A simulation method for long-distance near-infrared target turbulent imaging

Abstract

The present invention provides a simulation method for long-distance near-infrared target turbulent imaging, including: based on the correlation of the two-pixel offset when different two pixel points on the object-side near-infrared image are imaged to the same point on the image-side, obtaining the x-axis offset and y-axis offset that can modulate all pixels of the near-infrared output image to produce the turbulent random offset effect; based on the x-axis and y-axis offsets, processing the turbulent phase covariance function to obtain the atmospheric turbulence phase matrix; based on the atmospheric turbulence phase matrix, obtaining the point spread function that can modulate the output near-infrared image to produce the turbulent blur effect caused by high-order aberrations, and further obtaining the blur amount; adding the blur amount and the x-axis and y-axis offsets to the object-side input image to obtain the turbulent image. The present invention does not require step-by-step propagation, directly simulates the random offset and blur that cause the distortion of the turbulent image, conforms to the turbulent effect, has a fast calculation speed, and provides a basis for simulating the near-infrared image imaging under atmospheric turbulence.

Description

Technical Field

[0001] The present invention belongs to the technical field of turbulent imaging, and particularly relates to a simulation method for long-distance near-infrared target turbulent imaging. Background Art

[0002] Near-infrared band information has better penetration. In the complex environment of modern weapon operations, the near-infrared information of the afterflames of high-speed flying targets such as long-distance fighter jets and missiles can be obtained to locate the targets and then carry out optoelectronic countermeasure technologies such as optoelectronic reconnaissance, optoelectronic warning, and optoelectronic jamming. However, when imaging a long-distance near-infrared target, due to the existence of atmospheric turbulence disturbances in the imaging channel, the refractive index of the atmospheric medium changes unevenly with time and space randomly. Therefore, the quality of the captured long-distance near-infrared images will decline to varying degrees, mainly manifested as image blurring and image distortion. This impact will further reduce the accuracy of tracking and reconnaissance of long-distance near-infrared targets, with a large error. In order to still obtain clear long-distance near-infrared object information in the environment of atmospheric turbulence disturbances, currently, the main technical means to realize the reconstruction of near-infrared images of atmospheric turbulence is to use machine learning methods to train models. However, a large number of near-infrared image datasets of atmospheric turbulence disturbances are required during the training process. The image dataset is the basis for turbulent image reconstruction. Therefore, it is very necessary to quickly obtain a large number of datasets.

[0003] Generally, the methods for obtaining the dataset mainly include actually shooting turbulent images, simulating atmospheric turbulence images using a spatial light modulator, and calculating and simulating turbulent images using algorithms. The actually shot turbulent images are generally obtained by using an optical system composed of a telescope and a camera, and shooting for a long time and a long distance at different times and different weathers. However, this method is time-consuming and labor-intensive and cannot quickly obtain a large number of datasets; the principle of simulating atmospheric turbulence images using a spatial light modulator is to simulate an atmospheric turbulence phase screen and load it onto the spatial light modulator, so as to modulate the transmitted light wave to obtain an atmospheric turbulence image. However, this method is limited by the device accuracy;

[0004] The algorithm simulation has a fast calculation speed and the parameters can be flexibly adjusted. Therefore, using the algorithm to simulate and generate turbulent images has become the mainstream trend. The principle of the classic split-step propagation algorithm is to divide the atmospheric turbulence transmission channel into several phase screens, perform Fresnel diffraction integral calculations on each point of the input image, and then use the phase screen to represent the phase accumulation of the refraction of the light beam under the turbulent effect. Then, iterate in the direction of the receiving plane in turn, and finally obtain the atmospheric turbulence distorted image at the receiving plane. However, this method requires repeated Fourier transform and inverse Fourier transform operations, with a large amount of calculation and a long calculation time. Summary of the Invention

[0005] The object of the present invention is to propose a simulation method for long-distance near-infrared target turbulent imaging. This method does not require step-by-step propagation. Instead, it directly simulates the random offsets and blurs that cause image distortion to simulate the imaging of long-distance near-infrared targets in atmospheric turbulence. This method calculates faster, conforms to the actual turbulent imaging situation, can provide a large number of turbulent image datasets for turbulent image reconstruction, and thus effectively improves the accuracy of optoelectronic countermeasure technology.

[0006] To achieve the above object, the present invention provides a simulation method for long-distance near-infrared target turbulent imaging, including:

[0007] Based on the correlation of the two-pixel offset amounts when different two pixel points on the object-side near-infrared image are imaged to the same point on the image-side, obtain the x-axis offset amount and y-axis offset amount that can modulate all pixels of the near-infrared output image to produce the turbulent random offset effect;

[0008] Based on the x-axis offset amount and y-axis offset amount, process the turbulent phase covariance function to obtain the atmospheric turbulence phase matrix; based on the atmospheric turbulence phase matrix, obtain the point spread function that can modulate the output near-infrared image to produce the turbulent blur effect caused by high-order aberrations;

[0009] Based on the point spread function, obtain the atmospheric turbulence near-infrared image.

[0010] Optionally, obtaining the x-axis offset amount and y-axis offset amount includes:

[0011] Normalize the correlation of the offset amounts of different two pixels when they are imaged to the same point on the image-side on the object-side near-infrared image;

[0012] Sample the normalized offset amount correlation, and extract the x-axis offset amount and y-axis offset amount and apply them to the input object-side near-infrared image.

[0013] Optionally, the normalized offset amount correlation is:

[0014]

[0015] Wherein, represents the angle between the line connecting two pixels on the object-side near-infrared image and the imaging point and the normal, D is the aperture diameter of the receiving system, L is the transmission distance, λ is the optical wavelength, δ 0 is the object-side Nyquist pixel pitch, is the atmospheric coherence length, is the atmospheric refractive index structure constant υ is the abscissa in the form of the Bessel function, J ith (υ) represents the i-th of the first kind of Bessel function.

[0016] Optionally, there is a correlation between the turbulent phase covariance function and the atmospheric turbulence phase structure function:

[0017] F(s) = 2γ 2 -X(s)

[0018] where F(s) is the atmospheric turbulence phase structure function, γ 2 is the phase variance, and X(s) is the turbulent phase covariance function;

[0019] The atmospheric turbulence phase structure function following the Von-Karman model is expressed as:

[0020]

[0021] where s is the distance between two points in the phase screen, L 0 is the outer scale, r 0 is the atmospheric coherence length, Γ() is the gamma function, and K 5 / 6 (·) is the modified Bessel function of the third kind.

[0022] Optionally, obtaining the atmospheric turbulence phase matrix includes:

[0023] Removing the translation component, x-axis offset, and y-axis offset in the turbulent phase covariance function;

[0024] Obtaining the turbulent discrete power spectrum according to the processed turbulent phase covariance function;

[0025] Obtaining the atmospheric turbulence phase matrix according to the turbulent discrete power spectrum.

[0026] Optionally, the processed turbulent phase covariance function is:

[0027]

[0028] where P is the number of pixels in a row or column of the output image.

[0029] The turbulent discrete power spectrum is:

[0030]

[0031] where Φ(a′f',b′f') is the turbulent discrete power spectrum, a, b, a', b' are the number of sampling points, f' is the frequency domain sampling interval, x is the spatial domain sampling interval, and Q is the total number of sampling points;

[0032] The atmospheric turbulence phase matrix is:

[0033]

[0034] Among them, φ(ax,bx) is the atmospheric turbulence phase matrix, and h(a′,b′) is the random number matrix.

[0035] Optionally, obtaining the point spread function includes:

[0036] Dividing the atmospheric turbulence phase screen into several blocks, each block having a size of a preset number of pixels;

[0037] Performing an inverse Fourier transform on each divided phase screen with a size of a preset number of pixels to obtain a plurality of point spread functions with a size of a preset number of pixels.

[0038] Optionally, the point spread function is:

[0039]

[0040] Among them, x represents the position of the output image, P(u) is the pupil diameter function, |h(x)| 2 represents the point spread function.

[0041] Optionally, based on the point spread function, obtaining the near-infrared image of atmospheric turbulence includes:

[0042] Dividing the image with the turbulent random offset effect into several blocks, each block having a size of a preset number of pixels;

[0043] Convolving the several blocks, each with a size of a preset number of pixels of the point spread function and the near-infrared image with the turbulent random offset effect in a one-to-one correspondence to obtain a near-infrared image of atmospheric turbulence with a turbulent blur increment.

[0044] Optionally, the near-infrared image of atmospheric turbulence is:

[0045]

[0046] Among them represents the convolution operation, I(x) in represents the near-infrared image with the turbulent random offset effect, |h(x)| 2 represents the point spread function.

[0047] The present invention has the following beneficial effects:

[0048] The present invention adopts a method of directly calculating the turbulent random offset and blur that cause image distortion to simulate the imaging of a long-distance near-infrared image in atmospheric turbulence. According to the correlation of the offset amounts of two pixels when different two pixels on the object-side near-infrared image are imaged to the same point on the image-side, the turbulent random offset effect of all pixels of the near-infrared image is modulated and output. According to the fast Fourier transform method of calculating the discrete power spectrum based on the turbulent phase covariance function, the turbulent blur effect caused by high-order aberrations is modulated and output for the near-infrared image. This method does not require step-by-step propagation, has a fast calculation speed, conforms to the turbulent effect, and provides a basis for simulating the atmospheric turbulence imaging of a long-distance near-infrared target and generating a simulation dataset of atmospheric turbulence. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0050] Figure 1 It is a schematic flow chart of a simulation method for turbulent imaging of a long-distance near-infrared target according to an embodiment of the present invention;

[0051] Figure 2 It is a schematic diagram of an input long-distance near-infrared target image according to an embodiment of the present invention;

[0052] Figure 3 It is a schematic diagram of a long-distance near-infrared target with a turbulent random offset effect according to an embodiment of the present invention;

[0053] Figure 4 It is a schematic diagram of the atmospheric turbulence of an output long-distance near-infrared target according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail this application.

[0055] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that here.

[0056] As Figures 1 to 4 shown, this embodiment proposes a simulation method for turbulent imaging of a long-distance near-infrared target, including the following steps:

[0057] Step a. Set initial parameters;

[0058] Step b. According to the correlation of the two-pixel offsets when two different pixels on the object-side near-infrared image are imaged to the same point on the image-side, obtain the x-axis offset and y-axis offset that can modulate all pixels of the near-infrared output image to produce the turbulent random offset effect;

[0059] Step c. According to the fast Fourier transform method of calculating the discrete power spectrum based on the phase covariance function, obtain the atmospheric turbulence phase matrix;

[0060] Step d. According to the atmospheric turbulence phase matrix, obtain the point spread function that can modulate the output near-infrared image to produce the turbulent blur effect caused by higher-order aberrations; furthermore, obtain the blur amount. The blur amount is the blur on the overall image, which is the result of convolving the input image with the point spread function, and is the blur effect achieved step by step according to the text in the quotation marks. The blur can be described and represented by the point spread function. The specific blur amount is determined by the parameters actually given.

[0061] Step e. According to the point spread function, obtain the atmospheric turbulence near-infrared image.

[0062] The initial parameters in step a at least include the optical wavelength λ, the atmospheric refractive index structure constant the aperture size Δ′ of the receiving optical system, the transmission distance L, the object-side Nyquist pixel pitch δ 0 and the image-side Nyquist pixel pitch δ f , the number of image pixels P×P, the inner scale l 0 and the outer scale L 0 . In this embodiment, the optical wavelength λ is 1064 nm, the atmospheric refractive index structure constant is 7e-15, the aperture size Δ′ of the receiving optical system is 0.2 m, the transmission distance L is 7000 m, the object-side Nyquist pixel pitch δ 0 is 18.62 mm, the image-side Nyquist pixel pitch δ f is 2.66 μm, the number of image pixels is 256×256 pixels, the inner scale l 0 is 0.01 m, and the outer scale L 0 is 100 m.

[0063] Step b specifically includes:

[0064] Step 11: Normalize the correlation of the two-pixel offsets when two different pixels on the object-side near-infrared image are imaged to the same point on the image-side;

[0065] Step 12: Sample the offset correlation, and extract the x-axis offset and y-axis offset and apply them to the input near-infrared image.

[0066] In step 11, the normalized offset correlation can be expressed by the expectation as:

[0067]

[0068] Among them, is expressed as the angle between the line connecting two pixels on the object-side near-infrared image and the imaging point and the normal direction, D is the aperture diameter of the receiving system, L is the transmission distance, λ is the optical wavelength, and δ 0 is the object-side Nyquist pixel pitch, is the atmospheric coherence length, is the atmospheric refractive index structure constant, υ is the abscissa in the form of the Bessel function, and J ith (υ) represents the i-th of the Bessel functions of the first kind.

[0069] Step c specifically includes:

[0070] Step 21: Remove the translation component, x-axis offset, and y-axis offset in the turbulent phase covariance function;

[0071] Step 22: Obtain the turbulent discrete power spectrum according to the processed turbulent phase covariance function;

[0072] Step 23: Obtain the atmospheric turbulence phase matrix according to the turbulent discrete power spectrum.

[0073] In Step 21, there is a correlation between the turbulent phase covariance function and the atmospheric turbulence phase structure function:

[0074] F(s) = 2γ 2 -X(s)

[0075] where F(s) is the atmospheric turbulence phase structure function, γ 2 is the phase variance, and X(s) is the turbulent phase covariance function.

[0076] The atmospheric turbulence phase structure function following the Von-Karman model can be expressed as:

[0077]

[0078] where s is the distance between two points in the phase screen, L 0 is the outer scale, r 0 is the atmospheric coherence length, Γ() is the gamma function, and K 5 / 6 (·) is the modified Bessel function of the third kind.

[0079] In Step 21, the processed turbulent phase covariance function can be expressed as:

[0080]

[0081] Among them The number of pixels in a row or column of the P output image. In step 22, the discrete power spectrum can be expressed by the turbulent phase covariance function as:

[0082]

[0083] where a, b, a', b' are the number of sampling points, f' is the sampling interval in the frequency domain, x is the sampling interval in the spatial domain, and Q is the total number of sampling points.

[0084] In step 23, the atmospheric turbulence phase matrix can be expressed as:

[0085]

[0086] where h(a′, b′) is a random number matrix.

[0087] Step d specifically includes:

[0088] Step 31: Divide the atmospheric turbulence phase screen into 8×8 blocks, each block having a size of 32×32 pixels;

[0089] Step 32: Perform an inverse Fourier transform on each divided 32×32 pixel-sized phase screen to obtain 8×8 point spread functions of 32×32 pixels.

[0090] The point spread function in step 32 can be expressed as:

[0091]

[0092] where x represents the position of the output image, and P(u) is the pupil aperture function:

[0093] Step e specifically includes:

[0094] Step 41: Divide the image with the turbulent random offset effect into 8×8 blocks, each block being 32×32 pixels in size;

[0095] Step 42: Correlate and convolve the 8×8 blocks, each block being a 32×32 pixel-sized point spread function and the near-infrared image with the turbulent random offset effect to obtain an atmospheric turbulence near-infrared image with a turbulent blur increment.

[0096] In step 42, the output near-infrared turbulent image after convolution can be expressed as:

[0097]

[0098] where represents the convolution operation, and I(x) in represents the near-infrared image with the turbulent random offset effect.

[0099] Figure 1 is a flowchart of an embodiment; Figure 2 is to input a long-distance near-infrared target image, Figure 3 and Figure 4 is a long-distance near-infrared target image with a turbulent random offset effect and an atmospheric turbulence simulation image of the output long-distance near-infrared target obtained according to the above formula.

[0100] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A simulation method for long-distance near-infrared target turbulence imaging, characterized in that: include: Based on the correlation of the two-pixel offset when two different pixel points on the object-side near-infrared image are imaged to the same point on the image-side, the x-axis offset and y-axis offset of all pixels of the modulated near-infrared output image to produce a turbulent random offset effect are obtained; Based on the x-axis offset and the y-axis offset, the turbulence phase covariance function is processed to obtain an atmospheric turbulence phase matrix; based on the atmospheric turbulence phase matrix, a point spread function that can modulate the output near-infrared image to produce a turbulence blurring effect caused by high-order aberrations is obtained; Based on the point spread function, an atmospheric turbulence near-infrared image is acquired.

2. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 1, characterized in that: Obtaining the x-axis offset and the y-axis offset includes: Normalize the correlation of the offsets of two pixels when two different pixel points on the object-side near-infrared image are imaged to the same point on the image-side; The normalized offset correlation is sampled, and the x-axis offset and the y-axis offset are extracted and applied to the object-side near-infrared image.

3. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 2, characterized in that: The normalized offset correlation is: in, It is expressed as the angle between the line connecting two pixels and the imaging point on the object-side near-infrared image and the normal direction. D is the aperture diameter of the receiving system, L is the transmission distance, λ is the wavelength of light, δ0 is the object-side Nyquist pixel spacing, is the atmospheric coherence length, is the atmospheric refractive index structure constant, υ is the abscissa in the form of Bessel function, J ith (υ) denotes the i-th Bessel function of the first kind.

4. The method for simulating turbulence imaging of long-range near-infrared targets according to claim 1, characterized in that: The turbulence phase covariance function is related to the atmospheric turbulence phase structure function: F(s)=2γ 2 -X(s) Where F(s) is the atmospheric turbulence phase structure function, γ 2 is the phase variance, X(s) is the turbulence phase covariance function; The atmospheric turbulence phase structure function following the Von-Karman model is expressed as: Where s is the distance between two points in the phase screen, L0 is the outer scale, r0 is the atmospheric coherence length, Γ() is the gamma function, K 5 / 6 (·) is a modified Bessel function of the third kind.

5. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 1, characterized in that: Acquiring the atmospheric turbulence phase matrix comprises: Remove the translation component, x-axis offset and y-axis offset in the turbulence phase covariance function; According to the processed turbulence phase covariance function, a turbulence discrete power spectrum is obtained; The atmospheric turbulence phase matrix is ​​obtained according to the turbulence discrete power spectrum.

6. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 5, characterized in that: The turbulence phase covariance function after the processing is: in P is the number of rows or columns of pixels in the output image; The turbulence discrete power spectrum is: Where Φ(a′f',b′f') is the turbulence discrete power spectrum, a, b, a', b' are the number of sampling points, f' is the frequency domain sampling interval, x is the spatial domain sampling interval, and Q is the total number of sampling points; The atmospheric turbulence phase matrix is: Among them, φ(ax,bx) is the atmospheric turbulence phase matrix, and h(a′,b′) is the random number matrix.

7. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 1, characterized in that: Obtaining the point spread function includes: Divide the atmospheric turbulence phase screen into a number of blocks, each block having a size of a preset number of pixels; An inverse Fourier transform is performed on each phase screen divided into a preset number of pixels to obtain a plurality of point spread functions of the preset number of pixels.

8. The method for simulating turbulence imaging of long-range near-infrared targets according to claim 1, characterized in that: The point spread function is: Where x represents the position of the output image, P(u) is the pupil aperture function, h(x) 2 represents the point spread function.

9. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 1, characterized in that: Based on the point spread function, obtaining an atmospheric turbulence near-infrared image comprises: Divide the image with turbulent random offset effect into several blocks, each block is a preset number of pixels in size; A plurality of blocks, each of which is a point spread function of a preset number of pixels, and a near-infrared image containing a turbulence random offset effect are corresponded and convolved to obtain an atmospheric turbulence near-infrared image with a turbulence blur increment.

10. The method for simulating turbulent imaging of long-range near-infrared targets according to claim 1, characterized in that: The atmospheric turbulence near-infrared image is: in represents the convolution operation, I(x) in represents a near-infrared image with turbulent random offset effect, h(x) 2 represents the point spread function.

Images