A method for global enhancement of wide-field astronomical images

Abstract

The application discloses a wide-area astronomical image global enhancement method, which comprises the following steps: high-frame-frequency image acquisition is performed on a starry sky area; a reference star is converted from a star table position to an observation position, and a linear film model is solved; based on the solving result, geometric distortion correction is performed; based on the image after the geometric distortion correction, reference stars in a full field of view are matched and identified, so that the celestial coordinate information of all star images in the field of view is obtained; based on the celestial coordinate information, the poor astronomical effect of the star images is corrected, an ideal coordinate system with the center of the field of view as a tangent point is established, and the time-series observation images are projected to the ideal coordinate system; the processed images are translated, all the translated images are accumulated, the average value of all the pixel positions after the accumulation is obtained, the image superposition enhancement is realized, the dynamic range of a detector and the detection capability of a wide-area astronomical observation system are effectively improved; meanwhile, the signal-to-noise ratio of the star images can be effectively improved, and the star image centering and photometric precision can be further improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of astronomical image processing, in particular to a wide-area astronomical image global enhancement method. BACKGROUND

[0002] Limited by the dynamic range of the detector, the detector cannot improve the detection capability by infinitely delaying the exposure time, and the bright stars are more likely to appear saturation phenomenon, while the dark stars have the problem of insufficient signal-to-noise ratio. Through short exposure superposition is the most effective means to improve the dynamic range of the detector and the detection capability of the optical system. The premise of image shift superposition enhancement is that the target is high-precision registered in the sequence image, however, the influence of astronomical distortion and instrument distortion in the wide-area observation image makes it difficult to globally register the observation image; therefore, a wide-area astronomical image global enhancement method is needed to solve the above problems. SUMMARY

[0003] The purpose of the present application is to provide a wide-area astronomical image global enhancement method, which can effectively solve the problems existing in the prior art.

[0004] To solve the above technical problems, the present application adopts the following technical scheme: a wide-area astronomical image global enhancement method, comprising the following steps:

[0005] S1, high-frame-rate image acquisition is performed on the starry sky area;

[0006] S2, the reference star is converted from the star table position to the observation position, and a linear film model is solved, that is, the mapping relationship of the reference star celestial coordinates and the image pixel coordinates is solved, and the celestial sphere and the image plane theoretically only have a linear relationship of translation and rotation, so the position residual of the star image after the model mapping is the geometric distortion component;

[0007] S3, based on the solving result, the mean value of the residuals of all star images in the preset range is counted and 3sigma iteration is performed, and the distortion variable of any point in the field of view is calculated based on the statistical result, and the geometric distortion is corrected;

[0008] S4, based on the image after the geometric distortion correction, the reference stars in the whole field of view are matched and identified, and the celestial coordinate information of all star images in the field of view is obtained;

[0009] S5, based on the collected image and the celestial coordinate information of the star image, the following processing is performed:

[0010] The poor astronomical effect of the star image is corrected;

[0011] An ideal coordinate system with the center of the field of view as the tangent point is established, and the time sequence observation image is projected into the ideal coordinate system;

[0012] S5, the processed image is translated, and all the translated images are accumulated, the number of accumulations of each pixel position is counted, and the average value of all the accumulated pixel positions is taken to realize image superposition enhancement.

[0013] Preferably, the image acquisition of the starry sky region is carried out at no less than 10 Hz.

[0014] Preferably, the acquired multiple images are preprocessed, the median filtering is performed on the whole image through a window larger than 2 times the diameter of the star image, the star image (high-frequency signal) in the image is filtered out, the background image containing low-frequency signals such as thin clouds and optical system vignetting is obtained, and the original image is subtracted from the background image to remove the low-frequency background fluctuation such as thin clouds and optical system vignetting, and the fixed pattern noise such as dark current.

[0015] Preferably, the star image detection is performed on the acquired image.

[0016] The background mean and background noise of the image are counted, and the image is globally binarized segmented based on the statistical value;

[0017] The binarized image obtained by processing is connected domain labeled;

[0018] The connected domain labeling result is subjected to star region statistics and result output.

[0019] Preferably, the reference stars in the whole field of view are matched and identified, including:

[0020] Based on the pointing initial value information of the acquired image, the related region is searched in the star table to form a navigation star table;

[0021] Based on the navigation star table obtained by processing, a triangular arc length library is constructed;

[0022] Based on the star image detection result, an observation triangle library is constructed;

[0023] The navigation star library and the measured image are matched by using the principle of congruent triangles to obtain the celestial coordinate information of all star images in the field of view.

[0024] Preferably, the differential astronomical effect of the star image is corrected, including differential atmospheric refraction correction, and the specific process is as follows:

[0025] The observation direction vector of the target is calculated :

[0026] ;

[0027] Wherein, is the direction vector of the target in the absence of atmosphere, j=0, i represent the direction vector of the field of view center, the observed target i in the equatorial coordinate system, respectively; j=p represents the true direction of the pole, j=z represents the apparent direction vector of the zenith in the equatorial coordinate system; is the apparent zenith distance of the target, defined as the direction vector of the instantaneous zenith is the apparent direction vector of the target ; is the atmospheric refraction angle of the target

[0028] ;

[0029] Based on the observed direction vector , the polar coordinates of each pixel in the observed image are calculated and , wherein:

[0030] ;

[0031] ;

[0032] ;

[0033] ;

[0034] and are the unit direction vectors of the great circle arc and tangent line passing through ; the symbol is the direction vector of the vector ; represents the point of ; represents the point of ;

[0035] Based on the difference between the polar coordinates and , the polar coordinate changes and at the observation time are calculated, and the polar coordinate changes and are used for correction.

[0036] Preferably, the parallax astronomical effect of the star image is corrected, which also includes parallax aberration correction, specifically:

[0037] The observed direction vector affected by aberration is calculated:

[0038] ;

[0039] ;

[0040] where, is the direction vector of the observer's velocity, is the direction vector of the photon's arrival at the stationary observer, is the deflection angle, is the motion rate of the observer, and D is the distance from the point of the photon's arrival direction;

[0041] The celestial coordinates corresponding to each pixel are calculated based on the film model, and the differential aberration amount of each pixel position in the field of view is calculated through a theoretical model according to the celestial coordinates, and the correction of this effect is performed.

[0042] Preferably, the ideal coordinate system is established as follows: a rectangular coordinate system is established on the film with C point as the origin, the vertical axis is the projection of the declination circle, and the positive direction is taken as the direction of the increase of the declination; the horizontal axis is perpendicular to the axis, and the positive direction is taken as the direction of the increase of the right ascension;

[0043] wherein the ideal coordinates of the star image in the coordinate system , are one-to-one corresponding to the equatorial coordinates of the star image , and the corresponding relationship is:

[0044] ;

[0045] is the tangent equatorial coordinate of the intersection point of the field plane of the telescope and the celestial sphere;

[0046] The formula for calculating the equatorial coordinates from the ideal coordinates is:

[0047] .

[0048] Preferably, the film model parameters are solved by observing a dense star field, and the film model parameters are described by 4 parameters or 6 parameters after the geometric distortion correction and the differential astronomical effect correction; the solving formula is:

[0049] .

[0050] Preferably, the translation of the graph includes:

[0051] (1) referring to the bright stars in the image, the translation amount and the rotation amount of the time-series image relative to the reference image are calculated;

[0052] (2) the translation amounts dx and dy of the image in the sub-pixel scale are determined, and a blank image with the same size as the reference image is created to store the translated image;

[0053] (3), calculate the position of each pixel in the original image after translation; for each pixel position (x, y), the position after translation is (x+dx, y+dy);

[0054] (4), according to the pixel position after translation, find the corresponding adjacent pixels on the original image, adopt the bilinear interpolation method, take the weight value of the adjacent four pixel points of the target position, multiply the value of the adjacent pixels by the corresponding weight, and accumulate the result to calculate the pixel value after translation;

[0055] (5), assign the calculated pixel value after translation to the corresponding position of the blank reference image, repeat steps (2) to (4), process all pixels, and complete the translation of the image.

[0056] Beneficial effects: the present application corrects the geometric distortion before matching the certificate, matches more reference stars, calculates more accurate timing image registration quantity, corrects the poor astronomical effect of star image, eliminates the influence of atmospheric refraction and aberration, establishes an ideal coordinate system with the center of the field of view as the tangent point, projects the timing observation image to the ideal coordinate system, eliminates the star radius bending effect, obtains more accurate and effective images, and then completes the displacement and registration of the observation image in the sub-pixel scale, realizes the effective accumulation of the target signal in the timing image, and achieves the same enhancement effect of the observation image edge and the observation image center, effectively improves the dynamic range of the detector and the detection ability of the wide-area astronomical observation system; meanwhile, the present application can effectively improve the signal-to-noise ratio of the star image, and further improve the star image centering and photometric accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0057] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation of the present application.

[0058] In the drawings:

[0059] Figure 1 The flow chart of the present application wide-area astronomical image global enhancement method;

[0060] Figure 2 The polar coordinate diagram for calculating the poor atmospheric refraction of the present application;

[0061] Figure 3 The schematic diagram of the aberration effect of the present application;

[0062] Figure 4 The relationship diagram of the equatorial coordinate and the ideal coordinate of the present application;

[0063] Figure 5 The poor atmospheric refraction diagram of the present application;

[0064] Figure 6A schematic diagram of poor light aberration for the present application;

[0065] Figure 7 A schematic diagram of instrument distortion for the present application (left: original star map, right: distortion modeling result);

[0066] Figure 8 A comparison diagram of conventional shift overlay enhancement (left) and global enhancement (right) effects. DETAILED DESCRIPTION

[0067] The embodiments of the present application will be described below in conjunction with the accompanying drawings. The terms used in the embodiment part of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0068] Embodiment: As shown in the figure, a wide-area astronomical image global enhancement method comprises the following steps: Figure 1

[0069] S1, high-frame-rate (not less than 10 Hz) image acquisition is performed on the starry sky area, and the collected multiple frames of images are preprocessed to remove thin clouds, optical system vignetting, low-frequency background fluctuations, and fixed-pattern noise such as dark current;

[0070] S2, star image detection is performed on the star map obtained after preprocessing;

[0071] In this embodiment, the following method is used for star image detection:

[0072] S21, the background mean and background noise of the image are counted, and the image is globally binarized based on the counted values;

[0073] S22, the binarized image obtained by processing step S21 is labeled with a connected domain, and an 8-connected method is used;

[0074] S23, the connected domain labeling result obtained by processing step S22 is counted and outputted;

[0075] S3, the star map containing candidate star images is matched and identified with a reference star catalog, the reference stars in the entire field of view are matched and identified, and the matching and identification results are checked;

[0076] In this embodiment, the following method is used for matching and identification:

[0077] S31, based on the pointing initial value information of the collected image, the relevant area is searched in the star catalog to form a navigation star catalog;

[0078] S32, based on the navigation star catalog obtained by processing step S31, a triangular arc length library is constructed;

[0079] ​S33, constructing an observation triangle library based on the star image detection result obtained in step S2;

[0080] S34, matching the navigation star library with the measured image based on the principle of congruent triangles to obtain the celestial coordinate information of all star images in the field of view;

[0081] For the collected images, the following problems exist:

[0082] (1) In a large field of view astronomical observation image, the regions in the field of view are affected differently by atmospheric refraction, and the star image will change the angular distance relative to a reference point in the field of view, as shown in Figure 5 .

[0083] (2) Due to the finiteness of the speed of light and the motion of the observer relative to the target celestial body, the observation direction of the celestial body deviates from the actual photon arrival direction; as shown in Figure 6 .

[0084] (3) The wide-area astronomical observation field is large, and the distortion mode is complex, so geometric distortion correction

[0085] (4) In a fixed pointing observation device, the star image motion trajectory projected into the observation image is a curve, and the curvature radius of the motion trajectory of the star image with different declinations is different in the field of view.

[0086] Based on the above, in the embodiment, the following method is adopted:

[0087] S4, correcting the aberration astronomical effect of the star image based on the star image celestial coordinate information obtained in step S3;

[0088] S41, aberration atmospheric refraction correction; in the embodiment, the following symbols are defined:

[0089] is the direction vector of the target in the absence of atmosphere; j=0, i respectively represent the field of view center, the observation target i in the true equatorial coordinate system The direction vector of the view; j=p represents the true direction of the celestial pole, and j=z represents the view direction vector of the zenith in the true equatorial coordinate system; for the field of view center and all target stars in the field of view, the view direction vector mentioned here already includes the influence of the aberration; Since the aberration does not change the rotation center of the daily visual motion of the starry sky, the direction vector of the celestial pole here only uses its true direction.

[0090] is the observation direction vector of the target j (including the influence of atmospheric refraction);

[0091] Due to the influence of atmospheric refraction, the view direction vector of the celestial body before entering the atmosphere will become the observation direction vector . Located in the plane of the zenith direction vector Zenith direction vector The relationship between them can be expressed as:

[0092] ;

[0093] Where, is the target's apparent zenith distance, defined as the angle between the instantaneous zenith direction vector and the target's apparent direction vector . is the target's atmospheric refraction angle, which can be expressed as:

[0094] ;

[0095] Under standard atmospheric conditions, for yellow starlight with an equivalent wavelength of about 0.57 μm , , the corresponding refraction angle is , and Rj0 is the Rj under the assignment of k1 and k2; in the case of known apparent zenith distance, the above formula can be solved by iteration . When the atmospheric conditions at the observation site deviate significantly from the standard state, the actual refraction amount can be calculated using the following formula:

[0096] ;

[0097] Where T and P are the air temperature and air pressure at the time of observation, respectively.

[0098] The vector expression of the target's apparent zenith distance is:

[0099] ;

[0100] From the above formula, the observation directions of the celestial pole, the zenith, and the target can be calculated .

[0101] To describe the position of the target in the image field relative to the center of the field of view, a spherical polar coordinate system is established with the center of the field of view as the pole, as shown in Figure 2 , the coordinate vector is represented by ; here is the great circle arc length of the target's image to the center of the field of view, is the angular aperture of the target's image and the image of the north celestial pole at the center of the field of view;

[0102] Polar coordinates of the image of target i:

[0103] ;

[0104] ;

[0105] or

[0106] ;

[0107] wherein:

[0108] ;

[0109] ;

[0110] and are the unit direction vectors of the great circle arc and tangent respectively; the symbol represents the direction vector of the vector ; represents the direction vector of the vector ; represents the direction vector of the vector ; represents the direction vector of the vector

[0111] According to the polar coordinates of each pixel in the observed image and the polar coordinate changes and at the observation time can be obtained, and after correction, the observed image without the influence of differential atmospheric refraction relative to the center of the field of view can be obtained.

[0112] S42, differential aberration correction;

[0113] The influence of the celestial body observation position is represented by the following formula:

[0114] ;

[0115] ;

[0116] wherein, is the direction vector of the observer's speed, is the direction vector of the photon arrival at the stationary observer, is the observation direction vector after receiving the influence of the aberration, is the deflection angle, is the motion rate of the observer, and D is the distance from the point of arrival of the photon. Reference Figure 3 is shown, which is a schematic diagram of the influence of the aberration on the target observation direction;

[0117] In a large field of view astronomical observation image, the regions in the field of view are affected differently by the aberration, and the star image will produce an angular distance change relative to a reference point in the field of view; from Figure 3As can be seen from the aberration formula, the aberration displacement on the celestial sphere makes the observed direction of the celestial body displaced along the great circle arc passing through the celestial body and the vantage point towards the vantage point, and the size of the displacement is proportional to the speed of the observer and the sine of the distance from the vantage point. In the large field of view observation image, the distance from the vantage point of each pixel position is different, so the influence of the aberration is also different. The influence of the aberration on the observation position of the target in the field of view is very similar to the atmospheric refraction, and the correction method of the aberration effect is also the same, that is, the aberration amount and the refraction amount relative to the center of the field of view are calculated based on the pixel point of the celestial sphere coordinate and the aberration model, and the pixel position of the image is corrected.

[0118] S5, geometric distortion correction:

[0119] In the embodiment, the following method is used for geometric distortion correction:

[0120] S51, a batch of dense star field observation images are collected to ensure that the star image positions in the field of view have obvious changes;

[0121] S52, star image detection and reference star matching are performed on the observation image;

[0122] S53, the reference star is converted from the star table position to the observation position to avoid the distortion caused by the aberration atmospheric refraction and the aberration aberration;

[0123] S54, a linear film model is solved, and in the embodiment, 4 parameters are used to perceive all linear and nonlinear distortion terms;

[0124] S55, based on the solving result of the single frame linear film model, the measurement residual of all star images with good signal-to-noise ratio and non-saturation is counted;

[0125] S56, a suitable image grid (30*30 grid is used in the embodiment) is selected, the mean value of all star image residuals in the grid is counted, and 3sigma iteration is performed to remove;

[0126] S57, the mean value of the residuals of a large number of star images in each grid can absorb the random error of the position measurement, so as to standardize the position system error, that is, the geometric distortion of the grid point, as shown in Figure 7

[0127] S58, in the subsequent data processing process, based on the numerical template of the geometric distortion, the distortion amount of any point in the field of view is calculated by using the bilinear interpolation or the bicubic spline interpolation method, and the final geometric distortion correction is completed;

[0128] ​In still another embodiment, before the matching identification in step S2, the geometric distortion correction in step S5 is performed to calculate a more accurate time-series image registration amount; the geometric distortion correction in step S5 can be iterated with the reference star matching identification to match more reference stars, so that the areas lacking reference stars can be corrected, and more distortion correction can be matched to more reference stars, and generally 2-3 iterations can be stable;

[0129] In this embodiment, after the field distortion of the telescope image and the poor atmospheric refraction and aberration are corrected, an ideal coordinate system with the center of the field of view as a tangent point is established, the horizontal and vertical axes of the coordinate system are the increasing directions of the right ascension and declination, respectively, and the time-series observation image is projected into the ideal coordinate system, so that the star image motion trajectory caused by the earth rotation becomes a straight line in the ideal coordinate system, and the star diameter bending effect can be eliminated.

[0130] The ideal coordinate system establishment process is as follows:

[0131] The imaging rule of the astronomical telescope follows the central projection law, and therefore the ideal coordinate system is defined as a rectangular coordinate system on the photographic plate. Figure 4 In the ideal coordinate system, CO is the optical axis of the telescope, which is perpendicular to the focal plane and has a focal length F; the photographic plate is placed on the focal plane, and the center of the photographic plate is C; when CO is extended to the sky, the intersection of CO and the sky is C', which is the tangent point of the telescope field plane and the sky, and the equatorial coordinates of the tangent point are .

[0132] A rectangular coordinate system is established on the photographic plate with C as the origin, the vertical axis is the projection of the declination circle, and the increasing direction of the declination is taken as the positive direction; the horizontal axis is perpendicular to the axis, and the increasing direction of the right ascension is taken as the positive direction; C- is called the ideal coordinate system; for a photographic plate, the ideal coordinate system is unique, and the ideal coordinates , of the star image in the ideal coordinate system are one-to-one corresponding to the equatorial coordinates , and have the following relationship:

[0133] .

[0134] The formula for calculating the equatorial coordinates from the ideal coordinates is:

[0135] .

[0136] Because the measured coordinate axis is not completely parallel to the ideal coordinate axis, the scales of the two coordinate systems are also different, so the model parameters of the plate are solved by observing the dense star field, and after the effects of the image field distortion, poor atmospheric refraction and aberration are corrected, the measured coordinate and the ideal coordinate should be in linear relationship, which can be described by 4 parameters or 6 parameters.

[0137] .

[0138] S6, sub-pixel scale image translation registration;

[0139] In this embodiment, the following method is used for translation registration:

[0140] S61, after the astronomical distortion and instrument distortion are calibrated, the registration of the time sequence images can be completed by simple translation and rotation, because the time sequence images are short in interval, most of the bright stars in the images will not be out of the field of view, and the translation and rotation of the time sequence images relative to the reference image are calculated by taking the bright stars in the images as the reference;

[0141] S62, because the translation and rotation are not integer pixels, the sub-pixel scale needs to be completed; the key of sub-pixel image shift superposition is to make the discrete pixels continuous, or it can be understood as translating the image at the sub-pixel scale; the linear interpolation method is used to complete the shift of the image at the sub-pixel scale;

[0142] S63, the translation dx and dy of the image at the sub-pixel scale are determined, and a blank image with the same size as the reference image is created to store the translated image;

[0143] S64, the position of each pixel in the original image after translation is calculated. For each pixel position (x, y), the position (x+dx, y+dy) after translation is calculated;

[0144] S65, according to the pixel position after translation, the corresponding adjacent pixels in the original image are found, the double linear interpolation method is used to take the weight of the adjacent four pixel points of the target position, assuming that the area of the entire square region is 1, the value of point (x, y) can divide the block area into four, and the proportion of the area of each block region to the total area is the weight of the corresponding vertex, the value of the adjacent pixel is multiplied by the corresponding weight, and the result is accumulated to calculate the pixel value after translation;

[0145] S66, the calculated pixel value after translation is assigned to the corresponding position of the blank reference image, and steps S63-S65 are repeated to process all pixels, and the translation of the image is completed;

[0146] S7. Image overlay enhancement: Accumulate all translated images, count the number of accumulations at each pixel position, and take the average value of all accumulated pixel positions to complete the image overlay enhancement.

[0147] Based on the above methods, refer to Figure 8 The image shows a comparison between the global enhancement processing of the measured star map based on the method of this invention and the conventional shift and superposition processing. Figure 8 Both the left and right images are the results of superimposed enhancement of 100 frames of temporal images. The left image shows the result of conventional shift-superimposed enhancement, which shows that the star morphology in the central region of the field of view is good, but the star morphology in the edge regions is elongated in different directions. This is because astronomical and instrumental distortion effects have not been corrected. The right image shows the global enhancement result processed by this invention, which shows that the star morphology in both the edge and central regions of the image is good, achieving a high-quality global enhancement effect.

[0148] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. For those skilled in the art, after learning the contents described in the present invention, several equivalent changes and substitutions can be made without departing from the principle of the present invention. These equivalent changes and substitutions should also be considered to fall within the protection scope of the present invention.

Claims

1. A method for global enhancement of wide-field astronomical images, characterized in that, The method comprises the following steps: S1, high frame frequency acquisition is performed on the starry sky region; S2, the reference star is converted from the star table position to the observation position, and a linear film model is solved; S3, based on the solving result, the residual mean of all star images in a preset range is counted and 3sigma iteration is performed, the distortion variable of any point in the field of view is calculated based on the statistical result, and geometric distortion correction is performed; S4, based on the image after geometric distortion correction, the reference star in the whole field of view is matched and identified, and the celestial coordinate information of all star images in the field of view is obtained; S5, based on the collected image and the celestial coordinate information of the star image, the following processing is performed: The differential astronomical effect of the star image is corrected; An ideal coordinate system with the center of the field of view as the tangent point is established, and the time sequence observation image is projected to the ideal coordinate system; S6, the processed image is translated, and all the translated images are accumulated, the number of accumulations of each pixel position is counted, the average value of all the accumulated pixel positions is taken, and image superposition enhancement is realized; The differential astronomical effect of the star image is corrected, including differential atmospheric refraction correction, specifically: Computing a target's observation direction vector : ； wherein, is the direction vector of the target in the absence of atmosphere, j = 0, i respectively represent the central field of view, the observed target i in the true equatorial coordinate system direction vector; j = p represents the true direction of the celestial pole, j = z represents the apparent direction vector of the zenith in the true equatorial coordinate system; is the apparent zenith distance of the target, defined as the direction vector of the instantaneous zenith is the angle between the apparent direction vector of the target and the direction vector of the target; is the atmospheric refraction angle of the target; ； Based on the observation direction vector , the polar coordinates of each pixel in the observation image are calculated and wherein: ； ； ； ； and are the unit direction vectors of the great circle arc and tangent line, respectively; the symbol represents the direction vector of the vector ; ; denotes the point ; denotes the point ; polar coordinates and differential calculation of polar coordinate changes at observation times and and correction using polar coordinate changes and ​ 2. The method of claim 1, wherein: Image acquisition is performed on the starry sky region at a frequency of not less than 10 Hz.

3. The method of claim 2, wherein: The collected multiple images are preprocessed, the median filter is performed on the whole image through a window larger than 2 times the diameter of the star image, the star image in the image is filtered out, the background image containing low-frequency signals such as thin clouds and optical system vignetting is obtained, and the original image is subtracted from the background image to remove the low-frequency background fluctuation such as thin clouds and optical system vignetting, and the fixed pattern noise such as dark current.

4. The method of claim 1, wherein: Star image detection is performed on the collected image: The background mean and background noise of the image are counted, and the image is globally binarized based on the statistical value; The obtained binarized image is processed for connected domain labeling; The connected domain labeling result is subjected to star image region statistics and result output.

5. The method of claim 4, wherein: Matching and identifying the reference star in the whole field of view, including: Based on the pointing initial value information of the collected image, the related region is searched in the star table to form a navigation star table; Based on the obtained navigation star table, a triangular arc length library is constructed; Based on the star image detection result, an observation triangle library is constructed; The navigation star library and the measured image are matched by using the principle of congruent triangles to obtain the celestial coordinate information of all star images in the field of view.

6. The method of claim 1, wherein: The differential astronomical effect of the star image is also corrected, including differential aberration correction, specifically: The observation direction vector affected by the aberration is calculated: ； ； wherein is the direction vector of the observer's velocity, is the direction vector of the photon's arrival at the stationary observer, is the deflection angle, is the observer's motion rate, and D is the direction of the photon's arrival. Based on the film model, the celestial coordinates corresponding to each pixel are calculated, the differential aberration amount of each pixel position in the field of view is calculated through a theoretical model according to the celestial coordinates, and the differential astronomical effect is corrected.

7. The method of claim 1, wherein: The ideal coordinate system is established as follows: A rectangular coordinate system is created on the film with point C as the origin, and the vertical axis... This is the projection of the declination circle, with the direction of increased declination taken as the positive direction; the horizontal axis... Perpendicular to The axis is defined as the direction in which right ascension increases, taking it as the positive direction. Wherein, the ideal coordinate of the star in the coordinate system , ) and its equatorial coordinate are one-to-one correspondence, and the correspondence is: ； where: a - declination of the point of intersection of the plane of the field of view of the telescope with the celestial sphere; The formula for calculating the equatorial coordinate from the ideal coordinate is: 。 8. The method of claim 7, wherein: The film model parameters are solved by observing a dense star field, and after the geometric distortion correction and the differential astronomical effect correction, the film model parameters are described by 4 parameters or 6 parameters; the solving formula is: 。 9. The method of claim 1, wherein: The image is translated, including: (1) taking the bright star in the image as the reference, calculating the translation amount and rotation amount of the time sequence image relative to the reference image; (2) determining the translation amount dx and dy of the image in the sub-pixel scale, creating a blank image with the same size as the reference image, and storing the translated image in the blank image; (3) Calculate the position of each pixel in the original image after translation; for each pixel position (x, y), the position after translation is (x+dx, y+dy); (4) According to the position of the pixel after translation, find the corresponding adjacent pixels on the original image, and use the bilinear interpolation method to take the weight of the target position adjacent four pixel points, multiply the value of the adjacent pixels with the corresponding weight, and accumulate the results to calculate the pixel value after translation; (5) Assign the calculated pixel value after translation to the corresponding position of the blank reference image, repeat steps (2) to (4), process all pixels, and complete the translation of the image.

Images