confidence of the detection of the small signal target is evaluated after weighting.

[0034] The weighted formulas for signal-to-noise ratio, speed ratio, and energy concentration ratio are as follows:

[0035]

[0036] Among them, a1, a2, a3 are weighting coefficients, SNR is the signal-to-noise ratio of the system, and v r With v s are the moving speeds of the actual target and the simulated target in the image sequence, respectively, r With d s are the energy concentrations of the actual target and the simulated target in the image sequence, respectively.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention adopts differential image sequence matching based on target radiation and point spread function simulation to replace the traditional threshold detection method, which improves the detection capability of weak signals, reduces the requirements for detection, has high sensitivity and strong timeliness.

[0039] 2. The present invention constructs an intensity database of weak air and space targets in different environments, which can be applied in tracking and identifying targets such as aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of a method for evaluating the tiny signal limit perception caused by a positive and negative low temperature difference change according to the present invention;

[0041] Figure 2 A schematic diagram of a target differential image obtained by a simulated detector in an embodiment of the present invention;

[0042] Figure 3It is a schematic diagram of obtaining a target differential image under an actual test environment in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be pointed out that the embodiments described below are intended to facilitate the understanding of the present invention and do not have any limiting effect on the present invention.

[0044] like Figure 1 As shown, a method for evaluating the extreme perception of a tiny signal caused by a positive and negative low temperature difference change comprises the following steps:

[0045] S1, establish a full-link timing simulation model based on target parameters, atmospheric transmission path, environmental background simulation and camera imaging parameters to obtain image simulation sequences under different target radiation levels, different environmental backgrounds and different motion states in the imaging system.

[0046] S101, simulation of the original radiation state of the target. In this embodiment, the target is positioned as an ordinary aerospace observation target, and its surface can be regarded as a Lambertian radiation source, that is, the emission of the surface (such as the aircraft skin) in all directions is uniform, and its radiation characteristics comply with Planck's law, that is,

[0047]

[0048] Among them, M tar is the target spectral radiant emittance, ε tar is the spectral emissivity, c1 and c2 are the first and second radiation constants respectively. According to the known temperature data of the target, the ideal target radiation characteristics can be obtained.

[0049] S102, simulation of environmental background radiation status.

[0050] The environmental background radiation mainly includes the uniform atmospheric thermal radiation in the transmission path and the radiation value of the surrounding environment where the target is located. Based on the ground reflectivity model and the solar radiation model, according to the average statistical value of the public meteorological data, the relevant software is used for simulation as the input condition for background estimation. In this embodiment, the mid-latitude summer ocean atmosphere model is simulated by modtran software. Under the conditions of the mid-latitude summer atmosphere model, rural aerosol, visibility of 23km, solar altitude angle of 45°, the spectral radiance of the target under the background of standard water body is used as the background radiance before the entrance pupil.

[0051] S103, simulating image simulation sequences at different radiation levels.

[0052] Based on the calculation results in S101 and S102, the original radiation state of the target and the background radiation state are accumulated to obtain the theoretical radiation situation of the point target under different background radiation states. The relative position of the target in the background is changed based on the preset target speed, and the image simulation sequence under different radiation levels can be simulated.

[0053] S2, based on the test, obtain the detector pixel response model, capacitive coupling and other models. According to the spatiotemporal noise simulation and detector model correction, obtain the sub-pixel scale simulation sequence of the target signal. Based on the template matching of the actual differential image sequence, obtain the point target position. With the multi-frame target scale invariance as the constraint condition, realize the sub-pixel scale target scale solution and background and target energy separation.

[0054] S201, obtain the detector fixed point spread function through laboratory testing.

[0055] In the embodiment, for the detector output image, each pixel is divided into 5×5 sub-pixels, and the point spread function value of the imaging system used is calculated by fitting the convolution relationship based on the imaging system point spread function and the energy distribution of the fixed point source through micro-scanning in the pixel of the laboratory fixed point source. The point spread function model in this embodiment can be expressed by the following formula:

[0056]

[0057] Among them, IPS represents the intra-pixel sensitivity of the detector, which is represented by a model based on the detector charge diffusion and capacitive coupling mechanism and its test parameters. PSF is the point spread function of the optical system, which is represented by a Gaussian function with a half-height width of 1 pixel.

[0058] S202: Acquire a sub-pixel scale sequence of a target signal according to a preset working environment.

[0059] According to the actual working distance of the imaging system, the image simulation sequence of the preset target obtained in S1 combined with the background is convolved with the point spread function of the detection system, and the response data sequence based on the sub-pixel scale division described in S201 is calculated. At the same time, according to the noise model of the detector used for imaging, a certain random noise component is added. In this embodiment, random noise with a signal-to-noise ratio of 20 is added. A simulated target image obtained by differential simulation in steps S1-S2 is as follows Figure 2 shown.

[0060] S203, obtaining an image sequence collected by the remote sensing observation equipment in actual operation, performing a difference between two images with a fixed frame number difference in the time domain image sequence, using the same method as S202, to form a target sequence to be detected.

[0061] S204, the differential image obtained in S203 is matched with the differential image template sequence obtained in S202 to obtain the possible target point position. For the image sequence, a certain rough positioning target point position can be obtained in each frame. In order to reduce the processing burden, the differential image within a certain range around the target is intercepted as the region of interest to be evaluated. For the single frame differential image obtained for the actual point source target, Figure 3 shown.

[0062] S205, based on the equivalent point spread function obtained by the test in S201, target positioning and energy extraction are performed on the coarse positioning target point obtained by template matching in S204, that is, an optimization solution expressed by the following formula is performed.

[0063]

[0064] Among them, P i ′ ,j is the grayscale value of the actual pixel, B′ is the background estimation value, I′ is the target energy intensity to be determined, (x c ′ ,y c ′ ) is the centroid position of the target to be determined. Figure 3 The target center of mass is [0.5, 0.5].

[0065] S3, based on the sub-pixel position and energy of the target, a multi-parameter weighted small signal limit perception evaluation method is constructed based on the central pixel signal-to-noise ratio, speed, camera energy concentration, etc.

[0066] The possible weak signal target point is evaluated, and the response signal-to-noise ratio of the target, the target's movement speed, and the sub-pixel scale of the target are calculated, and compared with the differential image sequence obtained in S2. The confidence of the weak target detection is evaluated through the difference between the two. In this embodiment, the weighted formula including the signal-to-noise ratio, speed ratio, and energy concentration ratio is adopted as follows:

[0067]

[0068] Among them, a1, a2, a3 are weighting coefficients, which are related to the specific goals and scenarios in the actual implementation of this method. SNR is the signal-to-noise ratio of the system, as calculated in the formula, v r With v s are the moving speeds of the actual target and the simulated target in the image sequence, respectively, r With d s are the energy concentrations of the actual target and the simulated target in the image sequence, respectively.

[0069] The embodiments described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for evaluating the extreme perception of tiny signals caused by changes in positive and negative low temperature differences, characterized in that: The following steps are involved: (1) Establish a full-link timing simulation model based on target parameters, atmospheric transmission path, and environmental background simulation to obtain image simulation sequences under different radiation levels in the imaging system; (2) For the detector output image, each pixel is divided into sub-pixel scales, and a basic model of intra-pixel sensitivity is established based on the camera parameters. The point spread function of the imaging system used is obtained through testing; (3) according to the preset working environment, under the image simulation sequence of step (1) and the point spread function convolution of step (2), the sub-pixel scale sequence of the target signal is calculated; after further differential processing, a simulated differential image template sequence is obtained; (4) Obtaining an image sequence collected by the remote sensing observation equipment in actual operation, performing differential analysis on two images with a fixed frame difference in the image sequence, and obtaining a differential image sequence to be actually detected; (5) For the difference image sequence obtained in step (4) and the difference image template sequence obtained in step (3), the position of the roughly located target point is obtained by template matching; (6) Based on the point spread function obtained by the test in step (2), target positioning and energy extraction are performed on the coarse positioning target point obtained by template matching in step (5); (7) Based on the sub-pixel position and energy information of the target obtained in step (6), the signal-to-noise ratio, velocity ratio, and camera energy concentration ratio of the target center pixel are weighted to perform an extreme perception evaluation of the small signal target.

2. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 1 is characterized in that: The specific process of step (1) is as follows: (1-1) Simulation of the original radiation state of the target: According to the expected small signal target parameters, its emission characteristics are simulated to obtain the target's outgoing radiation in the observed environment; (1-2) Simulation of environmental background radiation status: Based on the ground reflectivity model and solar radiation model, simulation is performed according to the average statistical values ​​of public meteorological data to obtain the environmental background radiation, including the uniform atmospheric thermal radiation in the transmission path and the radiation value of the surrounding environment where the target is located; (1-3) Accumulating the original radiation state of the target and the environmental background radiation state to obtain the theoretical radiation state of the point target under different environmental background radiation states; The relative position of the target in the environmental background is changed based on the preset target speed, and image simulation sequences with different target radiation levels, different environmental backgrounds and different motion states are obtained.

3. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 1, characterized in that: The specific process of step (2) is as follows: For the detector output image, each pixel is divided into N×N sub-pixels. Through intra-pixel micro-scanning of a fixed point source in the laboratory, the point spread function of the imaging system used is calculated by fitting the convolution relationship based on the point spread function of the imaging system and the energy distribution of the fixed point source.

4. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 1, characterized in that: In step (2), the point spread function model is expressed by the following formula: Among them, IPS is the intra-pixel sensitivity of the detector, which is represented by a model based on the detector charge diffusion and capacitive coupling mechanism and its test parameters; PSF is the point spread function of the optical system; (Δx, Δy) is the distance between the position of the ideal point light source incident on the focal plane of the detector and the center of the target pixel.

5. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 1, characterized in that: The specific process of step (3) is as follows: According to the actual working distance of the imaging system, the response data sequence of the preset target is calculated based on the sub-pixel scale division of the target signal under the convolution of the image simulation sequence obtained in step (1) and the point spread function obtained in step (2); at the same time, a random noise component is added according to the noise model of the detector used for imaging; the image data of two adjacent frames are differentially processed to suppress the background signal and retain the target signal, thereby obtaining a new differential image template sequence.

6. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 1, characterized in that: In step (6), the coarsely positioned target point obtained by template matching in step (5) is subjected to target positioning and energy extraction, that is, an optimization solution expressed as follows is performed: Among them, P i ′ ,j is the grayscale value of the actual pixel, B′ is the background estimation value, I′ is the target energy intensity to be determined, (x c ′ ,y c ′ ) is the centroid position of the target to be determined, and ePSF is the point spread function model established in step (2).

7. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 1, characterized in that: The specific process of step (7) is as follows: For the target point with a small signal, the signal-to-noise ratio, speed, and energy concentration of the target are calculated based on the sub-pixel position and energy information of the target. After comparing with the actual differential image sequence to be detected obtained in step (4), the signal-to-noise ratio, speed ratio, and energy concentration ratio are obtained, and the confidence of the detection of the small signal target is evaluated after weighting.

8. The method for evaluating the tiny signal limit perception caused by the positive and negative low temperature difference change according to claim 7, characterized in that: The weighted formulas for signal-to-noise ratio, speed ratio, and energy concentration ratio are as follows: Among them, a1, a2, a3 are weighting coefficients, SNR is the signal-to-noise ratio of the system, and v r With v s are the moving speeds of the actual target and the simulated target in the image sequence, respectively, r With d s are the energy concentrations of the actual target and the simulated target in the image sequence, respectively.

Images