Automatic focusing method based on ground-based full-solar-surface solar telescope

Abstract

The invention discloses an automatic focusing method based on a ground-based full-solar-surface solar telescope. The method comprises the following steps: 1) determining a theoretical optimal focus of the ground-based full-solar-surface solar telescope, an adjustable range of the theoretical optimal focus and an adjusting step length; 2) focusing according to a set adjustment step length in the adjustable range; (3) shooting multiple frames of short exposure images in the atmosphere freezing time for the focus adjusted by each adjusting step length, and then selecting an image with the optimal image quality as a reference image of the corresponding focus by utilizing a spectrum ratio method; 4) obtaining a focusing image sequence according to the reference image of each focus; 5) screening the focusing image sequence by using a spectral ratio method to obtain a ratio result of different focuses at the limit spatial resolution of the ground-based full-solar-surface solar telescope as a focusing sequence image quality; 6) fitting the image quality of the focusing sequence to obtain an adjustable optimal focus; and 7) driving a focusing mechanism to drive an imaging device according to the obtained adjustable optimal focus, and adjusting the focus to the corresponding optimal focus.

Description

Technical Field

[0001] The present invention belongs to the technical field of optics and relates to an automatic focusing method based on a ground-based full-disk solar telescope. Background Art

[0002] The full disk refers to the entire observable solar disk area. The observation target of a ground-based full-disk solar telescope is the sun. During the process of observing the sun, due to the change in the solar altitude angle, there is a certain change in the focal length of the telescope's optical imaging system. Therefore, the telescope needs to adjust the focal length according to the change. In the past, most solar telescopes were manually focused by observers based on experience. This method has certain subjective errors, and the human errors of different observers are different, making it impossible to achieve a unified standard. In addition, with the advancement of telescope digitization, all functions have been automatically controlled, and manual focusing has instead become a major obstacle to digitization. Therefore, how to automatically focus the telescope has become an urgent problem to be solved.

[0003] Currently, automatic focusing technology is widely used in optical terminal devices in various fields. Its principle is that during the automatic focusing process, there are changes in the characteristics or mid-high frequency components of the image formed by the observation target at the imaging terminal. The sharpness evaluation / focusing function can be used to analyze the imaging quality at different positions. The image with more characteristic conformities or mid-high frequency components is regarded as the best focus imaging, and this position is the best imaging focus. However, different observation targets have unique physical characteristics, resulting in different image characteristics of the images formed at the imaging terminal. It is impossible to use a unified standard for automatic focusing. Without additional modification to the existing equipment, based on the current application situation of automatic focusing, its general technical solutions can be roughly divided into automatic focusing based on the gradient function as the focusing function, automatic focusing based on the spectrum function as the focusing function, and automatic focusing based on the entropy function as the focusing function. The theoretical bases of the three are different.

[0004] The evaluation basis of the gradient function is that the closer the image is to the best focus, the sharper its edges are, and the greater the gradient value on the gradient image. The spectrum function relies on the fact that the clearer the image, the more detailed information it contains, so there is more mid-high frequency information in the frequency domain. The premise for using the entropy function is that the entropy of the best focus image must be greater than the entropy of the defocused image. Under this premise, the best focus can be selected according to the entropy size.

[0005] The above three methods have different theories and different application scenarios.

[0006] Automatic focusing is technically mature in the civilian and industrial fields, but it is still in the exploratory stage in full-disk solar observations. The reasons can be attributed to the following three points:

[0007] 1. Influence of atmospheric turbulence on full-disk imaging

[0008] The observation target of the ground-based full-disk solar telescope is the full disk. The sunlight of the full disk needs to pass through the Earth's atmosphere before it can be detected by the telescope. During the process of passing through the atmosphere, the light emitted from the solar disk experiences absorption, radiation, scattering, and turbulence of the atmosphere. Among them, the last item has the greatest impact on the focal length analysis.

[0009] Atmospheric turbulence refers to the irregular motion of atmospheric turbulence. In terms of regions, it has an impact on the imaging of the telescope at various scales, including magnification, distortion, displacement, etc. And there is a very strong time-varying property, that is, a process of strong time-varying and multi-airspace changes. The obtained solar image is affected by various scales in addition to normal solar activities.

[0010] 2. Equipment errors of ground-based telescopes

[0011] During the operation of ground-based telescopes, there are various errors, and most of these errors will be reflected in the final image. For example, the tracking error of the telescope will cause the solar images at different times to jitter; the wavelength band drift of the filter causes changes in the characteristics of the solar image; the dirty spots and non-uniformity existing in the optical imaging system cause small-scale and large-scale intensity changes of the solar image, and so on.

[0012] That is, the strong time-varying and multi-airspace changes of environmental interference, and the multi-scale airspace changes introduced by the observation equipment. Due to the change of the image content, it is impossible to determine whether the image entropy of the optimal focal length position is the largest, so the evaluation of the entropy function is first excluded; for the gradient function, the strong time-varying and multi-airspace changing images make it very difficult to select the gradient area, and the noise effects in different regions at different times make it impossible to achieve repeatable and accurate calculations within the effective area; for spectral analysis, although it avoids the influence of multi-airspace changes, for images that change with time, the calculation results at different times show the law of atmospheric jitter, and this law itself is a random process. Therefore, it is impossible to accurately locate the optimal focus.

[0013] 3. Stronger disciplinary background

[0014] During the automatic focusing process of the full disk, in addition to data acquisition, data processing, motor control, etc., a lot of disciplinary professional knowledge reserves are also required, such as the influence of the environment and equipment on the full-disk image, the optimal focus range of the telescope optical system, the imaging resolution of the telescope, etc. Among them, the influence of the environment and equipment and the imaging resolution of the telescope are convenient for designers to design solutions according to the influencing factors, and the optical guidance of the optimal focus of the optical system determines the focusing range and step size. Therefore, it is impossible to simply apply mature solutions to achieve automatic focusing. Summary of the Invention

[0015] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide an automatic focusing method based on a ground-based full-disk solar telescope.

[0016] The present invention is guided by the theory of telescope optical design to obtain the ideal best focus position and focusing range, and prevent automatic focusing on other image planes of the telescope.

[0017] According to the influence of the atmosphere on the multi-airspace change of the full-disk solar image and the characteristic that it is impossible to select a characteristic area for analysis, the present invention selects the spectral ratio method as the image evaluation function for automatic focusing, and uses the spectral characteristics of the entire image as the evaluation index.

[0018] According to the influence of the atmosphere on the strong time-varying of the full-disk solar image and the inability to select long-integration images, the present invention continuously acquires multiple short-exposure images within the atmospheric freezing time at the same focus position, and performs frame selection according to the lucky imaging theory. With the spectral ratio method as the image evaluation function, the best image is selected as the reference image for this focus.

[0019] According to the characteristic that the theoretical focusing curve is approximately a Gaussian curve, the present invention selects the spectral ratio method as the image evaluation function for each focus reference image, performs Gaussian fitting on the results, and determines the best focus position according to the fitting results.

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

[0021] An automatic focusing method based on a ground-based full-disk solar telescope, the steps of which include:

[0022] 1) Determine the theoretical best focus and its adjustable range, as well as the adjustment step of the ground-based full-disk solar telescope;

[0023] 2) Perform focusing within the adjustable range of the theoretical best focus of the ground-based full-disk solar telescope according to the set adjustment step;

[0024] 3) For each focus adjusted by each adjustment step, use the ground-based full-disk solar telescope to take multiple short-exposure images within the atmospheric freezing time, and then use the spectral ratio method to screen the short-exposure image sequence corresponding to the current focus, and select the image with the best image quality as the reference image for the corresponding focus;

[0025] 4) Obtain a focusing image sequence according to the reference images of each focus;

[0026] 5) Use the spectral ratio method to screen the focusing image sequence, and obtain the ratio results of different foci at the limit spatial resolution of the ground-based full-disk solar telescope as the image quality of the focusing sequence;

[0027] 6) Use the Gaussian fitting method to fit the image quality of the focusing sequence, and then calculate the adjustable best focus according to the fitting function;

[0028] 7) Drive the imaging device of the optical system of the ground-based full-disk solar telescope by the obtained adjustable optimal focus to drive the focus adjustment mechanism, and adjust the focus of the ground-based full-disk solar telescope to the corresponding optimal focus.

[0029] Further, the image processing method of the spectral ratio method is as follows:

[0030] 21) Normalize each image in the image sequence to zero mean;

[0031] 22) Perform Fourier transform on each normalized image respectively to obtain the power spectrum of the image in the frequency domain;

[0032] 23) Convert the power spectrum in the frequency domain from Cartesian coordinate system to polar coordinate system;

[0033] 24) In the polar coordinate system, average the radius corresponding to each frequency along the angular direction to obtain an average curve representing the distribution of the power spectrum along the frequency direction;

[0034] 25) Use the distribution of the power spectrum of an image in the image sequence along the frequency direction as the reference power spectrum distribution L0(ρ) along the frequency direction, and compare the distributions L n (ρ) of the power spectra of the remaining images along the frequency direction with the reference power spectrum distribution along the frequency direction to obtain the image spectral ratio results corresponding to each image

[0035] 26) According to the image spectral ratio results C n (ρ) corresponding to each image, select the image corresponding to the optimal image spectral ratio result as the best image quality image of the image sequence.

[0036] Further, the image spectral ratio result = the distribution of the power spectrum of the image along the frequency direction / the distribution of the reference power spectrum along the frequency direction.

[0037] Further, according to the telescope optical system model of the ground-based full-disk solar telescope, determine the theoretical optimal focus and its adjustable range of the ground-based full-disk solar telescope.

[0038] Further, use the distribution of the power spectrum of the first image in the image sequence along the frequency direction as the reference.

[0039] The advantages of the present invention are as follows:

[0040] 1. From subjective judgment to a unified and relatively objective judgment standard, providing technical support for long-term stable solar observations.

[0041] 2. From manual to automatic process. Description of the Drawings

[0042] Figure 1It is the flowchart of the method of the present invention.

[0043] Figure 2 It is the schematic diagram of the processing results of each step of the present invention.

[0044] Figure 3 It is the flowchart of the spectral ratio method.

[0045] Figure 4 It is an example of spectral ratio in the automatic focusing of a telescope with an imaging size of 992 * 992. Specific implementation mode

[0046] The present invention will be further described in detail below with reference to the accompanying drawings. The examples given are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0047] Aiming at the problems existing in the current full-disk solar telescope automatic focusing, the present invention combines relevant disciplinary knowledge, uses the frame selection technology to select the best full-disk solar image within the atmospheric freeze time, reduces the influence of atmospheric time variability; selects the improved spectral analysis method - the spectral ratio method as the evaluation function to determine the image quality of images at different focus positions, that is, uses the spectral ratio method to suppress the influence of multi-airspace changes; finally, uses the Gaussian fitting method to achieve a more accurate calculation of the best focus, and drives the motor according to the result to move the imaging system to the best position.

[0048] First of all, according to the strong time-varying characteristics of atmospheric turbulence, it can be known that long-exposure images will smooth out the influence of turbulence at different times, resulting in a deterioration of the overall image quality, while short-exposure images within an extremely short time only retain the influence of turbulence at that moment, that is, the minimum time scale of the influence of atmospheric turbulence. This extremely short time is also called the atmospheric freeze time, generally less than 10 ms. Based on this phenomenon, astronomical experts have proposed the lucky imaging theory, that is, continuously shooting multiple frames of images within a short time, and the exposure time of each frame of image is less than the atmospheric freeze time. When evaluating the image quality of this series of images in the later stage, there is always an image whose image quality is close to the diffraction limit of the telescope. Based on this theory, the present invention takes multiple short-exposure images within the atmospheric freeze time at the same focus, and then uses the image evaluation function to select frames from this series of short-exposure images, and selects the image with the best image quality as the reference image for this focus, reducing the influence of the time variability of atmospheric turbulence. This method is adopted for each focus subsequently, so as to obtain the reference image of each focus.

[0049] Then, considering the multi-spatiotemporal variability of atmospheric turbulence, in the time-invariant case, each region on the image has a different instantaneous point spread function. However, within a certain region, its instantaneous point spread function is consistent, and this region is called the isoplanatic patch. Within the isoplanatic patch, the optical properties of the atmosphere are basically unchanged. The image formed by a point source passing through this isoplanatic patch is called a speckle pattern, and the images formed at the same moment are exactly the same. For the full-disk photosphere image, affected by the ground-based equipment environment, the typical value of the size of the isoplanatic patch is 3 - 5 arcseconds. For the full disk, its field of view is generally 34 arcminutes * 34 arcminutes. Calculated using a typical isoplanatic patch size of 5 arcseconds, the full-disk image contains 166,464 isoplanatic patches. Since the full-disk image contains hundreds of thousands of isoplanatic patches and each isoplanatic patch has random variations, it is impossible to select a specific region as the criterion for focusing. Analyzing based on the statistics of the random motion of the atmosphere, jointly analyzing multiple isoplanatic patches will reduce the regions with high image quality of the isoplanatic patches and increase the regions with low image quality of the isoplanatic patches. Therefore, choosing the entire region for image quality analysis can, to a certain extent, suppress the influence of atmospheric turbulence on the full-disk image. One can start from the spectral analysis method, compare the spectra of different full-disk images, and then analyze the comparison results in combination with the spatial resolution of the telescope to select the image with the best image quality, that is, use the spectral ratio method as the image quality evaluation function to determine the image with the best image quality.

[0050] The spectral ratio method is as follows:

[0051] 1. First, perform zero-mean normalization on each of the images to be compared to eliminate the influence of light intensity changes.

[0052] The zero-mean normalization formula is as follows:

[0053]

[0054] In the formula, x represents the horizontal coordinate of the image, y represents the vertical coordinate of the image, I(x, y) is the pixel value of the original image, μ is the mean of the original image, σ is the standard deviation of the original image, and f(x, y) is the pixel value of the image after zero-mean normalization.

[0055] 2. Then, perform Fourier transforms on the normalized images respectively to obtain the power spectra of the images in the frequency domain.

[0056] The two-dimensional discrete Fourier transform formula is as follows:

[0057]

[0058] In the formula, u represents the horizontal coordinate in the frequency domain, v represents the vertical coordinate in the frequency domain, M represents the width of the image, N represents the height of the image, j is the imaginary unit, and F(u, v) represents the result after the two-dimensional discrete Fourier transform.

[0059] The formula for calculating the power spectrum is as follows:

[0060] P(u, v) = Re(F(u, v)) 2 + Im(F(u, v)) 2

[0061] Wherein, Re represents the operation of taking the real part, Im represents the operation of taking the imaginary part, and P(u, v) is the final power spectrum result.

[0062] 3. Convert the power spectrum in the frequency domain from the Cartesian coordinate system, with the center of the frequency domain as the origin and the length of the frequency domain axis as the radius, to the polar coordinate system.

[0063] The polar coordinate conversion formula is as follows:

[0064]

[0065] Wherein, ρ represents the coordinate in the radius direction in the polar coordinate system, θ represents the coordinate in the angle direction in the polar coordinate system, Q represents the power spectrum in the polar coordinate system, and arctan is the arctangent function.

[0066] 4. In the polar coordinate system, average each radius (i.e., frequency) along the angle direction to obtain an average curve; this curve represents the distribution of the power spectrum along the frequency direction.

[0067] The formula is as follows:

[0068]

[0069] Wherein, represents taking the average along the angle direction, and L(ρ) represents the distribution of the power spectrum along the frequency direction.

[0070] 5. Use the distribution of the power spectrum of the first frame image along the frequency direction as the reference power spectrum distribution along the frequency direction, and compare the distribution of the power spectrum of the remaining images along the frequency direction with this reference power spectrum distribution along the frequency direction, that is:

[0071]

[0072] Wherein, L0(ρ) represents the distribution of the reference power spectrum along the frequency direction, L n (ρ) represents the distribution of the power spectrum of the current image along the frequency direction, and C n (ρ) represents the spectral ratio curve of the current image.

[0073] 6. Combine the spectral ratio results with the analysis of the spatial resolution of the telescope, such as Figure 4As shown, that is, the comparison of the telescope's spatial resolution mapped to the frequency domain; considering the influence of atmospheric jitter, if the angular resolution is 4 arcseconds, then the spatial resolution is in the range of 4 - 8 arcseconds. Integrate the data where the integral of the part from the horizontal axis 124 - 249 is greater than 1. The integral result of each line is used as the spectral ratio result corresponding to that position. According to the comparison result, if it is greater than 1, it is better than the reference result; otherwise, it is lower than the reference. The optimal result can be selected as the best image quality image of this sequence.

[0074] Finally, the change in image quality near the ideal best focus can be approximated as a Gaussian change. Therefore, according to this theoretical guidance, the error existing in the frame selection process can be further suppressed. Combining the above analysis, the final method is obtained:

[0075] 0. First, according to the telescope optical system model, determine the theoretical best focus, its adjustable range and step size, and avoid focusing on other image planes (determine the range of the best focal plane position).

[0076] 1. Continuously collect multiple short-exposure images within the atmospheric freeze time at the same focus. Use the spectral ratio method to compare the short-exposure images in this sequence to obtain the best image quality image at this focus as the focus reference image.

[0077] 2. For different foci, all use the method in step 1 to obtain the reference images at different foci as the focus adjustment image sequence.

[0078] 3. Use the spectral ratio method to compare the images in the focus adjustment image sequence to obtain the ratio results of different foci at the telescope's limit spatial resolution as the image quality of the focus adjustment sequence.

[0079] 4. Use the Gaussian fitting method to fit the image quality of the focus adjustment sequence, and then calculate the best focus adjustable for the device according to the fitting function.

[0080]

[0081] In the formula, a, b, and c represent the three coefficients of the Gaussian function, which are also the Gaussian fitting parameters. x represents different focus positions in this method, and G(x) represents the spectral ratio result.

[0082] 5. Drive the focus adjustment mechanism to drive the imaging device to the best focus.

[0083] Although specific embodiments of the present invention are disclosed for illustrative purposes, the purpose is to help understand the content of the present invention and implement it accordingly. Those skilled in the art can understand that: without departing from the spirit and scope of the present invention and the appended claims, various substitutions, changes, and modifications are possible. Therefore, the present invention should not be limited to the content disclosed in the best embodiments, and the scope of protection required by the present invention is defined by the scope of the claims.

Claims

1. An automatic focusing method based on a ground-based full-disk solar telescope, the steps of which include: 1) Determine the theoretical optimal focus, its adjustable range, and the adjustment step size of the ground-based full-disk solar telescope; 2) Perform focusing within the adjustable range of the theoretical optimal focus of the ground-based full-disk solar telescope according to the set adjustment step size; 3) For the focus after adjusting each adjustment step size, use the ground-based full-disk solar telescope to capture multiple short-exposure images within the atmospheric freeze time, and then use the spectral ratio method to screen the short-exposure image sequence corresponding to the current focus, and select the image with the best image quality as the reference image for the corresponding focus; 4) Obtain a focus adjustment image sequence based on the reference images of each focus; 5) Use the spectral ratio method to screen the focus adjustment image sequence, and obtain the ratio results of different foci at the limit spatial resolution of the ground-based full-disk solar telescope as the image quality of the focus adjustment sequence; 6) Use the Gaussian fitting method to fit the image quality of the focus adjustment sequence, and then calculate the adjustable optimal focus according to the fitting function; 7) Drive the imaging device of the optical system of the ground-based full-disk solar telescope according to the obtained adjustable optimal focus to drive the focus adjustment mechanism, and adjust the focus of the ground-based full-disk solar telescope to the corresponding optimal focus.

2. The method according to claim 1, wherein The image processing method of the spectral ratio method is: 21) Perform zero-mean normalization on each image in the image sequence; 22) Perform Fourier transform on each normalized image respectively to obtain the power spectrum of the image in the frequency domain; 23) Convert the power spectrum in the frequency domain from the Cartesian coordinate system to the polar coordinate system; 24) In the polar coordinate system, average the radius corresponding to each frequency along the angular direction to obtain an average curve representing the distribution of the power spectrum along the frequency direction; 25) Taking the distribution of the power spectrum of an image in the image sequence along the frequency direction as the reference power spectrum distribution L0(ρ) along the frequency direction, the distributions L n (ρ) of the power spectra of the remaining images are all compared with the distribution of the reference power spectrum along the frequency direction to obtain the image spectrum ratio results corresponding to each image 26) Select the image corresponding to the optimal image spectral ratio result as the best image quality image of the image sequence according to the image spectral ratio result C n (ρ) of each image.

3. The method according to claim 2, wherein Image spectral ratio result = Distribution of the power spectrum of the image along the frequency direction / Distribution of the reference power spectrum along the frequency direction.

4. The method according to claim 1 or 2 or 3, characterized in that, Determine the theoretical optimal focus and its adjustable range of the ground-based full-disk solar telescope according to the telescope optical system model of the ground-based full-disk solar telescope.

5. The method according to claim 1 or 2 or 3, characterized in that, Use the distribution of the power spectrum of the first image in the image sequence along the frequency direction as the reference.

Images