Satellite lightning data joint denoising method and system

Abstract

The application discloses a satellite lightning data joint denoising method and system, and the method comprises the following steps: acquiring a set of original lightning events of a satellite lightning imager in a target observation period and corresponding background meteorological cloud images, unifying the set of original lightning events and the background meteorological cloud images to the same projection coordinate system, and establishing a spatial index relationship between event points and background cloud image elements; the set of original lightning events comprises a plurality of event points; a space-time neighborhood is constructed with the event points as the center; the space-time density response values of the event points in the space-time neighborhood are calculated, the event points with the space-time density response values lower than a preset density threshold value are determined as isolated noise and removed; the flare risk geometry angle of the event points with the space-time density response values not less than the preset density threshold value is calculated, the inhibition weight of radiant energy is applied in combination with the local brightness gradient of the background meteorological cloud image, an effective lightning event set after joint denoising is output, and efficient and accurate purification of satellite lightning original data is realized.

Description

Technical Field

[0001] This invention relates to the field of meteorological remote sensing data processing and computer data visualization technology, and in particular to a method and system for joint denoising of satellite lightning data. Background Technology

[0002] Currently, the Lightning Imager (LMI) carried by geostationary meteorological satellites (such as the Fengyun-4 series) can continuously observe lightning activity over large areas, providing crucial data support for lightning warnings and convective cloud monitoring. However, due to limitations imposed by the satellite's on-orbit operating environment and optical imaging geometry, the raw lightning pulse event set contains various types of spurious signals, severely restricting the signal purity and reliability of subsequent lightning dynamic cloud image generation. Therefore, a joint denoising method that comprehensively utilizes spatiotemporal statistical features and optical geometric constraints is urgently needed to achieve efficient and accurate purification of raw satellite lightning data.

[0003] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for joint denoising of satellite lightning data, which solves the technical problems in the prior art, such as the presence of various types of false signals mixed in the original lightning pulse event set, which seriously restricts the signal purity and reliability of subsequent lightning dynamic cloud map generation, thereby achieving efficient and accurate purification of the original satellite lightning data.

[0005] To achieve the above objectives, this invention provides a joint denoising method for satellite lightning data, comprising the following steps:

[0006] S1. Obtain the original lightning event set and the corresponding background meteorological cloud image during the target observation period by the satellite lightning imager. Unify the original lightning event set and the background meteorological cloud image to the same projection coordinate system and establish a spatial index relationship between event points and background cloud image elements. The original lightning event set contains multiple event points, and each event point includes at least spatial coordinates, occurrence time and radiation energy.

[0007] S2. Construct a spatiotemporal neighborhood centered on the event point;

[0008] S3. Calculate the spatiotemporal density response value of each event point in the spatiotemporal neighborhood, and determine the event points with spatiotemporal density response values ​​lower than the preset density threshold as isolated noise and remove them.

[0009] S4. Calculate the flare risk geometry of event points whose spatiotemporal density response value is not less than a preset density threshold, and apply a suppression weight to the radiation energy in combination with the local brightness gradient of the background meteorological cloud map, and output the effective lightning event set after joint denoising.

[0010] In one embodiment of the present invention, calculating the spatiotemporal density response value of each event point in the spatiotemporal neighborhood includes: calculating the spatiotemporal density response value of the event point using a spatial kernel function and a temporal kernel function, wherein both the spatial kernel function and the temporal kernel function are truncated Gaussian kernel functions.

[0011] In one embodiment of the present invention, the spatially truncated Gaussian kernel function is:

[0012]

[0013] The time-truncation Gaussian kernel function is:

[0014]

[0015] in, This represents the spatial distance between the current event point and its adjacent event points. This represents the spatial neighborhood radius, used to limit the maximum spatial distance involved in the spatial kernel function calculation. For spatial kernel functions, For time kernel function, For spatial scale parameters, For time scale parameters, The time neighborhood threshold, The time truncation threshold is represented by the spatiotemporal density response value, which is the weighted cumulative contribution of the spatially truncated Gaussian kernel function and the time-truncated Gaussian kernel function to all adjacent event points in the spatiotemporal neighborhood.

[0016] In one embodiment of the present invention, the method further includes: calculating the flare risk geometric angle based on the solar zenith angle, satellite observation zenith angle, and relative azimuth angle between the solar azimuth angle and the satellite observation azimuth angle of the location of the event point whose spatiotemporal density response value is not less than a preset density threshold; and applying a suppression weight to the radiation energy of the event point when the flare risk geometric angle corresponding to the event point is less than a preset angle threshold and the local brightness gradient of the background meteorological cloud image is higher than a preset gradient threshold.

[0017] In one embodiment of the present invention, the formula for calculating the suppression weight is:

[0018]

[0019] in, For the geometric angle of flare risk, For the preset angle threshold, To preset the gradient threshold, This is the flare suppression intensity coefficient. To prevent extremely small positive numbers with a denominator of zero.

[0020] In one embodiment of the present invention, the effective energy weight of the event point is determined by the product of the spatiotemporal density discrimination result and the flare suppression weight; when the spatiotemporal density discrimination result is retained, the effective energy weight is equal to the flare suppression weight; when the spatiotemporal density discrimination result is eliminated, the effective energy weight is zero; the corrected event radiation energy is equal to the product of the original radiation energy and the effective energy weight.

[0021] In one embodiment of the present invention, unifying the original lightning event set and the background meteorological cloud image into the same projection coordinate system includes: if the original lightning event points are represented by latitude and longitude, then their projection is converted to a preset output plane; if the original lightning event points are represented by satellite payload probe coordinates, row and column numbers, or scan coordinates, then they are converted into unified projection coordinates according to satellite positioning parameters, payload geometric parameters, and the Earth ellipsoid model; the background meteorological cloud image is one or more of visible light cloud images, infrared cloud images, and enhanced cloud images.

[0022] In one embodiment of the present invention, the spatial neighborhood radius of the spatiotemporal neighborhood ranges from 1 to 5 output grid sizes, and the time threshold ranges from 100 milliseconds to 1000 milliseconds; the preset density threshold is set according to the load noise level, observation area, day and night conditions, or historical statistical data.

[0023] This invention also provides a joint denoising system for satellite lightning data, comprising:

[0024] The data acquisition unit is used to acquire the original lightning event set and the corresponding background meteorological cloud image during the target observation period by the satellite lightning imager, unify the original lightning event set and the background meteorological cloud image to the same projection coordinate system, and establish a spatial index relationship between event points and background cloud image elements; the original lightning event set contains multiple event points, and each event point includes at least spatial coordinates, occurrence time and radiation energy.

[0025] A neighborhood construction unit is used to construct a spatiotemporal neighborhood centered on the event point;

[0026] The first calculation unit is used to calculate the spatiotemporal density response value of each event point in the spatiotemporal neighborhood, and to determine and remove event points whose spatiotemporal density response value is lower than a preset density threshold as isolated noise.

[0027] The second calculation unit is used to calculate the flare risk geometry of event points whose spatiotemporal density response value is not less than a preset density threshold, and to apply a suppression weight to the radiation energy in combination with the local brightness gradient of the background meteorological cloud map, and output the effective lightning event set after joint denoising.

[0028] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described satellite lightning data joint denoising method.

[0029] Compared with the prior art, the satellite lightning data joint denoising method and system according to the present invention has at least the following beneficial effects:

[0030] First, this invention can improve the purity of satellite lightning event data and reduce the impact of false lightning signals on subsequent applications. Instead of filtering lightning events using only a single energy or brightness threshold, this invention combines the spatiotemporal density characteristics of the lightning event itself with the optical geometric relationship between the sun, satellite, and pixel to jointly denoise the original lightning events. On the one hand, by constraining the spatiotemporal density, isolated random noise caused by high-energy particle impacts on detectors and transient instrument responses can be effectively eliminated. On the other hand, by jointly judging the flare risk geometry angle and the local brightness gradient of the background cloud image, optical false alarms caused by daytime solar flares, high-brightness cloud top reflections, and strong cloud boundary reflections can be suppressed.

[0031] Second, the denoising strategy of this invention employs weighted suppression rather than simple deletion, which can suppress false signals while preserving real lightning events. Event points with insufficient spatiotemporal density are removed, and energy suppression weights are applied to event points located in high-risk flare regions, thus avoiding the problem of excessively removing real daytime lightning events.

[0032] Third, this invention, through a dual constraint mechanism of spatiotemporal density discrimination and flare suppression, can adapt to noise characteristics under different observation conditions. In nighttime observations, isolated noise dominates, and spatiotemporal density constraints play a major role; in daytime observations, false alarms from flares increase significantly, and optical geometric constraints play a key role, with the two mechanisms working synergistically and complementing each other. Attached Figure Description

[0033] Figure 1 This is a flowchart of a satellite lightning data joint denoising method according to an embodiment of the present invention;

[0034] Figure 2 This is a structural diagram of a satellite lightning data joint denoising system according to an embodiment of the present invention. Detailed Implementation

[0035] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0036] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0037] like Figure 1 As shown, a satellite lightning data joint denoising method according to a preferred embodiment of the present invention includes the following steps:

[0038] S1. Obtain the original lightning event set and the corresponding background meteorological cloud image during the target observation period by the satellite lightning imager. Unify the original lightning event set and the background meteorological cloud image to the same projection coordinate system and establish a spatial index relationship between event points and background cloud image elements. The original lightning event set contains multiple event points, and each event point includes at least spatial coordinates, occurrence time and radiation energy.

[0039] Specifically, let the set of raw lightning pulse events acquired by the satellite lightning imager during the observation period be:

[0040] Among them, any original lightning pulse event Represented as:

[0041]

[0042] In the formula, and These represent the spatial coordinates of the event point in the unified projection plane or the latitude-longitude mapping plane, respectively. Indicates the time when the event occurred. This indicates the lightning radiation energy or intensity corresponding to the event.

[0043] Background meteorological cloud images are represented as two-dimensional grayscale matrices in the same projected coordinate system.

[0044] in This represents the two-dimensional cell or grid index in the output image. This indicates the frame number of the dynamic image. The background meteorological cloud image can be a visible light cloud image, an infrared cloud image, an enhanced cloud image, or other satellite cloud images that can reflect the spatial distribution and brightness texture characteristics of clouds.

[0045] The original lightning event set and the background meteorological cloud image are unified to the same projected coordinate system. Specifically, each lightning event point is mapped to the corresponding pixel position in the output grid, and a spatial index relationship between the event point and the background cloud image pixels is established. This index relationship is used to subsequently extract the background brightness, local gradient, and solar-satellite-pixel geometric parameters at the location of the event point.

[0046] The following mapping relationship is constructed between the original lightning event set and the background meteorological cloud image:

[0047]

[0048] in, For the first The final dynamic cloud map after fusing lightning radiation features and background cloud texture features.

[0049] Based on event points Construct a spatiotemporal neighborhood centered on the event point. The spatial location is:

[0050]

[0051] And with event points Constructing a spatiotemporal neighborhood around the center:

[0052]

[0053] in, The radius of the spatial neighborhood. Indicates the first The spatial location of each event point; Indicates the first The time of occurrence of each event point The time neighborhood threshold, It represents Euclidean distance.

[0054] Calculate the spatiotemporal density response value of each event point within its spatiotemporal neighborhood. Define the event point. The spatiotemporal density response value is:

[0055] in, For spatial kernel functions, This is the time kernel function.

[0056] Based on the above spatiotemporal density response values, a spatiotemporal density discriminant function is constructed:

[0057]

[0058] in, This is the spatiotemporal density discrimination threshold, i.e., the preset density threshold. When When, it indicates the event point. Lacking support from adjacent discharge events, it was judged as isolated noise and discarded; when When, it indicates the point in time of the event. The spatiotemporal correlation characteristics required for real lightning events are retained.

[0059] For event points detected by spatiotemporal neighborhood density, i.e., event points whose spatiotemporal density response value is not less than a preset density threshold, flare risk judgment and event energy correction are further performed.

[0060] Based on the spatial location of the event point and the observation time, obtain the solar zenith angle corresponding to that location. Satellite observation zenith angle And the relative azimuth angle between the solar azimuth angle and the satellite observation azimuth angle. And calculate the flare risk geometry. :

[0061]

[0062] in, The smaller the value, the closer the solar incident direction is to the satellite observation direction, which is a mirror reflection sensitive geometric condition, and the higher the probability of a false alarm of optical flare in the corresponding region.

[0063] Simultaneously, using the index relationship established in step S1, the local brightness gradient near the background cloud image pixel corresponding to the event point is extracted. When the flare risk geometry angle corresponding to the event point... If the angle is less than a preset threshold and the local brightness gradient of the background cloud map is higher than a preset gradient threshold, the event point is considered to be located in a high-risk area of ​​solar flare or strong cloud top reflection, and a suppression weight is applied to its radiation energy.

[0064] After completing the spatiotemporal density discrimination and flare suppression, a set of effective lightning events after joint denoising is formed.

[0065] This embodiment further defines the specific calculation method for the spatiotemporal density response value in step S3.

[0066]

[0067] The time kernel function uses a truncated Gaussian kernel:

[0068]

[0069] In the formula, For spatial scale parameters, For time scale parameters, This represents the spatial distance between the current event point and its adjacent event points, preferably Euclidean distance. This represents the spatial neighborhood radius or spatial truncation threshold, used to limit the maximum spatial distance involved in the spatial kernel function calculation. For spatial kernel functions, For time kernel function, The time neighborhood threshold, The time truncation threshold is used to limit the maximum time interval for participating in the calculation of the time kernel function. The spatiotemporal density response value is the weighted cumulative contribution of the spatially truncated Gaussian kernel function and the time-truncated Gaussian kernel function to all adjacent event points in the spatiotemporal neighborhood.

[0070] Spatiotemporal density response value The weighted cumulative contribution of the spatially truncated Gaussian kernel function and the temporally truncated Gaussian kernel function to all adjacent event points in the spatiotemporal neighborhood is the specific expansion of equation (7).

[0071] The design of the truncated Gaussian kernel function makes the contribution of neighboring event points that are closer to the current event point greater, and the contribution of event points that are farther away smaller. Event points that exceed the truncation radius or time threshold contribute zero, thereby reducing computational overhead while ensuring the locality of density estimation.

[0072] This embodiment further defines the specific implementation method of flare suppression in step S4. The local brightness gradient of the background cloud map is defined as:

[0073]

[0074] Set event point The cell position mapped to the background cloud image is When a target pixel is under strong flare geometry conditions and the background brightness gradient is abnormally increased, flare suppression weights are triggered. The formula for calculating the suppression weights is:

[0075]

[0076] in, The geometric angle threshold for flare risk, i.e., the preset angle threshold. This is the background brightness gradient threshold, i.e., the preset gradient threshold. This is the flare suppression intensity coefficient. To prevent extremely small positive numbers with a denominator of zero.

[0077] From equation (13), it can be seen that when When the background brightness gradient does not exceed the threshold, the exponential term approaches zero. When the value is close to 1, the event energy is basically preserved; when And when the background brightness gradient exceeds the threshold, It decays exponentially as the risk of flares increases, thereby reducing the impact of optical false alarm events on the generation of subsequent lightning cloud maps.

[0078] The design of this suppression weight ensures that the smaller the flare risk geometry angle (i.e., the closer it is to the solar reflection geometry), the stronger the suppression; conversely, the larger the background brightness gradient (i.e., the more likely there is a bright reflective surface), the stronger the suppression. The combined effect of these two factors achieves dual suppression through geometric constraints and radiative gradients.

[0079] Furthermore, by integrating spatiotemporal density constraints and flare suppression constraints, the original event point is defined. The effective energy weight is:

[0080]

[0081] When the spatiotemporal density discrimination result When the value is 1 (reserved), the effective energy weight is equal to the flare suppression weight. When the spatiotemporal density discrimination result When the value is 0 (removed), the effective energy weight is zero.

[0082] The corrected event radiation energy is:

[0083]

[0084] That is, the corrected radiation energy is equal to the product of the original radiation energy and the effective energy weight. Through this weighting mechanism, isolated noise events are completely eliminated (weight is zero), false alarm events of flares are moderately suppressed (weight is greater than zero but less than 1), and real lightning events maintain a high weight, thereby achieving refined joint denoising.

[0085] The set of valid lightning events after joint denoising is represented as follows:

[0086]

[0087] That is, only event points with corrected radiated energy greater than zero are retained in the effective lightning event set.

[0088] In a preferred embodiment, unifying the original lightning event set and the background meteorological cloud image to the same projected coordinate system includes:

[0089] If the original lightning event points are represented in latitude and longitude, they are projected onto a preset output plane. If the original lightning event points are represented in satellite payload detector coordinates, row and column numbers, or scan coordinates, they are converted to unified projection coordinates based on satellite positioning parameters, payload geometric parameters, and the Earth ellipsoid model. Simultaneously, the background meteorological cloud image is resampled, cropped, or stitched to ensure it has the same spatial extent, projection method, pixel size, and image dimensions as the output image. The background meteorological cloud image can be one or more of the following: visible light cloud image, infrared cloud image, or enhanced cloud image.

[0090] In a preferred embodiment, the spatial neighborhood radius The range can be from 1 to 5 output grid sizes, with a time threshold. The range can be from 100 milliseconds to 1000 milliseconds. Preset density threshold. The settings are based on the load noise level, observation area, day / night conditions, or historical statistical data.

[0091] like Figure 2 As shown, the present invention also provides a joint denoising system for satellite lightning data, comprising:

[0092] The data acquisition unit 201 is used to acquire the original lightning event set and the corresponding background meteorological cloud image of the satellite lightning imager during the target observation period, unify the original lightning event set and the background meteorological cloud image to the same projection coordinate system, and establish a spatial index relationship between event points and background cloud image elements; the original lightning event set contains multiple event points, and each event point includes at least spatial coordinates, occurrence time and radiation energy.

[0093] Neighborhood construction unit 202 is used to construct a spatiotemporal neighborhood centered on the event point;

[0094] The first calculation unit 203 is used to calculate the spatiotemporal density response value of each event point in the spatiotemporal neighborhood, and to determine and remove event points whose spatiotemporal density response value is lower than a preset density threshold as isolated noise.

[0095] The second calculation unit 204 is used to calculate the flare risk geometric angle of event points whose spatiotemporal density response value is not less than a preset density threshold, and to apply a suppression weight to the radiation energy in combination with the local brightness gradient of the background meteorological cloud map, and output the effective lightning event set after joint denoising.

[0096] The specific implementation methods of each unit are the same as the implementation methods of the corresponding steps in the method embodiment, and will not be repeated here.

[0097] In one embodiment, the system further includes a storage unit for storing the original lightning event set, background meteorological cloud image, spatiotemporal density response value, flare risk geometry, suppression weight, and the effective lightning event set after joint denoising.

[0098] In one embodiment, the system further includes an output unit for outputting the combined denoised effective lightning event set to the subsequent lightning dynamic cloud map generation module.

[0099] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A joint denoising method for satellite lightning data, characterized in that, Includes the following steps: S1. Obtain the original lightning event set and the corresponding background meteorological cloud image during the target observation period by the satellite lightning imager. Unify the original lightning event set and the background meteorological cloud image to the same projection coordinate system and establish a spatial index relationship between event points and background cloud image elements. The original lightning event set contains multiple event points, and each event point includes at least spatial coordinates, occurrence time and radiation energy. S2. Construct a spatiotemporal neighborhood centered on the event point; S3. Calculate the spatiotemporal density response value of each event point in the spatiotemporal neighborhood, and determine the event points with spatiotemporal density response values ​​lower than the preset density threshold as isolated noise and remove them. S4. Calculate the flare risk geometry of event points whose spatiotemporal density response value is not less than a preset density threshold, and apply a suppression weight to the radiation energy in combination with the local brightness gradient of the background meteorological cloud map, and output the effective lightning event set after joint denoising.

2. The method according to claim 1, characterized in that, Calculating the spatiotemporal density response value of each event point within the spatiotemporal neighborhood includes: calculating the spatiotemporal density response value of the event point using a spatial kernel function and a temporal kernel function, wherein both the spatial kernel function and the temporal kernel function are truncated Gaussian kernel functions.

3. The method according to claim 2, characterized in that, The spatially truncated Gaussian kernel function is: The time-truncation Gaussian kernel function is: in, This represents the spatial distance between the current event point and its adjacent event points. Represents the spatial neighborhood radius. For spatially truncated Gaussian kernel functions, For time-truncated Gaussian kernel functions, For spatial scale parameters, For time scale parameters, The time neighborhood threshold, The time truncation threshold is represented by the spatiotemporal density response value, which is the weighted cumulative contribution of the spatially truncated Gaussian kernel function and the time-truncated Gaussian kernel function to all adjacent event points in the spatiotemporal neighborhood.

4. The method according to claim 1, characterized in that, The method further includes: calculating the flare risk geometric angle based on the solar zenith angle, satellite observation zenith angle, and relative azimuth angle between the solar azimuth angle and the satellite observation azimuth angle of the location of the event point whose spatiotemporal density response value is not less than a preset density threshold; and applying a suppression weight to the radiation energy of the event point when the flare risk geometric angle corresponding to the event point is less than a preset angle threshold and the local brightness gradient of the background meteorological cloud image is higher than a preset gradient threshold.

5. The method according to claim 4, characterized in that, The formula for calculating the suppression weight is: in, For the geometric angle of flare risk, For the preset angle threshold, To preset the gradient threshold, This is the flare suppression intensity coefficient. To prevent extremely small positive numbers with a denominator of zero.

6. The method according to claim 1, characterized in that, The method further includes: the effective energy weight of the event point is determined by the product of the spatiotemporal density discrimination result and the flare suppression weight; when the spatiotemporal density discrimination result is retained, the effective energy weight is equal to the flare suppression weight; when the spatiotemporal density discrimination result is eliminated, the effective energy weight is zero; the corrected event radiation energy is equal to the product of the original radiation energy and the effective energy weight.

7. The method according to claim 1, characterized in that, Unifying the original lightning event set and background meteorological cloud image into the same projection coordinate system includes: if the original lightning event points are represented by latitude and longitude, then their projection is converted to a preset output plane; if the original lightning event points are represented by satellite payload probe coordinates, row and column numbers, or scan coordinates, then they are converted into unified projection coordinates according to satellite positioning parameters, payload geometric parameters, and the Earth ellipsoid model; the background meteorological cloud image is one or more of visible light cloud images, infrared cloud images, and enhanced cloud images.

8. The method according to claim 1, characterized in that, The spatial neighborhood radius of the spatiotemporal neighborhood ranges from 1 to 5 output grid sizes, and the time threshold ranges from 100 milliseconds to 1000 milliseconds; the preset density threshold is set according to the load noise level, observation area, day and night conditions, or historical statistical data.

9. A joint denoising system for satellite lightning data, characterized in that, include: The data acquisition unit is used to acquire the original lightning event set and the corresponding background meteorological cloud image during the target observation period by the satellite lightning imager, unify the original lightning event set and the background meteorological cloud image to the same projection coordinate system, and establish a spatial index relationship between event points and background cloud image elements; the original lightning event set contains multiple event points, and each event point includes at least spatial coordinates, occurrence time and radiation energy. A neighborhood construction unit is used to construct a spatiotemporal neighborhood centered on the event point; The first calculation unit is used to calculate the spatiotemporal density response value of each event point in the spatiotemporal neighborhood, and to determine and remove event points whose spatiotemporal density response value is lower than a preset density threshold as isolated noise. The second calculation unit is used to calculate the flare risk geometry of event points whose spatiotemporal density response value is not less than a preset density threshold, and to apply a suppression weight to the radiation energy in combination with the local brightness gradient of the background meteorological cloud map, and output the effective lightning event set after joint denoising.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the satellite lightning data joint denoising method as described in any one of claims 1 to 8.

Images