A method for distortion correction of images captured by an intelligent recognition SWIR 6-mm lens

Abstract

The present invention discloses a distortion correction method for images of an intelligent recognition SWIR 6-mm lens, including: S1, obtaining original image data through the SWIR 6-mm lens; S2, calculating a distortion compensation factor in the original image data according to the optical system characteristics of the SWIR 6-mm lens; S3, using the distortion compensation factor to correct geometric distortion and obtaining geometrically corrected image data; S4, performing geometric correction on the edge region of the geometrically corrected image data according to the optical system characteristics of the SWIR 6-mm lens to obtain corrected image data; S5, applying an adaptive denoising algorithm to the corrected image data, dynamically adjusting the noise suppression intensity according to the distribution of optical distortion, and performing detail enhancement processing; S6, outputting the image data after distortion correction, denoising, and detail enhancement. The present invention has the advantages of good detail retention and high-efficiency real-time processing ability, etc.

Description

Technical Field

[0001] The present invention relates to the technical field of image correction, and particularly to a method for correcting the distortion of an image of an intelligent recognition SWIR 6-mm lens. Background Art

[0002] With the wide application of short-wave infrared (SWIR) imaging technology, especially in the fields of aerospace, military reconnaissance, security monitoring, and industrial inspection, the advantages of SWIR lenses in low-light and complex environments have been increasingly emphasized. However, with the improvement of requirements, traditional SWIR lens technology often faces the problem of image distortion during high-resolution and large-field-of-view imaging. Especially under the conditions of a large field of view and a short focal length, the distortion problem is more serious. In particular, the geometric distortion and aberration in the edge region seriously affect the imaging quality and the accuracy of the image.

[0003] In the prior art, although some methods have been proposed to reduce and correct the geometric distortion problem of SWIR lenses, these methods still have many defects and limitations and cannot meet the challenges of higher performance requirements. Specifically, the existing methods have obvious deficiencies in the following aspects:

[0004] 1. Insufficient accuracy of distortion correction: Traditional methods mainly reduce distortion through physical optical design optimization. Although some distortions can be effectively reduced, under the conditions of a large field of view and a short focal length, the geometric distortion and aberration in the edge region are still difficult to completely eliminate, and the imaging accuracy is difficult to guarantee.

[0005] 2. Poor adaptability of denoising algorithms: Existing noise suppression methods usually rely on global noise estimation to suppress high-frequency noise in images. However, these methods lack adaptive adjustment for local regions and distortion compensation regions in images, are difficult to cope with complex image content, have unsatisfactory denoising effects, and may affect the retention of details.

[0006] 3. Insufficient detail enhancement ability: Traditional methods mainly rely on simple image processing techniques for denoising and restoration, lack the enhancement of local details, resulting in serious loss of details in the denoised image, especially in the high-contrast region of the image, affecting the quality of the image and the accuracy of information.

[0007] 4. Slow processing speed: Many existing image distortion correction and noise suppression methods rely on cumbersome calculation steps, resulting in insufficient real-time processing ability, unable to meet the requirements of fast response and efficient processing, and limiting their wide use in practical applications.

[0008] Therefore, how to provide a method for correcting the distortion of an image of an intelligent recognition SWIR 6-mm lens is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0009] An object of the present invention is to provide a method for distortion correction of images of a smart-identifying SWIR 6-mm lens. By dynamically adjusting the distortion compensation factor, using a weighted average algorithm to correct pixel coordinates, applying an adaptive denoising algorithm to process noise, and combining detail enhancement techniques, this method effectively solves the problems of insufficient distortion correction accuracy, incomplete noise suppression, and detail loss during the imaging process of the SWIR lens, and has the advantages of high-precision image correction ability, excellent noise removal effect, good detail retention, and efficient real-time processing ability.

[0010] A method for distortion correction of images of a smart-identifying SWIR 6-mm lens according to an embodiment of the present invention includes the following steps:

[0011] S1. Obtain original image data through a SWIR 6-mm lens, and perform distortion analysis on the obtained original image data to identify geometric distortion in the original image data;

[0012] S2. According to the optical system characteristics of the SWIR 6-mm lens, including focal length, aperture, and optical system configuration, calculate the distortion compensation factor in the original image data, establish a mathematical model for distortion correction, and apply the mathematical model to the original image data to generate corrected image data;

[0013] S3. Use the distortion compensation factor to adjust the pixel coordinates in the corrected image data by using a geometric transformation algorithm to correct geometric distortion and obtain geometrically corrected image data;

[0014] S4. According to the optical system characteristics of the SWIR 6-mm lens, perform geometric correction on the edge region of the geometrically corrected image data to reduce geometric distortion caused by the edge effect of the optical system and obtain corrected image data;

[0015] S5. Apply an adaptive denoising algorithm to the corrected image data, dynamically adjust the noise suppression intensity according to the distribution of optical distortion, eliminate noise introduced by correction, and perform detail enhancement processing;

[0016] S6. Output the image data after distortion correction, denoising, and detail enhancement.

[0017] Optionally, the S1 specifically includes:

[0018] S11. Obtain original image data through a SWIR 6-mm lens. The original image data is a short-wave infrared image in the wavelength band of 900 nm to 1700 nm, and the original image data is in the form of a two-dimensional matrix, including pixel position and gray value information;

[0019] S12. Preprocess the acquired original image data, including cropping, image scaling, and grayscale value normalization, to obtain standard image data;

[0020] S13. According to the optical design parameters of the SWIR 6-mm lens, establish an optical distortion model and perform geometric distortion analysis on the preprocessed standard image data, including radial distortion analysis and tangential distortion analysis;

[0021] Radial distortion analysis:

[0022] ;

[0023] where, is the radial distortion value, is the distance from the pixel point in the standard image data to the optical axis, , , are the radial distortion coefficients;

[0024] Tangential distortion analysis:

[0025] ;

[0026] where, is the tangential distortion value, , is the coordinate of the pixel point in the standard image data, , , are the tangential distortion coefficients;

[0027] S14. Adopt a feature point matching algorithm to identify and extract the pixel positions corresponding to the feature points of the standard image data in the original image data, and quantify the geometric distortion in the original image data by comparing their position deviations.

[0028] Optionally, the specific steps of S2 include:

[0029] S21. According to the optical system characteristics of the SWIR 6-mm lens, obtain the focal length, aperture, lens curvature, refractive index of the optical element, and the relative positions between the optical elements;

[0030] S22. According to the focal length, aperture, lens curvature, and refractive index of the optical element, establish the optical transfer function of the lens:

[0031] ;

[0032] where, is the optical transfer function corresponding to the spatial frequency in the frequency domain, is the cut-off frequency of the lens, is the attenuation index of the lens, is the spatial frequency in the frequency domain of the standard image data;

[0033] S23. Calculate the distortion compensation factor in the original image data based on the optical transfer function , and the distortion compensation factor is used to correct the geometric distortion in the original image data:

[0034] ;

[0035] Among them, is the distortion compensation factor, is the optical transfer function, is the distance from the pixel point in the standard image data to the optical axis, , , is the radial distortion coefficient;

[0036] S24. Establish a mathematical model for distortion correction based on the calculated distortion compensation factor, combined with the optical transfer function of the SWIR6 mm lens and the geometric characteristics of the original image data:

[0037] ;

[0038] Among them, is the corrected image data, is the original image data, is the distortion compensation factor;

[0039] S25. Apply the mathematical model for distortion correction to the original image data to correct the geometric distortion of the original image data and generate the corrected image data.

[0040] Optionally, the specific steps of S3 include:

[0041] S31. Use the distortion compensation factor to adjust the pixel coordinates in the corrected image data. The adjustment operation includes transforming the coordinates of each pixel point in the image data to correct the geometric distortion;

[0042] S32. According to the distortion compensation factor and the original pixel coordinates of the corrected image data, establish a pixel coordinate transformation formula through a geometric transformation algorithm:

[0043] ;

[0044] ;

[0045] Among them, and is the pixel coordinate after geometric correction, and is the original pixel coordinate of the corrected image data, and is the pixel coordinate adjustment amount, is the distortion compensation factor;

[0046] S33. According to the distortion compensation factor of each pixel point in the corrected image data, use the weighted average algorithm to calculate the pixel coordinate adjustment amount of each pixel point:

[0047] ;

[0048] ;

[0049] Among them, is the weighting factor:

[0050] ;

[0051] Among them, and are the coordinates of adjacent pixel points, is the distortion compensation factor of adjacent pixel points, is a small constant used to avoid the weighting factor being zero;

[0052] S34. Through the coordinate transformation formula and the weighted average algorithm, adjust all pixel points in the image data to eliminate geometric distortion;

[0053] S35. According to the adjusted pixel coordinates, use the bilinear interpolation algorithm to recalculate the corrected image data;

[0054] S36. Output the image data after geometric correction.

[0055] Optionally, the specific steps of S4 include:

[0056] S41. According to the optical system characteristics of the SWIR 6 mm lens, identify the edge area in the image data after geometric correction. The edge area is the area within 10 to 20 pixels from the image edge in the image data after geometric correction;

[0057] S42. According to the geometric characteristics of the image data after geometric correction, calculate the distortion compensation factor of the edge area. The distortion compensation factor of the edge area is used to correct the geometric distortion caused by the edge effect of the optical system:

[0058] ;

[0059] Among them, is the distortion compensation factor for the edge region, and are the original pixel coordinates of the corrected image data, is the distance from the pixel point in the standard image data to the optical axis, is the coefficient for controlling the compensation intensity;

[0060] S43. According to the distortion compensation factor of the edge region, use the geometric transformation algorithm to adjust the pixel coordinates of the edge region;

[0061] S44. According to the adjusted pixel coordinates, use the bilinear interpolation method to recalculate the pixel values of the edge region;

[0062] S45. Correct all the pixels in the entire image to eliminate the geometric distortion caused by the edge effect of the optical system;

[0063] S46. Output the image data after geometric correction and edge region correction.

[0064] Optionally, the specific steps of S5 include:

[0065] S51. Perform noise analysis on the corrected image data, and the noise analysis includes identifying the noise introduced during the distortion correction process in the image;

[0066] S52. Based on the local optical distortion distribution of the corrected image data, dynamically adjust the noise suppression intensity in the denoising algorithm :

[0067] ;

[0068] Among them, is the noise suppression intensity, is the adjustment factor, is the distortion compensation factor for each pixel point;

[0069] S53. Apply the adaptive denoising algorithm based on the adjusted noise suppression intensity to suppress the noise in the corrected image data through local noise estimation, and obtain the denoised image data;

[0070] S54. After the denoising process, perform detail enhancement according to the local detail information of the denoised image data, and use the enhancement coefficient to adjust the texture details of the image:

[0071] ;

[0072] Among them, is the image data after detail enhancement, is the gradient value of the image data at the pixel point, is the detail enhancement coefficient;

[0073] S55. Output the image data after denoising and detail enhancement processing.

[0074] Optionally, the S53 specifically includes:

[0075] S531. Apply an adaptive denoising algorithm based on the adjusted noise suppression intensity The denoising algorithm includes denoising the corrected image data through a wavelet transform method;

[0076] S532. Use wavelet transform to decompose the image data, and obtain the low-frequency sub-band and high-frequency sub-band by low-pass filtering and high-pass filtering the corrected image data respectively: :

[0077] ;

[0078] ;

[0079] Among them, and are the low-pass filtering and high-pass filtering operations respectively;

[0080] S533. According to the noise suppression intensity , dynamically adjust the noise coefficient in the high-frequency sub-band :

[0081] ;

[0082] Among them, is the original high-frequency sub-band, is the noise suppression intensity;

[0083] S534. Reconstruct the adjusted low-frequency and high-frequency sub-bands through inverse wavelet transform to obtain the denoised image data :

[0084] ;

[0085] Among them, is the denoised image data;

[0086] S535. Output the denoised image data.

[0087] The beneficial effects of the present invention are:

[0088] (1) By combining a dynamic distortion compensation factor, an adaptive denoising algorithm, a weighted average algorithm, and a detail enhancement technique, the present invention successfully solves the problems of geometric distortion, noise suppression, and detail loss in SWIR 6-mm lens images, thereby significantly improving the quality and accuracy of the images. This method can effectively eliminate edge distortion in large field-of-view and short focal length imaging, ensuring high resolution and accuracy of the images, while maintaining the details and information accuracy of the images.

[0089] (2) By adaptively adjusting the noise suppression intensity and applying a wavelet transform denoising method, the present invention precisely removes the noise caused by geometric correction, especially high-frequency noise, significantly improving the quality of the image data, especially under low light conditions. This method retains more image details after denoising and further enhances the usability of the image through detail enhancement, being particularly suitable for SWIR imaging in high resolution and complex environments.

[0090] (3) By combining multiple processing steps of geometric correction and adaptive denoising, the present invention provides a comprehensive optimization of the imaging quality of SWIR lenses, significantly improving the processing speed and real-time response ability. This enables the method to be widely applied to real-time image processing tasks under high-precision imaging requirements, especially in the fields of aerospace, security monitoring, and industrial inspection, with the advantages of high efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0092] Figure 1 is the overall flowchart of a distortion correction method for an image of an intelligent recognition SWIR 6-mm lens proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0093] Now, the present invention will be further described in detail with reference to the drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0094] Refer to Figure 1 , a distortion correction method for an image of an intelligent recognition SWIR 6-mm lens, comprising the following steps:

[0095] S1. Obtain the original image data through a SWIR 6-mm lens, and perform distortion analysis on the obtained original image data to identify the geometric distortion in the original image data;

[0096] In this embodiment, S1 specifically includes:

[0097] S11. Obtain the original image data through a SWIR 6-mm lens. The original image data is a short-wave infrared image in the wavelength band of 900 nm to 1700 nm, and the original image data is in the form of a two-dimensional matrix, containing pixel position and gray value information;

[0098] S12. Preprocess the obtained original image data, including cropping, image scaling, and gray value normalization, to obtain standard image data;

[0099] S13. According to the optical design parameters of the SWIR 6-mm lens, establish an optical distortion model, and perform geometric distortion analysis on the preprocessed standard image data, including radial distortion analysis and tangential distortion analysis;

[0100] Radial distortion analysis:

[0101] ;

[0102] Among them, is the radial distortion value, is the distance from the pixel point in the standard image data to the optical axis, , , are the radial distortion coefficients;

[0103] Tangential distortion analysis:

[0104] ;

[0105] Among them, is the tangential distortion value, , is the coordinate of the pixel point in the standard image data, , , are the tangential distortion coefficients;

[0106] S14. Adopt a feature point matching algorithm to identify and extract the pixel positions corresponding to the feature points of the standard image data in the original image data, and quantify the geometric distortion in the original image data by comparing their position deviations.

[0107] In this embodiment, the original image data is obtained by using a SWIR 6-mm lens, and the distortion analysis is performed on the obtained original image data, and the geometric distortion problem in the image is successfully identified. This method can accurately identify the distortion features in the image, especially for the edge distortion caused by large field of view angle and short focal length imaging, which greatly improves the accuracy of distortion detection.

[0108] S2. According to the optical system characteristics of the SWIR 6-mm lens, including focal length, aperture, and optical system configuration, calculate the distortion compensation factor in the original image data, establish a mathematical model for distortion correction, and apply the mathematical model to the original image data to generate corrected image data;

[0109] In this embodiment, S2 specifically includes:

[0110] S21. According to the optical system characteristics of the SWIR 6-mm lens, obtain the focal length, aperture, lens curvature, refractive index of the optical elements, and relative positions between the optical elements;

[0111] S22. According to the focal length, aperture, lens curvature, and refractive index of the optical elements, establish the optical transfer function of the lens:

[0112] ;

[0113] Among them, is the optical transfer function corresponding to the spatial frequency in the frequency domain, is the cut-off frequency of the lens, is the attenuation index of the lens, is the spatial frequency in the frequency domain of the standard image data;

[0114] S23. Based on the optical transfer function, calculate the distortion compensation factor in the original image data, and the distortion compensation factor is used to correct the geometric distortion in the original image data:

[0115] ;

[0116] Among them, is the distortion compensation factor, is the optical transfer function, is the distance from the pixel point in the standard image data to the optical axis, , , is the radial distortion coefficient;

[0117] S24. According to the calculated distortion compensation factor, combine the optical transfer function of the SWIR 6-mm lens with the geometric characteristics of the original image data to establish a mathematical model for distortion correction:

[0118] ;

[0119] Among them, is the corrected image data, is the original image data, is the distortion compensation factor;

[0120] S25. Apply the mathematical model for distortion correction to the original image data, and correct the geometric distortion of the original image data to generate corrected image data.

[0121] In this embodiment, by combining the optical system characteristics of the SWIR 6-mm lens, the distortion compensation factor in the original image data is calculated, and a mathematical model for distortion correction is established. This method can accurately quantify the geometric distortion in the image according to the optical characteristics of the lens, and effectively correct the original image data through the mathematical model to generate high-quality corrected image data.

[0122] S3. Use the distortion compensation factor to adjust the pixel coordinates in the corrected image data by using a geometric transformation algorithm to correct the geometric distortion and obtain geometrically corrected image data;

[0123] In this embodiment, S3 specifically includes:

[0124] S31. Use the distortion compensation factor , to adjust the pixel coordinates in the corrected image data. The adjustment operation includes transforming the coordinates of each pixel point in the image data to correct the geometric distortion;

[0125] S32. According to the distortion compensation factor and the original pixel coordinates of the corrected image data , establish a pixel coordinate transformation formula through a geometric transformation algorithm:

[0126] ;

[0127] ;

[0128] where, and are the pixel coordinates after geometric correction, and are the original pixel coordinates of the corrected image data, and are the pixel coordinate adjustment amounts, is the distortion compensation factor;

[0129] S33. According to the distortion compensation factor of each pixel point in the corrected image data , use the weighted average algorithm to calculate the pixel coordinate adjustment amount of each pixel point:

[0130] ;

[0131] ;

[0132] where, is the weighting factor:

[0133] ;

[0134] where and are the coordinates of adjacent pixel points, is the distortion compensation factor of adjacent pixel points, is a small constant used to avoid the weighting factor being zero;

[0135] S34. Adjust all pixel points in the image data through the coordinate transformation formula and the weighted average algorithm to eliminate geometric distortion;

[0136] S35. Recalculate the corrected image data using the bilinear interpolation algorithm according to the adjusted pixel coordinates;

[0137] S36. Output the image data after geometric correction.

[0138] In this embodiment, by using the distortion compensation factor and combining the geometric transformation algorithm to adjust the pixel coordinates in the corrected image data, the geometric distortion is accurately corrected to obtain the geometrically corrected image data. This method can effectively eliminate the image distortion caused by lens distortion, especially the geometric deformation in the edge region and complex scenes.

[0139] S4. According to the optical system characteristics of the SWIR6 mm lens, perform geometric correction on the edge region of the geometrically corrected image data to reduce the geometric distortion caused by the edge effect of the optical system and obtain the corrected image data;

[0140] In this embodiment, S4 specifically includes:

[0141] S41. According to the optical system characteristics of the SWIR6 mm lens, identify the edge region in the geometrically corrected image data, and the edge region is the region within 10 to 20 pixels from the image edge in the geometrically corrected image data;

[0142] S42. Calculate the edge region distortion compensation factor according to the geometric characteristics of the geometrically corrected image data, and the edge region distortion compensation factor is used to correct the geometric distortion caused by the edge effect of the optical system:

[0143] ;

[0144] where is the edge region distortion compensation factor, and are the original pixel coordinates of the corrected image data, is the distance from the pixel point in the standard image data to the optical axis, is the coefficient for controlling the compensation intensity;

[0145] S43. According to the edge region distortion compensation factor, use the geometric transformation algorithm to adjust the pixel coordinates of the edge region;

[0146] S44. According to the adjusted pixel coordinates, use the bilinear interpolation method to recalculate the pixel values of the edge region;

[0147] S45. Correct all the pixels in the entire image to eliminate the geometric distortion caused by the edge effect of the optical system;

[0148] S46. Output the image data after geometric correction and edge region correction.

[0149] In this embodiment, by combining the optical system characteristics of the SWIR6 mm lens, precise geometric correction is performed on the edge region of the geometrically corrected image data, significantly reducing the geometric distortion caused by the edge effect of the optical system. Through this process, the distortion of the image edge can be effectively repaired, and the overall accuracy and clarity of the image can be improved, especially in the edge region of the image.

[0150] S5. Apply the adaptive denoising algorithm to the corrected image data, dynamically adjust the noise suppression intensity according to the distribution of optical distortion, eliminate the noise introduced by the correction, and perform detail enhancement processing;

[0151] In this embodiment, S5 specifically includes:

[0152] S51. Perform noise analysis on the corrected image data, and the noise analysis includes identifying the noise introduced in the image during the distortion correction process;

[0153] S52. Dynamically adjust the noise suppression intensity in the denoising algorithm based on the local optical distortion distribution of the corrected image data :

[0154] ;

[0155] where, is the noise suppression intensity, is the adjustment factor, is the distortion compensation factor of each pixel point;

[0156] S53. Apply the adaptive denoising algorithm based on the adjusted noise suppression intensity to suppress the noise in the corrected image data through local noise estimation, and obtain the denoised image data;

[0157] S54. After denoising, perform detail enhancement based on the local detail information of the denoised image data, and use the enhancement coefficient to adjust the texture details of the image:

[0158] ;

[0159] wherein, is the image data after detail enhancement, is the gradient value of the image data at the pixel point, is the detail enhancement coefficient;

[0160] S55. Output the image data after denoising and detail enhancement.

[0161] The specific steps of S53 include:

[0162] S531. Apply an adaptive denoising algorithm based on the adjusted noise suppression intensity , and the denoising algorithm includes denoising the corrected image data by means of wavelet transform;

[0163] S532. Use wavelet transform to decompose the image data, and obtain the low-frequency sub-band and the high-frequency sub-band respectively by low-pass filtering and high-pass filtering of the corrected image data :

[0164] ;

[0165] ;

[0166] wherein, and are the low-pass filtering and high-pass filtering operations respectively;

[0167] S533. According to the noise suppression intensity , dynamically adjust the noise coefficient in the high-frequency sub-band :

[0168] ;

[0169] wherein, is the original high-frequency sub-band, is the noise suppression intensity;

[0170] S534. Reconstruct the adjusted low-frequency and high-frequency sub-bands through inverse wavelet transform to obtain the denoised image data :

[0171] ;

[0172] wherein, is the denoised image data;

[0173] S535. Output the denoised image data.

[0174] In this embodiment, by applying an adaptive denoising algorithm, the noise suppression intensity is dynamically adjusted according to the distribution of optical distortion, successfully eliminating the noise introduced by distortion correction, and performing detail enhancement on the basis of denoising. This method can effectively remove the noise in the image data, especially the high-frequency noise generated during the correction process, while maintaining and enhancing the details of the image, ensuring the smoothness and clarity of the image quality.

[0175] S6. Output the image data after distortion correction, denoising and detail enhancement.

[0176] In this embodiment, by outputting the image data after distortion correction, denoising and detail enhancement processing, the quality and accuracy of the image are successfully improved. This method can effectively eliminate the image distortion caused by lens distortion, noise and other influencing factors, while retaining and enhancing the detail information in the image.

[0177] Example:

[0178] To verify the effectiveness of the present invention, the technology is selected to be applied in an aerospace mission, which is executed by a high-resolution satellite dedicated to Earth observation. The mission of the satellite is to conduct large-scale ground monitoring to assist in resource exploration, environmental monitoring and disaster prediction. The SWIR6 mm lens carried by the satellite must meet the imaging requirements of high resolution and large field of view, so there are extremely high requirements for the imaging quality. Especially during the remote sensing imaging process, distortion and noise will seriously affect the quality and accuracy of the data, thus affecting subsequent data analysis and decision-making. Therefore, it is particularly important to apply the image distortion correction method proposed by the present invention to improve the quality and accuracy of the image.

[0179] The satellite mission team first used the SWIR6 mm lens to conduct large-area ground observations and obtained the original image data. Since the lens adopts a wide-angle short-focus design, the geometric distortion in the image, especially in the edge area, is very obvious, making the image unable to be directly used for precise analysis. During this process, traditional image correction methods cannot effectively solve the distortion problem under the conditions of large field of view and short focal length, resulting in the image quality not meeting the actual requirements. Therefore, the satellite mission team decided to use the image distortion correction method proposed by the present invention for processing.

[0180] First, the original image data identifies geometric distortions therein, especially in the edge regions of the image, through distortion analysis. Through the distortion compensation factor calculation and geometric transformation algorithm of the present invention, the geometric distortions of the image are effectively corrected. The coordinates of each pixel point are precisely adjusted. After geometric correction, the distortions in the image are significantly reduced, and the geometric accuracy of the image is greatly improved.

[0181] Next, for the noise introduced due to distortion correction in the image, an adaptive denoising algorithm is used. Based on the noise distribution in each region, the noise suppression intensity is dynamically adjusted, and wavelet transform is applied to suppress the noise in the image data, successfully removing high-frequency noise. At the same time, the denoised image is further optimized through a detail enhancement algorithm. By enhancing the gradient information of the image, details are effectively retained.

[0182] After this series of processes, the generated image data is evaluated in real time. Data analysis shows that the corrected image has been significantly improved in terms of clarity, detail fidelity, and geometric accuracy. Especially in the edge regions, after geometric correction, image distortion is effectively eliminated, and the quality of the image is uniform and precise. The specific data results are as follows:

[0183] Table 1: Image Quality Assessment Report

[0184]

[0185] As can be seen from the data in Table 1, after implementing the method of the present invention, the degree of distortion of the image is significantly reduced, the noise is effectively removed, the details of the image are fully enhanced, and the resolution and clarity of the image are both significantly improved. In specific application scenarios, using this technology can significantly improve the image quality and meet the requirements of satellites for high-resolution and low-distortion imaging.

[0186] The present invention successfully solves the problems of distortion, noise, and detail loss faced by existing SWIR lenses in high-resolution and large field-of-view imaging by combining a dynamic distortion compensation factor, an adaptive denoising algorithm, and a detail enhancement technique. Through this method, the satellite can provide more accurate and reliable image data, greatly enhancing the image analysis capabilities in fields such as resource exploration and environmental monitoring. Specifically, this method not only significantly improves the geometric accuracy and clarity of the image, but also effectively removes noise and enhances the details in the image, providing more accurate support for subsequent remote sensing image processing and analysis.

[0187] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes, should be covered by the protection scope of the present invention.

Claims

1. An image distortion correction method for intelligently identifying a SWIR 6-mm lens, characterized in that, It includes the following steps: S1. Obtain the original image data through a SWIR 6-mm lens, perform distortion analysis on the obtained original image data, and identify the geometric distortion in the original image data; S2. According to the optical system characteristics of the SWIR 6-mm lens, including focal length, aperture, and optical system configuration, calculate the distortion compensation factor in the original image data, establish a mathematical model for distortion correction, and apply the mathematical model to the original image data to generate corrected image data; S3. Use the distortion compensation factor to adjust the pixel coordinates in the corrected image data by using a geometric transformation algorithm, correct the geometric distortion, and obtain geometrically corrected image data; S4. According to the optical system characteristics of the SWIR 6-mm lens, perform geometric correction on the edge region of the geometrically corrected image data to reduce the geometric distortion caused by the optical system edge effect, and obtain corrected image data; S5. Apply an adaptive denoising algorithm to the corrected image data, dynamically adjust the noise suppression intensity according to the distribution of optical distortion, eliminate the noise introduced by the correction, and perform detail enhancement processing; S6. Output the image data after distortion correction, denoising, and detail enhancement; The specific content of S5 includes: S51. Perform noise analysis on the corrected image data, and the noise analysis includes identifying the noise introduced in the image during the distortion correction process; S52. Dynamically adjust the noise suppression intensity in the denoising algorithm based on the local optical distortion distribution of the corrected image data : ； Among them, is the noise suppression intensity, is the adjustment factor, is the distortion compensation factor for each pixel point; S53. Apply an adaptive denoising algorithm based on the adjusted noise suppression intensity to suppress the noise in the corrected image data through local noise estimation, and obtain denoised image data; S54. After denoising processing, perform detail enhancement based on the local detail information of the denoised image data, and use an enhancement coefficient to adjust the texture details of the image: ； Among them, is the image data after detail enhancement, is the gradient value of the image data at the pixel point, is the detail enhancement coefficient; S55. Output the image data after denoising and detail enhancement processing; The specific content of S53 includes: S531. Apply the adaptive denoising algorithm based on the adjusted noise suppression intensity The denoising algorithm includes denoising the corrected image data by means of wavelet transform method; S532. Decompose the image data using wavelet transform to obtain the corrected image data Obtain the low-frequency sub-band and high-frequency sub-band respectively through low-pass filtering and high-pass filtering: ； ； Among them, and are low-pass filtering and high-pass filtering operations respectively; S533. Dynamically adjust the noise factor in the high-frequency subband according to the noise suppression intensity : ； Among them, is the high-frequency sub-band, is the noise suppression intensity; S534. Reconstruct the adjusted low-frequency and high-frequency subbands through inverse wavelet transform to obtain the denoised image data : ； Among them, is the denoised image data; S535. Output the denoised image data.

2. The distortion correction method for the image of an intelligent recognition SWIR 6-mm lens according to claim 1, wherein The specific content of S1 includes: S11. Obtain the original image data through a SWIR 6-mm lens. The original image data is a short-wave infrared image in the 900 nm to 1700 nm band, and the original image data is in the form of a two-dimensional matrix, containing pixel position and gray value information; S12. Perform preprocessing on the obtained original image data, including cropping, image scaling, and gray value normalization, to obtain standard image data; S13. According to the optical design parameters of the SWIR 6-mm lens, establish an optical distortion model, and perform geometric distortion analysis on the preprocessed standard image data, including radial distortion analysis and tangential distortion analysis; Radial distortion analysis: ； Among them, is the radial distortion value, is the distance from the pixel point in the standard image data to the optical axis, , , are the radial distortion coefficients; Tangential distortion analysis: ； Among them, is the tangential distortion value, , are the coordinates of the pixel points in the standard image data, , , are the tangential distortion coefficients; S14. Use a feature point matching algorithm to identify and extract the pixel positions in the original image data corresponding to the feature points of the standard image data, and quantify the geometric distortion in the original image data by comparing their position deviations.

3. A distortion correction method for images of an intelligent recognition SWIR 6-mm lens according to claim 1, characterized in that, The specific content of S2 includes: S21. According to the optical system characteristics of the SWIR 6-mm lens, obtain the focal length, aperture, lens curvature, refractive index of optical elements, and relative positions between optical elements of the lens; S22. According to the focal length, aperture, lens curvature, and refractive index of optical elements, establish the optical transfer function of the lens: ； Among them, is the optical transfer function corresponding to the spatial frequency in the frequency domain, is the cut-off frequency of the lens, is the attenuation index of the lens, is the spatial frequency in the frequency domain of the standard image data; S23. Calculate the distortion compensation factor in the original image data based on the optical transfer function , where the distortion compensation factor is used to correct the geometric distortion in the original image data: ； Among them, is the aberration compensation factor, is the optical transfer function, is the distance from the pixel point in the standard image data to the optical axis, , , is the radial aberration coefficient; S24. Based on the calculated distortion compensation factor, establish a mathematical model for distortion correction by combining the optical transfer function of the SWIR 6-mm lens and the geometric characteristics of the original image data: ； Among them, is the corrected image data, is the original image data, is the distortion compensation factor; S25. Apply the mathematical model for distortion correction to the original image data, and perform geometric distortion correction on the original image data to generate corrected image data.

4. The distortion correction method for the image of an intelligent recognition SWIR 6-mm lens according to claim 1, characterized in that, The specific steps of S3 are as follows: S31. Using the distortion compensation factor , adjust the pixel coordinates in the corrected image data. The adjustment operation includes transforming the coordinates of each pixel point in the image data to correct geometric distortion; S32. According to the distortion compensation factor and the original pixel coordinates of the corrected image data , establish a pixel coordinate transformation formula through a geometric transformation algorithm: ； ； Among them, and are the pixel coordinates after geometric correction, and are the original pixel coordinates of the corrected image data, and is the pixel coordinate adjustment amount, is the distortion compensation factor; S33. According to the distortion compensation factors of each pixel point in the corrected image data , use the weighted average algorithm to calculate the pixel coordinate adjustment amount of each pixel point: ； ； Among them, is a weighting factor: ； Among them, and are the coordinates of adjacent pixel points, is the distortion compensation factor of adjacent pixel points, is a small constant used to avoid the weighting factor being zero; S34. Adjust all pixel points in the image data through the coordinate transformation formula and the weighted average algorithm to eliminate geometric distortion; S35. According to the adjusted pixel coordinates, recalculate the corrected image data using the bilinear interpolation algorithm; S36. Output the image data after geometric correction.

5. The distortion correction method for the image of an intelligent recognition SWIR 6-mm lens according to claim 1, characterized in that, The specific steps of S4 are as follows: S41. According to the optical system characteristics of the SWIR 6-mm lens, identify the edge region in the geometrically corrected image data. The edge region is the region within 10 to 20 pixels from the image edge in the geometrically corrected image data; S42. Calculate the edge region distortion compensation factor according to the geometric characteristics of the image data corrected geometrically. , where the edge region distortion compensation factor is used to correct the geometric distortion caused by the edge effect of the optical system. ； Among them, is the edge region distortion compensation factor, and are the original pixel coordinates of the corrected image data, is the distance from the pixel point in the standard image data to the optical axis, is the coefficient for controlling the compensation intensity; S43. According to the edge region distortion compensation factor, adjust the pixel coordinates of the edge region using the geometric transformation algorithm; S44. According to the adjusted pixel coordinates, recalculate the pixel values of the edge region using the bilinear interpolation method; S45. Correct all pixels in the entire image to eliminate the geometric distortion caused by the edge effect of the optical system; S46. Output the image data after geometric correction and edge region correction.

Images