A method and system for evaluating restoration of remote sensing images in orbit based on stationary orbit

Abstract

The present application relates to a kind of on-orbit remote sensing image restoration evaluation method and system based on stationary orbit, the method applicable system successively includes: integrating sphere, edge image target, parallel light tube, space camera principle prototype, imaging processor and fast view equipment;The imaging processor is respectively with space camera principle prototype and fast view equipment between being equipped with the cable of connection;This method includes the following steps: step one, image acquisition;Step two, calculate the transmission function of Nyquist frequency;Step three, calculate signal-to-noise ratio;Step four, fusion evaluation.The present application is based on stationary orbit's on-orbit remote sensing image restoration evaluation method and system, for the face array camera of stationary orbit gaze imaging, when on-orbit restoration, the contradiction that the transmission function of Nyquist frequency increases while signal-to-noise ratio drops, and the limitation of existing image restoration evaluation method, can be fused according to overall index requirement, the index of transmission function of Nyquist frequency and signal-to-noise ratio is evaluated, effectively guide the improvement of restoration algorithm.

Description

Technical Field

[0001] This invention relates to the field of remote sensing image processing and remote sensing image quality assessment, specifically to an on-orbit remote sensing image restoration and evaluation method and system based on geostationary orbit. Background Technology

[0002] With the continuous development of space remote sensing technology, space remote sensing imaging systems have been widely used in commercial, military, and civilian applications. The demands on the observation capabilities and functions of optical remote sensors are constantly increasing, leading space remote sensing imaging systems to develop towards higher resolution, lower cost, and greater clarity. However, affected by factors such as satellite platform vibration, imaging environment, and defocusing, the quality of remote sensing images degrades significantly, resulting in blurred images, low clarity and contrast, and substantial loss of detail, posing considerable difficulties for target analysis and interpretation. Remote sensing image restoration technology can effectively address this challenge, but the evaluation of the restoration quality directly impacts the restoration effect. Furthermore, effective evaluation of on-orbit remote sensing image restoration results can guide the improvement of image restoration algorithms.

[0003] In the field of image processing, metrics such as peak signal-to-noise ratio (PSNR), information entropy, and contrast are generally used to evaluate the image restoration effect. However, using this image processing evaluation method to evaluate the quality of on-orbit remote sensing images has limitations and cannot accurately reflect the quality attributes of on-orbit images. The generally accepted evaluation criteria for on-orbit remote sensing image quality typically use the Nyquist frequency modulation transfer function (Nyquist transfer function) and signal-to-noise ratio (SNR) as two key indicators. However, after on-orbit image restoration, the Nyquist transfer function generally increases, while the SNR decreases. This contradiction between the two indicators complicates the restoration evaluation process. Summary of the Invention

[0004] The present invention aims to solve the technical problems in the prior art by providing an on-orbit remote sensing image restoration and evaluation method and system based on geostationary orbit.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] A method for evaluating the restoration of in-orbit remote sensing images based on geostationary orbit includes the following steps:

[0007] Step 1: Acquire images;

[0008] Step 2: Calculate the niff transfer function;

[0009] Step 3: Calculate the signal-to-noise ratio;

[0010] Step 4: Integration and Evaluation.

[0011] In the above technical solution, step one specifically includes:

[0012] The image restoration function of the imaging processor is turned off, and raw images H are continuously acquired and stored, namely I1, I2...I... K K = H;

[0013] Activate the image restoration function of the imaging processor, continuously acquire and store H restored images, namely J1, J2...J L L = H.

[0014] In the above technical solution, step two specifically includes:

[0015] Using the edge image, the naphth transfer function of the original image for a specified region is calculated using the edge method. The naphth transfer function of a single original image is MTF. K If K∈[1,H], then the average Navigational Transfer Function of the original image is:

[0016]

[0017] The Nyquist transfer function of the restored single-frame image is MTF. L If L∈[1,H], then the average Navigational transfer function of the restored image is:

[0018]

[0019] The boost in the naphth transfer function of the restored image is expressed by the following formula:

[0020] δ mtf =(J mtf -I mtf ) / I mtf .

[0021] In the above technical solution, step three specifically includes:

[0022] Select H original images and H restored images corresponding to uniform regions, each with a size of m×n; obtain the signal-to-noise ratio of each pixel in the original image of the corresponding region as:

[0023]

[0024] Where average(I) K (i,j)) represents the mean DN value of the pixel at (i,j) in H original images, std(I K (i,j)) represents the mean square error of the DN value of the pixel at (i,j) in the restored H-image;

[0025] The signal-to-noise ratio of each pixel in the restored image of the corresponding region is:

[0026]

[0027] Where average(J) L (i,j)) represents the mean DN value of the pixel at (i,j) in the H-image restored image, std(J L (i,j)) represents the mean square error of the DN value of the pixel at (i,j) in the restored H-image;

[0028] The decrease in signal-to-noise ratio of the restored image can be expressed by the following formula:

[0029]

[0030] In the above technical solution, step four specifically includes:

[0031] Enable the image restoration function, acquire H original images, and calculate the signal-to-noise ratio (SNR) of the image by selecting a uniform region. The SNR calculation is as follows:

[0032] δ=ηδ mtf -γδ snr ;

[0033] Where η and γ correspond to the weights for improving the Navier-to-Navier-to-Transmission Function and decreasing the Signal-to-Noise Ratio, respectively.

[0034] In the above technical solution, the weights for the upscaling of the naphth function and the decrease in the signal-to-noise ratio satisfy the following:

[0035] When the improvement of the Navier-to-frequency transfer function and the decrease in the signal-to-noise ratio are considered to be of equal importance, η = γ;

[0036] When it is believed that the increase in attention to the Nyquist-transfer function is greater than the decrease in the signal-to-noise ratio, η < γ;

[0037] When it is considered that the increase in attention to the Nyquist-transfer function is less than the decrease in the signal-to-noise ratio, η > γ;

[0038] When the improvement in the Navier-to-Nutrition Function (NJF) is comparable to the decrease in the Signal-to-Noise Ratio (SNR), δ = 0.

[0039] When the improvement in Navier-to-Navier-to-transfer function is more significant than the decrease in signal-to-noise ratio, δ > 0;

[0040] When the decrease in signal-to-noise ratio is more significant than the increase in Navier-Frequency transfer function, δ < 0.

[0041] A system to which the above-mentioned geostationary orbit-based on-orbit remote sensing image restoration and evaluation method is applicable includes: an integrating sphere, an edge image target, a collimator, a space camera prototype, an imaging processor, and a fast-viewing device; the imaging processor is connected to the space camera prototype and the fast-viewing device by cables respectively.

[0042] The edge image target is used to calculate the Navigational Transfer Function of the original image of a specified region using the edge image and the edge method.

[0043] The imaging processor is used to continuously acquire and store the original image when the image restoration function is turned off and / or on, and to select a uniform region of the image to calculate the signal-to-noise ratio of the image;

[0044] The integrating sphere is used to provide a uniform light source;

[0045] The collimator is used to convert the beam output from the edge image target into a parallel beam, which is then incident on the focal plane detector of the space camera prototype.

[0046] The space camera prototype is used for imaging;

[0047] The fast-viewing device is used to display and store images, providing data for subsequent image processing and analysis.

[0048] The present invention has the following beneficial effects:

[0049] The present invention provides an on-orbit remote sensing image restoration evaluation method and system based on geostationary orbit. Addressing the contradiction that the naphth transfer function increases while the signal-to-noise ratio decreases during on-orbit restoration of area array cameras using geostationary orbit staring imaging, as well as the limitations of existing image restoration evaluation methods, the present invention can integrate the two indicators, naphth transfer function and signal-to-noise ratio, for evaluation based on overall performance requirements, thereby effectively guiding the improvement of restoration algorithms. Attached Figure Description

[0050] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0051] Figure 1 This is a schematic diagram of the system to which the on-orbit remote sensing image restoration and evaluation method based on a geostationary orbit of the present invention is applicable.

[0052] The reference numerals in the figure are:

[0053] 1-Integrating sphere; 2-Edge image target; 3-Columnar tube; 4-Space camera prototype; 5-Imaging processor; 6-Fast-viewing device; 7-Cable. Detailed Implementation

[0054] The present invention will now be described in detail with reference to the accompanying drawings.

[0055] The geostationary orbit-based on-orbit remote sensing image restoration and evaluation method of the present invention is applicable to a ground-based simulation platform for geostationary orbit remote sensing images, the system of which is as follows: Figure 1As shown, the system sequentially includes: an integrating sphere 1, an edge image target 2, a collimator 3, a space camera prototype 4, an imaging processor 5, and a fast-viewing device 6; it also includes cables 7 connecting the imaging processor 5 to the space camera prototype 4 and the fast-viewing device 6. The edge image target 2 is used to calculate the Navigation Function (NfF) of the original image of a specified area using the edge image and the edge method; the imaging processor 5 is used to continuously acquire and store the original image when the image restoration function is off and / or on, and to select a uniform region of the image to calculate the signal-to-noise ratio; the integrating sphere 1 provides a uniform light source; the collimator 3 converts the beam output from the edge image target 2 into a parallel beam, which is then incident on the focal plane detector of the space camera prototype 4; the space camera prototype 4 is used for imaging; and the fast-viewing device 6 is used to display and store the image, providing data for subsequent image processing and analysis. Because the area array camera in a geostationary orbit is a staring imaging system, it does not require image registration, thus achieving high matching accuracy for corresponding areas of the image during continuous imaging.

[0056] The on-orbit remote sensing image restoration and evaluation method based on geostationary orbit of the present invention includes the following steps:

[0057] I. Image Acquisition

[0058] The image restoration function of imaging processor 5 is turned off, and H original images are continuously acquired and stored, namely I1, I2...I... K K = H, activate the image restoration function of imaging processor 5, continuously acquire and store H restored images, namely J1, J2...J L L = H.

[0059] II. Calculating the Navier-Frequency Transfer Function

[0060] 1. Using the edge image, calculate the naphth transfer function of the original image of a specified area using the edge method. The naphth transfer function of a single original image is MTF. K If K∈[1,H], then the average Navigational Transfer Function of the original image is:

[0061]

[0062] 2. Similarly, the Nyqubit transfer function of a single restored image is MTF. L If L∈[1,H], then the average Navigational transfer function of the restored image is:

[0063]

[0064] 3. The boost in the naphth transfer function of the restored image is expressed by the following formula:

[0065] δ mtf =(Jmtf -I mtf ) / I mtf (3)

[0066] III. Calculating the signal-to-noise ratio

[0067] 1. Select a uniform region of size m×n corresponding to H original images and H restored images. The signal-to-noise ratio of each pixel in the original image of the corresponding region can be obtained as:

[0068]

[0069] Where average(I) K (i,j)) represents the mean DN value of the pixel at (i,j) in H original images, std(I K (i,j)) represents the mean square error of the DN value of the pixel at (i,j) in the restored H-image.

[0070] 2. The signal-to-noise ratio of each pixel in the restored image of the corresponding region is:

[0071]

[0072] Where average(J) L (i,j)) represents the mean DN value of the pixel at (i,j) in the H-image restored image, std(J L (i,j)) represents the mean square error of the DN value of the pixel at (i,j) in the restored H-image.

[0073] 3. The decrease in signal-to-noise ratio of the restored image can be expressed by the following formula:

[0074]

[0075] IV. Integration Evaluation

[0076] Enable the image restoration function, acquire H original images, and calculate the signal-to-noise ratio (SNR) of the image by selecting a uniform region. The SNR calculation is as follows:

[0077] δ=ηδ mtf -γδ snr (7)

[0078] Where η and γ correspond to the weights for improving the Navier-Stokes transfer function and decreasing the signal-to-noise ratio.

[0079] The weight can be adjusted according to the overall indicator requirements. When the importance of improving the NAND flash function and decreasing the signal-to-noise ratio is considered to be equal, η = γ; when the importance of improving the NAND flash function is considered to be greater than decreasing the signal-to-noise ratio, η < γ; otherwise, η > γ.

[0080] When the improvement in Navier-to-Navier-to-Transfer Function (NAV) is comparable to the decrease in Signal-to-Noise Ratio (SNR), δ = 0; when the improvement in NAV is more significant than the decrease in SNR, δ > 0; otherwise, δ < 0.

[0081] The present invention provides an on-orbit remote sensing image restoration evaluation method and system based on geostationary orbit. Addressing the contradiction that the naphth transfer function increases while the signal-to-noise ratio decreases during on-orbit restoration of area array cameras using geostationary orbit staring imaging, as well as the limitations of existing image restoration evaluation methods, the present invention can integrate the two indicators, naphth transfer function and signal-to-noise ratio, for evaluation based on overall performance requirements, thereby effectively guiding the improvement of restoration algorithms.

[0082] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for evaluating the restoration of on-orbit remote sensing images based on geostationary orbit, characterized in that, The applicable system includes, in sequence along the optical path: an integrating sphere (1), an edge image target (2), a collimator (3), a space camera prototype (4), an imaging processor (5), and a fast-viewing device (6); the imaging processor (5) is connected to the space camera prototype (4) and the fast-viewing device (6) by cables (7). The imaging processor (5) is used to continuously acquire the original image and the restored image when the image restoration function is turned off and / or on; The integrating sphere (1) is used to provide a uniform light source; The collimator (3) is used to convert the beam output from the edge image target (2) into a parallel beam and incident it onto the focal plane detector of the space camera prototype (4). The prototype space camera (4) is used for imaging; The fast-viewing device (6) is used to display and store images, providing data for subsequent image processing and analysis; The evaluation method includes the following steps: Step 1: Acquire images; Step one specifically includes: With image restoration function disabled, H original images were continuously acquired, respectively. ; Turn on the image restoration function and continuously acquire H restored images, respectively. ; Step 2: Calculate the niff transfer function; Step two specifically includes: Using the edge image, the naphth transfer function of the original image for a specified region is calculated using the edge method. The naphth transfer function of a single original image is: , The average Navigational Transfer Function of the original image is: The Navier-Frequency transfer function of the restored single-image image is , The average Navigational Transfer Function of the restored image is: The boost in the naphth transfer function of the restored image is expressed by the following formula: ； Step 3: Calculate the signal-to-noise ratio; Step three specifically includes: Select H original images and H restored images corresponding to uniform regions, each with a size of m×n; obtain the signal-to-noise ratio of each pixel in the original image of the corresponding region as: in H original images The mean DN value of the pixel at that location. H original images The mean square error of the DN value of the pixel at that location; The signal-to-noise ratio of each pixel in the restored image of the corresponding region is: in Image after H-frame restoration The mean DN value of the pixel at that location. Image after H-frame restoration The mean square error of the DN value of the pixel at that location; The decrease in signal-to-noise ratio of the restored image can be expressed by the following formula: ； Step 4: Integration and Evaluation; Step four specifically includes: increasing the napyr function of the restored image calculated in step two. The signal-to-noise ratio of the restored image calculated in step three decreases. The fusion evaluation is performed using the following formula: ； in, and These correspond to the weights for improving the frequency transfer function and decreasing the signal-to-noise ratio, respectively.

2. The on-orbit remote sensing image restoration and evaluation method based on geostationary orbit according to claim 1, characterized in that, The weights for the upscaling of the naphth function and the decrease in the signal-to-noise ratio satisfy the following: When the improvement in Navier-Stokes transfer function and the decrease in signal-to-noise ratio are considered to be of equal importance... ; When it is believed that the increase in NAND flash memory is more significant than the decrease in signal-to-noise ratio, ; When it is believed that the increase in NAND flash memory is less significant than the decrease in signal-to-noise ratio, ; When the improvement in Navier-to-Navier-to-Transfer Function (NRF) is comparable to the decrease in Signal-to-Noise Ratio (SNR). ; When the improvement in Navier-to-transfer function is more significant than the decrease in signal-to-noise ratio. ; When the decrease in signal-to-noise ratio is more significant than the increase in Navier-Stokes transfer function. .

Images