Cloud flare detection method based on SMAT configuration telescope

Abstract

The invention discloses a clouded flare detection method based on an SMAT configuration telescope. The clouded flare detection method comprises the following steps: 1) eliminating an image rotation difference between a photosphere image P and a color sphere image C; converting the photosphere images P into photosphere grey-scale maps and aligning the photosphere grey-scale maps; converting the color ball images C into color ball grey-scale images and aligning the color ball grey-scale images; then identifying the solar panel in each light ball grey-scale map and each color ball grey-scale map, and adjusting the image resolution of each light ball grey-scale map and each color ball grey-scale map to be consistent; 2) obtaining a normalized photosphere image light variation curve according to the processing of each photosphere grey-scale map, and obtaining a normalized color sphere image light variation curve according to the processing of each color sphere grey-scale map; 3) dividing the normalized chromatic ball image light variation curve by the normalized chromatic ball image light variation curve to obtain a light variation curve after the influence of the cloud layer is eliminated; 4) performing cloud layer elimination on the corresponding photosphere image P and the corresponding color sphere image C in the step 1) by using the obtained light variation curve; and 5) carrying out solar flare detection according to the light ball image P and the color ball image C after the cloud layer is eliminated.

Description

Technical Field

[0001] The present invention belongs to the technical field of astronomical observation, and relates to a method for detecting cloud-covered solar flares based on a telescope with a SMAT configuration. Background Art

[0002] The sun, as the closest star to the earth, its activities are not only directly related to the changes in the earth's climate, but also deeply affect the earth's space environment and many aspects of human society. Solar flares, as a kind of intense electromagnetic radiation burst event, the energy released by it is equivalent to the simultaneous explosion of billions of megaton nuclear bombs. These high-energy particles and electromagnetic waves spread out in all directions at extremely high speeds. When they reach the earth, they will have a significant impact on the earth's magnetic field, ionosphere and even the entire geomagnetic environment. For example, strong solar winds can cause satellite orbit shifts, damage satellite electronic equipment, interfere with radio communications, and even cause serious consequences such as power grid failures. Therefore, accurately predicting and timely monitoring solar flare activities is of extremely important significance for ensuring space safety, protecting communication facilities, and maintaining the country's critical infrastructure.

[0003] However, ground-based solar observations face many challenges, especially the limitations of weather conditions. The existence of the atmosphere not only absorbs sunlight, weakening or distorting the observation signal, but also introduces additional noise, making it difficult to identify and analyze solar flares. Traditional methods often have difficulty distinguishing the brightness changes caused by cloud cover from actual solar flare activities, which directly affects the reliability of the observation results.

[0004] Common cloud removal methods include homomorphic filtering, wavelet transform, median filtering, etc. Although these methods have achieved good results, they all affect the brightness of the original image after cloud removal and cannot restore the true brightness without clouds. The above methods cannot eliminate the influence of clouds on the brightness of solar flares, and require a large amount of computing resources and a long time. Therefore, they cannot be used for real-time observation of solar flare systems.

[0005] For example, the cloud removal algorithm for Ha images:

[0006] 1. First, select an image without cloud cover and obtain a standard limb darkening profile;

[0007] 2. Subtract this standard limb darkening profile from the image obscured by clouds to obtain a relatively flat solar disk;

[0008] 3. Then use median filtering to eliminate various solar activity phenomena on this image to obtain a cloud image;

[0009] 4. Finally, subtract the cloud image from the original image to obtain a restored solar disk image.

[0010] The above method has a good cloud removal effect for a single image. However, the median filtering method will affect the cloud information while removing solar activities. The above cloud removal method has cumbersome steps, requires a long time, and the restored image data cannot be used for flare monitoring.

[0011] Regarding the Ha cloud removal method based on the processing of the quiet solar chromosphere background, this method is also designed for the SMAT telescope. A special background image is constructed through the remaining information of the quiet solar limb darkening in the cloud coverage data, and then a flat-field data closer to the 'cloud absorption' feature is constructed by combining the original data with this background image to repair the image. This method is mainly used to repair the cloud flare data, but the repaired flare data cannot be used for flare monitoring, that is, the brightness value of the flare data cannot be restored. Summary of the Invention

[0012] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a method for detecting cloudy flares based on the SMAT configuration telescope, which is a fast and effective method for identifying cloudy flares, can greatly reduce the computational cost, improve the computational speed, and monitor flare eruptions in real time.

[0013] The present invention is applicable to the SMAT or a binocular telescope with a similar structure that can observe different solar levels, and a system for real-time flare monitoring can be built. By dividing the chromospheric light variation curve by the photospheric light variation curve, the present invention can effectively reduce the influence of clouds on the chromospheric brightness change.

[0014] The technical solution of the present invention is as follows:

[0015] A method for detecting cloudy flares based on the SMAT configuration telescope, the steps of which include:

[0016] 1) For the photospheric image P and chromospheric image C observed by the SMAT telescope, first eliminate the image rotation difference between the photospheric image P and the chromospheric image C; then convert each photospheric image P into a grayscale image to obtain a photospheric grayscale image, and align each photospheric grayscale image; convert each chromospheric image C into a grayscale image to obtain a chromospheric grayscale image, and align each chromospheric grayscale image; then identify the solar disk in each photospheric grayscale image and chromospheric grayscale image, and adjust the image resolution of each photospheric grayscale image and chromospheric grayscale image to be consistent;

[0017] 2) Process each photospheric grayscale image processed in step 1) to obtain a normalized photospheric image light variation curve; according to step 1)

[0018] Process each chromospheric grayscale image processed to obtain a normalized chromospheric image light variation curve;

[0019] 3) Divide the normalized chromosphere image light curve by the normalized photosphere image light curve to obtain the light curve F(t) after eliminating the influence of clouds;

[0020] 4) Use the obtained light curve F(t) to perform cloud elimination on the corresponding photosphere image P and chromosphere image C in step 1);

[0021] 5) Perform solar flare detection based on the photosphere image P and chromosphere image C after cloud elimination in step 4).

[0022] Furthermore, the light curve F(t) = C(t)·C3; where C1, C2, and C3 are all constants, C2 = P(t), T C (t) is the modulation curve of the cloud on the photosphere image P, T P (t) is the modulation curve of the cloud on the chromosphere image C, P(t) is the photosphere grayscale image, and C(t) is the chromosphere grayscale image.

[0023] Furthermore, the light curve where NPT(t) is the normalized photosphere image light curve and NCT(t) is the normalized chromosphere image light curve.

[0024] Furthermore, if using the obtained light curve F(t) to perform cloud elimination on the corresponding photosphere image P and chromosphere image C in step 1) cannot restore the peak value of the flare brightness; then in step 1), after identifying the solar disk in each photosphere grayscale image and chromosphere grayscale image, crop the edge part outside the solar disk in the image to improve the signal-to-noise ratio; then adjust the image resolutions of the cropped photosphere grayscale images and chromosphere grayscale images to be consistent.

[0025] Furthermore, the method for aligning each photosphere grayscale image is: select one image in the photosphere grayscale images as the photosphere reference image, and align the remaining photosphere grayscale images with the photosphere reference image; the method for aligning each chromosphere grayscale image is: select one image in the chromosphere grayscale images as the chromosphere reference image, and align the remaining chromosphere grayscale images with the chromosphere reference image; where for two images P1 and P2 to be aligned, perform the forward Fourier transform on them to obtain P1 and P2, and calculate R = P1*P1 * , P1 * is the complex conjugate of P1, P2 * is the complex conjugate of P2; calculate the inverse Fourier transform image r of R, and use the coordinates corresponding to the peak position of the image r as the translation offset between the two images P1 and P2.

[0026] Further, the method for identifying the solar disk in each photosphere grayscale image and chromosphere grayscale image is as follows: First, use the Canny edge detection algorithm to calculate the strong edges in the photosphere grayscale image and chromosphere grayscale image to obtain the corresponding edge images; then use the Hough transform to detect the circles in the edge images as the solar disk.

[0027] Further, by rotating the photosphere image P, the image rotation difference between the photosphere image P and the chromosphere image C is eliminated.

[0028] A server, comprising a memory and a processor, wherein the memory stores a computer program, and the computer program is configured to be executed by the processor, and the computer program includes instructions for executing the above method.

[0029] A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program realizes the above method when executed by a processor.

[0030] The advantages of the present invention are as follows:

[0031] The present invention has the characteristics of simple calculation steps and fast speed, can effectively restore the true brightness change when there are clouds, and can be used to monitor solar flares when there are clouds. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flowchart of the present invention.

[0033] Figure 2 are the photosphere and chromosphere images simultaneously observed by the SMAT telescope;

[0034] (a) Chromosphere image, (b) Photosphere image.

[0035] Figure 3 are the photosphere and chromosphere images with the same adjusted image resolution;

[0036] (a) Adjusted chromosphere image, (b) Adjusted photosphere image.

[0037] Figure 4 are the cropped photosphere and chromosphere images;

[0038] (a) Cropped chromosphere image, (b) Cropped photosphere image.

[0039] Figure 5 is the normalized light variation curve graph.

[0040] Figure 6 is the true curve graph for restoring the brightness change of the chromosphere image caused by flares. DETAILED DESCRIPTION OF THE INVENTION

[0041] 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.

[0042] The present invention proposes a method for monitoring cloud-covered solar flares based on the SMAT telescope configuration, aiming to effectively eliminate the influence of clouds on ground-based solar flare monitoring and improve the accuracy and reliability of observation data. The SMAT telescope has adjacent Ha telescope and magnetic field telescope, which observe the chromosphere at and the photosphere at respectively. Since the two telescopes are adjacent, we believe that the cloud occlusion is the same for the photospheric image and the chromospheric image. In addition, a solar flare is a violent energy release process in the solar atmosphere, mainly occurring in the chromosphere and corona of the sun, and causing drastic changes in the brightness and flux of the chromosphere and corona. For the solar photosphere, solar flares generally do not affect the brightness of the photosphere. Therefore, it can be considered that the influence of clouds on the SMAT photospheric image and chromospheric image is the same. By using the characteristic that the true brightness of the photosphere does not change during the occurrence of solar flares, the chromospheric brightness change affected by clouds is restored to achieve precise monitoring of solar flares in the presence of clouds. The processing steps of this method are as Figure 1 shown and are mainly divided into three parts:

[0043] 1. Obtain the photospheric and chromospheric data simultaneously observed by the SMAT telescope. Rotate the photospheric image 180 degrees clockwise to eliminate the image rotation difference between the chromosphere and the photosphere, and convert the image to a grayscale image. Align the chromospheric (hereinafter referred to as C) image and the photospheric (hereinafter referred to as P) image respectively, and then identify the edge of the solar disk and crop the image boundary outside the solar disk. Finally, adjust the resolution of the C and P images to ensure consistency. Since the size of the sun in the photospheric and chromospheric images is not the same, first crop the 'black background' part on the four sides of the solar disk to obtain Figure 3 , so that when cropping the flare occurrence area later, it can ensure that the cropping sizes of the C and P images are the same. In addition, after edge recognition and cropping of the 'black background' for the image data, it also achieves a certain degree of alignment effect (i.e., the C image and the P image data are aligned), making the sun in the C and P images look the same size.

[0044] 2. Calculate the light variation curves of the photospheric image and the chromospheric image respectively, and perform normalization processing.

[0045] 3. Divide the light variation curve of the chromospheric image by the light variation curve of the photospheric image to obtain the light variation curve after eliminating the influence of clouds.

[0046] The following are the detailed steps and explanations:

[0047] The specific processing method of step 1 is as follows:

[0048] 1.1. Obtain the photosphere and chromosphere image data simultaneously observed by the SMAT telescope. The observation bands are (photosphere) and (chromosphere), and the time resolution is about 12 s. Denote the chromosphere image data as C and the photosphere image data as P. The image data is as Figure 2 shown.

[0049] 1.2. The program reads the C and P image data, rotates the photosphere image clockwise by 180 degrees, and converts the image data into grayscale images. The formula for RGB to grayscale conversion is:

[0050] GrayValue = a1·R + a2·G + a3·B

[0051] where a1, a2, and a3 are coefficients. There is no unified standard for the coefficients, and it is only necessary to ensure that the coefficients are the same when performing grayscale image conversion on the C and P images.

[0052] 1.3. Due to factors such as the axis system movement of the ground telescope, the images appear jittery and offset. Image alignment is required. This method uses image correlation alignment. The specific method is as follows: Select the first image in the C and P image data as the reference image respectively, and align the remaining images with the reference image. Assume that F1 and F2 are the forward Fourier transforms of two images P1 and P2, and P1 * and P2 * are the complex conjugates of P1 and P2. Calculate R = P1 * P1 * . The inverse Fourier transform of R is r. Then the coordinates corresponding to the peak position of r are the translational offset between the two images.

[0053] 1.4. Use the Canny edge detection algorithm to calculate the edges of the images. The Canny edge detection algorithm is a classic edge detection method widely used in the fields of image processing and computer vision. It can effectively detect strong edges in images through multi-stage processing steps. Then use the Hough transform to detect circles in the edge images. The parameters of the detected circles include the center coordinates and the radius. The Hough Transform is a feature extraction technique used to detect simple shapes (such as lines, circles, etc.) from images. It realizes the detection of specific shapes by mapping the points in the image space to curves in the parameter space. After detecting the solar disk in the grayscale images corresponding to the C and P images, crop the image boundaries outside the solar disk in each grayscale image, and then adjust the image resolution sizes of each grayscale image to make them the same. Through the processing of the above steps, the grayscale images corresponding to the C and P images are as Figure 3As shown. If edge recognition is directly used for cropping, the effect is not good, and edge recognition needs to be performed on each image, which greatly increases the computational load. In the present invention, after the images are aligned, only one of the grayscale images corresponding to the C and P images needs to be subjected to edge recognition for cropping, and the same cropping can be performed on the remaining images, thus reducing the computational load and being convenient and fast.

[0054] 1.5. The increases in chromospheric brightness caused by different flares vary. For some flares with relatively weak peak brightness, after being affected by the Earth's atmosphere in various ways, their original signals are greatly distorted and the signal-to-noise ratio is too low. As a result, the chromospheric brightness changes reconstructed by this method cannot reflect the brightness changes caused by flares. Since the spatial scale of flares is much smaller than the solar disk, a small area of the flare occurrence region can be intercepted to improve the signal-to-noise ratio and the reconstruction effect of this method. Since the above steps have already aligned and edge-cropped the C and P data, when cropping the flare occurrence region, the cropping areas and positions of the grayscale images corresponding to the C and P images are kept consistent. The cropped effect is as Figure 4 shown.

[0055] The specific processing method of step 2 is as follows:

[0056] 2.1. Calculate the light curve, which is to calculate the change curve of the grayscale value of each image. The total grayscale value of a single image is the sum of the grayscale values of all pixel points.

[0057] For chromospheric images:

[0058]

[0059] For photospheric images:

[0060]

[0061] Under cloudless and atmosphere-free conditions, the brightnesses of the chromospheric image and the photospheric image change with time as C(t) and P(t) respectively. If there are clouds, assume that the modulation effects of the clouds on the chromospheric image and the photospheric image are T C (t) and T P (t) respectively. Therefore, in the case of clouds, the detected brightness changes of the chromosphere and the photosphere can be expressed as:

[0062] For chromospheric images: CT(t) = C(t)·T C (t)

[0063] For photospheric images: PT(t) = P(t)·T P (t).

[0064] 2.2. Normalize the CT(t) and PT(t) data. The processing method is to divide by the maximum value of the curve. For example, NCT(t) = CT(t) / max(CT(t)), NPT(t) = PT(t) / max(CT(t)). The normalized light curve is as follows Figure 5 shown. The two black vertical dashed lines in the figure are the corresponding time positions of the start and end of the flare. The red vertical dashed line is the position corresponding to the peak of the X-ray flux of the flare.

[0065] The specific processing method for step 3 is as follows:

[0066] 3.1. Compare the observed normalized curves NCT(t) and NPT(t) to obtain

[0067] 3.2. Based on the previous assumption: The influence of the cloud layer on the chromosphere and photosphere images is the same. Then it can be obtained that the ratio of the modulation functions of the cloud layer for the two is a constant, that is At the same time, considering that the photosphere brightness remains basically unchanged during the flare, that is, P(t) = C2, so where C1, C2, and C3 are all constants,

[0068] 3.3. The change of the obtained F(t) curve is only related to the chromosphere brightness change C(t) in the cloud-free case. That is, during the flare, F(t) can effectively remove the influence of cloud layer occlusion on the flare brightness. And through the above method, the brightness of an M1.5-class flare that occurred on July 3, 2024 was successfully restored correctly, proving the effectiveness of this method. As Figure 6 shown.

[0069] 3.4. The restored F(t) (true chromosphere brightness change) needs to be verified. The verification method is to compare the GOES satellite X-ray flux data and check whether there is a peak near the position corresponding to the peak of the X-ray flux of F(t). It is also possible to compare the Ha observation data of the Xihe satellite. If there is, it proves successful restoration. If not, it means the restoration effect is not good, and the flare occurrence area needs to be cropped, or it means that the observation data cannot be restored by this method. (This step is used to verify the feasibility of the method; when applied to flare monitoring, this step is not required) Similar to Figure 5 that, the two black vertical dashed lines in the figure are the corresponding time positions of the start and end of the flare. The red vertical dashed line is the position corresponding to the peak of the X-ray flux of the flare.

[0070] Although specific embodiments of the present invention are disclosed for illustrative purposes, which are intended to assist in understanding the content of the present invention and implementing it accordingly, those skilled in the art can understand that various substitutions, changes, and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the present invention should not be limited to the content disclosed in the preferred embodiments, and the scope of protection claimed by the present invention shall be defined by the scope defined in the claims.

Claims

1. A cloud flare detection method based on a SMAT configuration telescope, the steps comprising: 1) For the photospheric image P and the chromospheric image C observed by the SMAT telescope, first eliminate the image rotation difference between the photospheric image P and the chromospheric image C; then convert each photospheric image P into a grayscale image to obtain a photospheric grayscale image, and align each photospheric grayscale image; convert each chromospheric image C into a grayscale image to obtain a chromospheric grayscale image, and align each chromospheric grayscale image; then identify the solar disk in each photospheric grayscale image and chromospheric grayscale image, and adjust the image resolution of each photospheric grayscale image and chromospheric grayscale image to be consistent; 2) According to the grayscale images of the photospheric image processed in step 1), a normalized light curve of the photospheric image is obtained; According to the grayscale images of the chromosphere processed in step 1), a normalized light curve of the chromosphere image is obtained; 3) Divide the normalized chromospheric light curve by the normalized photospheric light curve to obtain the light curve F(t) after eliminating the cloud effect; 4) Using the obtained light curve F(t), the corresponding photospheric image P and chromospheric image C in step 1) are cloud-free; 5) Perform solar flare detection on the photospheric image P and the chromospheric image C after removing the clouds according to step 4).

2. The method according to claim 1, characterized in that Light curve F(t) = C(t)·C3; where C1, C2, and C3 are all constants. C2=P(t),T C (t) is the modulation curve of the cloud layer on the photospheric image P, T P (t) is the modulation curve of the cloud layer on the chromospheric image C, P(t) is the photospheric grayscale image, and C(t) is the chromospheric grayscale image.

3. The method according to claim 2, characterized in that Light curve Among them, NPT(t) is the normalized photospheric light curve, and NCT(t) is the normalized chromospheric light curve.

4. The method according to claim 1, 2 or 3, characterized in that: If the obtained light curve F(t) is used to remove the clouds from the corresponding photospheric image P and chromospheric image C in step 1), the peak brightness of the flare cannot be restored; then in step 1), after identifying the solar disk in each photospheric grayscale image and chromospheric grayscale image, the edge portion outside the solar disk in the image is cropped to improve the signal-to-noise ratio; then the image resolution of the cropped photospheric grayscale image and chromospheric grayscale image is adjusted to be consistent.

5. The method according to claim 1, 2 or 3, characterized in that: The method for aligning the photospheric grayscale images is as follows: one image is selected from the photospheric grayscale images as the photospheric reference image, and the remaining photospheric grayscale images are aligned with the photospheric reference image; the method for aligning the chromosphere grayscale images is as follows: one image is selected from the chromosphere grayscale images as the chromosphere reference image, and the remaining chromosphere grayscale images are aligned with the chromosphere reference image; For the two images P1 and P2 to be aligned, a positive Fourier transform is performed to obtain P1 and P2, and R = P1*P1 is calculated. * , P1 * is the complex conjugate of P1, P2 * is the complex conjugate of P2; calculate the inverse Fourier transform image r of R, and use the coordinates corresponding to the peak position of image r as the translation offset between the two images P1 and P2.

6. The method according to claim 1, 2 or 3, characterized in that: The method for identifying the solar disk in each photospheric grayscale image and chromospheric grayscale image is as follows: first, the strong edges in the photospheric grayscale image and the chromospheric grayscale image are calculated using the Canny edge detection algorithm to obtain the corresponding edge image; then, the circle detected in the edge image using the Hough transform is used as the solar disk.

7. The method according to claim 1, 2 or 3, characterized in that: By rotating the photospheric image P, the image rotation difference between the photospheric image P and the chromospheric image C is eliminated.

8. A server, characterized in that: The invention comprises a memory and a processor, wherein the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program comprises instructions for executing the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images