Angle-gated scattering imaging method

Abstract

This invention discloses a scattering imaging method based on angle gating. It utilizes the angle gating principle to shield large-angle scattered light, improving the signal-to-noise ratio before camera acquisition. A time-domain minimization pixel channel filtering algorithm is developed, combined with image enhancement and denoising algorithms to achieve image restoration in foggy conditions. The single-frame image obtained by this invention shows a significant improvement in signal-to-noise ratio compared to traditional imaging methods; the restoration results of multi-frame images using the designed algorithm are superior to those of traditional stacking and averaging; and overall, it can achieve scattering imaging performance exceeding three times the visibility distance.

Description

Technical Field

[0001] This invention belongs to the field of scattering imaging, particularly passive scattering imaging under conditions of no active illumination, specifically a scattering imaging method based on angle gating. Background Technology

[0002] Imaging equipment struggles to acquire clear images in harsh environments such as fog, haze, smoke, and dust, primarily due to the scattering effect of particles in the atmosphere, which reduces the signal-to-noise ratio during sensor acquisition and consequently degrades image quality. Firstly, the target's reflected light is absorbed and scattered by suspended particles in the atmosphere, causing an attenuation of its energy and reducing the brightness of the target scene in the imaging system's results. Secondly, sunlight or other ambient light is scattered by particles in the atmosphere, creating stray light that reduces contrast. In environments with strong scattering, the intensity of this stray light can even be much greater than the target's reflected light, ultimately resulting in poor contrast or even complete blurring of the image acquired by the imaging system.

[0003] In foggy conditions, the atmospheric scattering imaging model is described as follows:

[0004] I(x)=J(x)e -rd(x) +A(1-e -rd(x) )

[0005] Where x is the spatial coordinate of an image pixel, I(x) is the acquired foggy image, J(x) is the clear target scene image to be recovered, r represents the atmospheric scattering coefficient, d represents the target scene distance, and A is the global atmospheric light, i.e., the atmospheric scattering light at infinity, which is usually assumed to be a global constant and is independent of the spatial coordinate x.

[0006] In the above formula, e -rd(x) Let t(x) represent the transmittance at coordinate space x, and we obtain the following formula:

[0007] I(x)=J(x)t(x)+A(1-t(x))

[0008] Therefore, the image dehazing process is the process of solving J(x) based on I(x). In order to solve J(x), it is necessary to solve for the transmittance t(x) and the global atmospheric light A based on I(x).

[0009] Traditional image dehazing methods are mainly divided into image enhancement based on image features and image restoration based on physical models. Image enhancement techniques improve display quality by stretching image contrast through algorithms, and can be used as a general comparison method to verify the effectiveness of image restoration techniques. Image restoration techniques mostly solve problems based on the aforementioned atmospheric scattering models or by measuring and estimating the scattering objects, such as dark channel dehazing, polarization dehazing, and wavefront correction. As can be seen from the above formulas, the reasons for image degradation are: firstly, multiplicative noise of the target scene signal with respect to attenuation; and secondly, additive noise of atmospheric scattered light that has not been modulated by the object. For self-luminous objects, the former dominates, while in passive detection environments illuminated by sunlight, the latter dominates.

[0010] The patent "Cloud-penetrating Imaging Method Based on Wavefront Correction (CN 108594429 A)" proposes a cloud-penetrating imaging method based on wavefront correction, which includes: correcting the wavefront damaged by cloud / fog by modulating a spatial light modulator, and then using a lens and detector to achieve point imaging of a point source object. At this point, the wavefront emitted by that point has been recovered. Due to the optical memory effect, the light waves emitted by points within a certain range near that point are also corrected to varying degrees, thereby achieving cloud-penetrating imaging. However, this method has the following problems:

[0011] 1. Due to the limitations of optical memory effect, the imaging range is not large;

[0012] 2. It requires measuring the damaged wavefront and using it for correction, which is time-consuming;

[0013] 3. Dynamic scattering environments require repeated measurements and calibrations, which are limited by the time of a single measurement and the decoherence time due to environmental changes. When the former is greater than the latter, the method fails.

[0014] 4. Relying on the measurement and recalibration of the scatterer, the signal-to-noise ratio of the acquisition was not improved. Summary of the Invention

[0015] To address the aforementioned problems and shortcomings, this invention proposes an angle-gated scattering imaging method. Utilizing the incident angle characteristics of signal and stray light, an angle-gated device is used to improve the signal-to-noise ratio (SNR) during acquisition, and algorithms are further used to optimize the image display effect. Therefore, it is not limited by the range of optical memory effects or the dynamic characteristics of the environment, requires no measurement or calibration, incurs no extra time costs, and, as this technology focuses on improving the SNR at the acquisition end, it can be transferred to other scattering imaging techniques as a universal SNR enhancement method.

[0016] In environments filled with scattering particles such as fog, haze, smoke, and dust, the scattering effect of these particles reduces the reflected light energy from the target scene, while increasing stray sunlight entering the camera through these particles. Under such conditions, the signal-to-noise ratio of the image acquired by the imaging device's sensor is significantly reduced, rendering the target scene unrecognizable.

[0017] In the atmospheric scattering imaging model, the noise-causing component A(1-t(x)) is represented as a constant value with respect to distance x. However, in real-world environments, the signal light satisfies the geometric imaging model and enters the camera sensor at a fixed angle of incidence. For the noise-causing skylight, its angle of incidence is random and varies over time. Therefore, the model can be modified as follows:

[0018] I(x,t)=J(x)t(x)+A(1-t(x))D(t)

[0019] Where D(t) represents the angular transmittance with respect to time t.

[0020] The technical solution of the present invention is as follows:

[0021] An angle-gated scattering imaging method utilizes the incident angle characteristics of the target signal light and stray light to shield large-angle stray light through angle gating technology, thereby obtaining a high signal-to-noise ratio original image. Based on the gating principle, a time-domain minimization pixel channel filtering algorithm is developed to construct an original image with an even higher signal-to-noise ratio from multiple frames. Finally, image enhancement and denoising algorithms are combined for further optimization to recover a clear target scene image in the super-visibility range.

[0022] Furthermore, the specific steps are as follows:

[0023] ① Install an angle gating device between the camera and lens of the imaging system to shield large-angle stray light and acquire a high signal-to-noise ratio original fog map I(x,t);

[0024] ② Modified atmospheric scattering model:

[0025] I(x,t)=J(x)t(x)+A(1-t(x))D(t)

[0026] In the formula, J(x) is the clear target scene image to be recovered, x is the spatial coordinate of the image pixel, t(x) represents the attenuation coefficient at position x, A is the global atmospheric light, that is, the atmospheric scattered light at infinity, and D(t) represents the corrected atmospheric scattered light angular transmittance with respect to time t.

[0027] ③ Based on the high signal-to-noise ratio (SNR) images of multiple frames passing through the angle gating system within a short time t, the original fog map I with the highest SNR during the acquisition time is constructed by using temporal minimization pixel channel filtering. re :

[0028]

[0029] The minimized pixel channel filtering process involves processing each pixel channel of a multi-frame image over a short period of time, extracting the pixel with the least superimposed noise during that time period to construct the fog map I to be recovered. re ;

[0030] ④ The original fog map I to be recovered with the highest signal-to-noise ratio re Visual restoration is performed, including contrast stretching using adaptive histogram equalization, denoising the stretched image using DCT filtering, and outputting a clear restored image of the target scene.

[0031] This invention relates to optical design for improving signal-to-noise ratio before image acquisition, aberration correction of optical systems, and algorithm processing after image acquisition.

[0032] The imaging lens, angle gating device, camera, and other components are coaxially placed to form the core imaging system. The field of view of the imaging lens is set to be smaller than the working field of view of the angle gating device to ensure minimal light intensity attenuation in the target scene. The invention can operate like a traditional imaging system under clear conditions. In foggy weather, only the addition of the angle gating device is needed to achieve scattering imaging. Since adding components to the optical system can cause optical path differences leading to defocusing, an electronically controlled focusing device is required for compensation and correction.

[0033] Preferably, the imaging lens is a folding telescope with a focal length of 2800mm, a light-gathering aperture of 280mm, and a field of view of 0.273° for the selected camera sensor size.

[0034] Preferably, the angle-gating device is a liquid crystal tunable filter with an adjustable range of 420–730 nm and a field of view of approximately 6°. It mainly utilizes the angle-gating characteristic: the larger the angle, the lower the transmittance.

[0035] Preferably, the focusing position of the electronically controlled focuser is recorded with numerical values, and the specified value can be returned with one key, which facilitates blind focusing in foggy or invisible conditions.

[0036] Preferably, the camera is a 16-bit high-sensitivity low-noise camera with a readout noise of 1.0 med e- and a dark current of less than 0.5 e- / pixel / s.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] 1. From an optical design perspective, improving the signal-to-noise ratio at the acquisition end can be universally applied to other defogging methods, thereby increasing the recovery limit;

[0039] 2. Compared with wavefront correction, transmission matrix measurement and other methods, the present invention does not require complex measurement and modulation, and is not limited by the range of memory effect or the dynamic characteristics of scattering medium;

[0040] 3. The angle gating device is designed as a detachable module, which can observe the target scene normally under clear conditions without causing energy loss or image degradation;

[0041] 4. The defocusing error caused by the switching angle gating device can be compensated by the electronically controlled focusing device, and blind focusing in foggy weather can be achieved with only one measurement;

[0042] 5. The signal-to-noise ratio of the raw image obtained by a single exposure is significantly improved, and the image is clearer under the same or even lower visibility conditions;

[0043] 6. Utilizing the angle gating principle of the invention, under the premise of frame rate allowance, the acquired multi-frame images are constructed with the developed temporal-domain minimized pixel channel filtering to construct images with higher signal-to-noise ratio, and the restoration results are better than the results of traditional superposition averaging.

[0044] 7. By improving the signal-to-noise ratio of the acquired image through front-end optical design, it can be used as a universal optimization scheme for other scattering imaging systems. Combined with the optimization processing of the developed time-domain minimized pixel channel filtering algorithm, image enhancement algorithm, and image denoising algorithm, the overall scattering imaging effect can be achieved by more than 3 times the visibility distance. Attached Figure Description

[0045] Figure 1 This is a schematic diagram illustrating the principle of the angle-gated scattering imaging method of the present invention.

[0046] Figure 2 It is a flowchart of the process for constructing multi-frame images with the highest signal-to-noise ratio to be restored and image enhancement processing.

[0047] Figure 3 This is a schematic diagram of an angle-gated scattering imaging system.

[0048] Figure 4 This is a defocus analysis diagram before and after the installation of the angle selection device. Detailed Implementation

[0049] This invention provides a scattering imaging method based on angle gating. By employing an angle gating device to shield large-angle stray light, the signal-to-noise ratio during image acquisition is improved. Based on the dynamic characteristics of the scattering environment, a time-domain minimization pixel channel filtering algorithm is developed in conjunction with this angle gating principle to construct a higher signal-to-noise ratio original image to be restored from multiple frames of images. Finally, an adaptive histogram equalization algorithm is used to improve contrast, and a DCT filtering denoising algorithm is used to optimize the image, achieving a scattering imaging restoration result with a visibility distance of more than 3 times.

[0050] In this embodiment of the invention, the application scenario is when traditional imaging faces the problem of image blurring due to scattering effects caused by water vapor, dust, and other various pollutant particles in the air. For example... Figure 1As shown, scattering particles in the air cause some of the reflected light from the target scene to be blocked, resulting in a decrease in signal strength. On the other hand, sunlight enters the camera sensor directly through the scattering effect of scattering particles. This part is called sky light, atmospheric scattered light, or stray light, which adds a relatively uniform background noise. The consequence is that the signal-to-noise ratio is reduced when the camera collects data in foggy weather, resulting in low image contrast and inability to distinguish the target scene.

[0051] In this embodiment of the invention, the incident angle characteristics of signal light and stray light are utilized to achieve a certain degree of stray light shielding through an angle gating device. In an imaging environment full of scattering particles, the target scene always satisfies the geometric imaging relationship. However, stray light is incident on the camera sensor at any angle, and light outside the field of view will also leave noise at the sensor after being reflected multiple times by the inner wall of the lens barrel.

[0052] The invention utilizes an angle-gating device—using a liquid crystal tunable filter as an example—to demonstrate the effectiveness of the method. In the embodiment, any wavelength can be selected because the angle-gating characteristic of the device is utilized; that is, the larger the angle, the lower the light transmittance. In particular, the light transmittance is extremely low beyond an effective field of view of 6°. Preferably, the lens is a telescope with a field of view of only 0.273° to ensure the transmittance of the target scene signal. Since stray light is at any angle, and most of it is greater than the angle of the signal light, although the introduction of optical elements will reduce the target scene signal, it shields large-angle stray light to a greater extent, thus improving the signal-to-noise ratio of the acquired image compared to direct imaging.

[0053] In this embodiment of the invention, instead of using the traditional method of stacking and averaging multiple consecutively acquired images with improved signal-to-noise ratio, a temporal minimization pixel channel filtering algorithm is developed to construct the original image to be restored. Specifically, as follows... Figure 2 As shown, under the same visibility, the target scene attenuation is constant, so the acquired fog image signal intensity is consistent. For multi-frame images, the target scene signal intensity is consistent for each pixel in time sequence, but the angle of random stray light superimposed on different pixels at different times is different. Under the same exposure time, due to the angle gating effect, the overall stray light of each pixel is lower than that of traditional imaging, and the fluctuation is greater in the time domain. For a specific pixel channel in the time domain, the larger the gray value, the more superimposed stray light noise there is. Therefore, under the condition that the frame rate allows, taking the minimum value of each pixel channel of multi-frame images over a period of time can construct the image with the highest signal-to-noise ratio of each pixel channel during this period, which is better than the traditional multi-frame superposition method. After obtaining the constructed high signal-to-noise ratio image, in order to facilitate human visual observation, adaptive histogram equalization is used to stretch the image contrast, and then DCT denoising algorithm is used to optimize the image display effect, ultimately achieving a scattering imaging effect of more than 3 times the visibility distance.

[0054] In this embodiment of the invention, the specific implementation steps are as follows:

[0055] 1. Connect the imaging lens, adapter, and camera without installing the LCD tunable filter;

[0056] 2. Connect the camera and computer power supply, open the camera operation software, set the exposure time, and open the real-time preview interface;

[0057] 3. Connect the theodolite power supply and handle, and locate the target scene;

[0058] 4. Adjust the electronic focus lever handle to focus on the target scene;

[0059] 5. Record the reading of the electronically controlled focuser as 1;

[0060] 6. Take scene images under clear conditions as the true values ​​for comparison with the restored image;

[0061] 7. Install an LCD tunable filter;

[0062] 8. Refocus;

[0063] 9. Record the reading of the electronically controlled focuser as 2;

[0064] 10. Remove the LCD tunable filter and restore the electronic focuser to a reading of 1;

[0065] 11. Normal monitoring of the target scene under clear conditions;

[0066] 12. Wait for foggy weather conditions;

[0067] 13. When fog occurs, turn on the visibility meter to record data and synchronize the time with the computer;

[0068] 14. Install a liquid crystal tunable filter and preheat it upon power-on;

[0069] 15. After the LCD tunable filter has finished warming up, the electronically controlled focuser is set to a reading of 2.

[0070] 16. Reset the exposure time to maintain an average grayscale value similar to the ground truth image at a similar intensity, and continuously acquire images;

[0071] 17. Use the temporal minimum pixel channel filtering algorithm to construct the original image to be restored from the ten foggy images collected above:

[0072]

[0073] 18. Adaptive histogram equalization is used to stretch the contrast of the above images;

[0074] 19. Use DCT filtering to denoise the above image;

[0075] 20. Output the restored image and the true image to the display window;

[0076] 21. Export visibility data and label the visibility values ​​when restoring the image.

[0077] In this embodiment of the invention, the analysis of defocus compensation for adding a liquid crystal tunable filter is as follows: Figure 4 As shown, x is the object distance, h is the lens diameter, i is the incident angle, f is the focal length, d is the thickness of the liquid crystal tunable filter with an equivalent refractive index of n, θ and y are the angular and longitudinal position differences of the emitted light before and after the liquid crystal tunable filter is installed, and ΔV is the focal plane position difference before and after the installation.

[0078] Optical calculations yield the following:

[0079] Offset angle:

[0080] Vertical offset: y = d(tan(i) - tan(θ))

[0081] Focal plane position difference:

[0082] When the imaging lens remains unchanged, the incident angle i remains unchanged. This offset is determined only by the thickness d and the equivalent refractive index n of the added device. This means that when the device is fixed, the defocusing amount is a fixed value when changing different scenes. Therefore, only one measurement is needed to record the difference in focusing readings before and after the addition. When changing different scenes, blind focusing can be achieved under foggy and invisible conditions.

[0083] During imaging, when there are a large amount of water vapor, dust, or other pollutant particles in the air, on the one hand, the reflected light energy of the imaging target scene is lost due to line-of-sight obstruction; on the other hand, sunlight is scattered by these particles and enters the camera sensor from various angles, resulting in strong background stray light. The invention utilizes the angle gating principle to shield large-angle scattered light, improving the signal-to-noise ratio before camera acquisition, and develops a time-domain minimization pixel channel filtering algorithm, combined with image enhancement and denoising algorithms to achieve image restoration in foggy weather. The single-frame image obtained by the invention shows a significant improvement in signal-to-noise ratio compared to traditional imaging; the multi-frame image restoration results using the designed algorithm are superior to the results of traditional stacking and averaging; overall, it can overcome the scattering imaging effect at three times the visibility distance.

[0084] The above description is only a preferred embodiment of the present invention. All equivalent changes and modifications made within the scope of the claims of the present invention should be included in the scope of the present invention.

Claims

1. An angle-gated based scatter imaging method, comprising an imaging system, characterized in that, The method comprises the following steps: ① An angle gating device is installed between the camera and the lens of the imaging system to shield large-angle stray light and collect high signal-to-noise ratio original fog map I(x, t); ② The atmospheric scattering model is corrected as follows: I(x, t) = J(x) t(x) + A(1-t(x)) D(t) In the formula, J(x) is a clear target scene image to be restored, x is the spatial coordinates of the image pixels, t(x) represents the attenuation coefficient at position x, A is the global atmospheric light, that is, the atmospheric scattering light at infinite distance, and D(t) represents the corrected atmospheric scattering light angle transmittance with respect to time t; ③According to the high signal-to-noise ratio images of multiple frames of the angle gating system in a short time t, the highest signal-to-noise ratio to-be-recovered original fog image I in the short time is constructed through time domain minimum pixel channel filtering re : The minimum pixel channel filtering is a pixel channel processing of multi-frame images in a short time, and a pixel point with minimum superimposed noise of each pixel in the short time is extracted to construct a to-be-recovered fog image I re ; IV. The highest signal-to-noise ratio is restored to the original fog map I re Visual recovery is performed, including contrast stretching using adaptive histogram equalization, and denoising processing on the stretched image using DCT filtering, to output a clear target scene image after recovery.

2. The angle-gated scatter imaging method of claim 1, wherein: The angle gating device is a liquid crystal tunable filter, a grating or a metasurface.

3. The angle-gated scatter imaging method of claim 1, wherein: The imaging system is provided with an electrically controlled focusing device, has a function of marking a focusing position and digitally returning to a specified reading position, and realizes blind focusing in foggy weather.

4. The angle-gated scatter imaging method of claim 1, wherein: The lens of the imaging system is a return type telescope lens, and the field of view angle of the lens is smaller than that of the angle gating device.

Images