A matching pursuit method and device for a satellite cloud cluster at a current time

Abstract

The matching tracking method for adjacent time satellite cloud clusters disclosed in the application comprises the following steps: acquiring a first observed cloud image and a first brightness temperature base image at a current time, and a second observed cloud image and a second brightness temperature base image at a next time; performing cloud cluster region identification on the cloud images according to a map at the current time, and performing cloud cluster region identification on the cloud images according to the base image at the next time, to obtain a cloud cluster region set at the current time and a cloud cluster region set at the next time; extracting a first feature parameter set corresponding to the cloud cluster region set at the current time from the first observed cloud image, and extracting a second feature parameter set corresponding to the cloud cluster region set at the next time from the second observed cloud image; performing cloud cluster region matching processing on the cloud cluster region set at the current time and the cloud cluster region set at the next time according to the first feature parameter set and the second feature parameter set, to obtain a target matching cloud cluster pair set; and generating a cloud cluster tracking result according to the target matching cloud cluster pair set. The method can more accurately perform cloud cluster tracking.

Description

Technical Field

[0001] This application relates to the field of satellite image technology, specifically to a matching and tracking method and apparatus for phase-temporal satellite cloud clusters. Background Technology

[0002] Currently, geostationary satellite cloud images are a crucial tool for monitoring the dynamic evolution of cloud clusters. Their continuous observation capabilities are of core value in tracing the life cycle of cloud clusters and improving the accuracy of weather forecasts. Existing methods typically rely on forecasters manually observing and comparing cloud images from adjacent time periods to match and track cloud clusters in different satellite images at nearby moments. This method is not only inefficient and highly subjective, but also struggles to handle complex morphological changes such as splitting and merging that occur during cloud movement. This results in severely insufficient accuracy in cloud tracking results, failing to meet the demands of automated, high-precision operational applications. Summary of the Invention

[0003] In view of the above problems, this application provides a matching and tracking method and apparatus for phase-temporal satellite cloud clusters, which can solve the problem of insufficient accuracy of existing cloud cluster tracking results.

[0004] Firstly, this application provides a matching and tracking method for phase-temporal satellite cloud clusters, including: Obtain the first observation cloud image and the first brightness temperature background image at the current moment, and the second observation cloud image and the second brightness temperature background image at the next moment; wherein, the first observation cloud image and the first brightness temperature background image correspond spatially, and the second observation cloud image and the second brightness temperature background image correspond spatially. Based on the first brightness temperature background map, cloud cluster regions are identified in the first observed cloud map to obtain the current set of cloud cluster regions. Based on the second brightness temperature background map, cloud cluster regions are identified in the second observed cloud map to obtain the next set of cloud cluster regions. Based on the first observed cloud map, extract the first set of feature parameters corresponding to the cloud cluster region set at the current time, and based on the second observed cloud map, extract the second set of feature parameters corresponding to the cloud cluster region set at the next time. Based on the first set of feature parameters and the second set of feature parameters, cloud cluster region matching processing is performed on the current cloud cluster region set and the next cloud cluster region set to obtain a target matching cloud cluster pair set. Cloud cluster tracking results are generated based on the target matching cloud cluster pair set.

[0005] In the above technical solution, the method can determine the cloud cluster identification benchmark at different times through the brightness temperature base map and the observed cloud map, ensuring the accuracy and spatial consistency of the extraction of each cloud cluster region; at the same time, quantitative matching based on the feature parameter set can replace manual comparison to realize the automation of cloud cluster tracking, thereby effectively solving the matching problem caused by cloud cluster splitting and merging, and thus significantly improving the accuracy and reliability of cloud cluster tracking results.

[0006] In some embodiments, the step of identifying cloud regions in the first observed cloud image based on the first brightness temperature background image to obtain the current set of cloud regions, and identifying cloud regions in the second observed cloud image based on the second brightness temperature background image to obtain the next set of cloud regions, includes: Based on the first brightness temperature background map, cloud detection processing is performed on the first observed cloud map to obtain the first cloud detection result, and the first cloud detection result is binarized to obtain the first binarized result map; Based on the second brightness temperature base map, cloud detection processing is performed on the second observed cloud map to obtain the second cloud detection result, and the second cloud detection result is binarized to obtain the second binarized result map; Perform connected component identification processing on the first binarized result image to obtain the set of cloud regions at the current time; The second binarized result image is processed for connected component identification to obtain the set of cloud regions at the next time step.

[0007] In the above technical solution, the method can complete accurate cloud detection and binarization segmentation based on the brightness temperature base map, and extract independent cloud cluster regions through connected component identification, effectively eliminating non-cloud interference and ensuring the integrity and independence of the cloud cluster region set at the current and next time moments.

[0008] In some implementations, the step of performing cloud region matching processing on the current cloud region set and the next cloud region set based on the first feature parameter set and the second feature parameter set to obtain a target matching cloud pair set includes: Based on the first set of feature parameters and the second set of feature parameters, the area overlap method is used to perform cloud region matching processing on the current cloud region set and the next cloud region set to obtain a preliminary set of matched cloud pairs. Based on the preliminary matched cloud cluster pair set, determine the first set of unmatched cloud cluster regions remaining in the current cloud cluster region set and the second set of unmatched cloud cluster regions remaining in the next cloud cluster region set; The first set of remaining cloud regions and the second set of remaining cloud regions are matched based on individual feature similarity to obtain a set of supplementary matching cloud pairs. The total set of cloud cluster pairs, consisting of the initial set of matched cloud cluster pairs and the supplementary set of matched cloud cluster pairs, is integrated and optimized to obtain the target set of matched cloud cluster pairs.

[0009] In the above technical solution, the method can effectively solve the matching ambiguity problem caused by cloud splitting, merging and edge blurring through hierarchical matching and integration optimization strategy. Thus, while ensuring the high efficiency of matching the main cloud, it can achieve accurate matching of weakly related cloud, thereby significantly reducing the omission and mismatch rate, and finally obtaining cloud matching results with both integrity and consistency.

[0010] In some implementations, the step of performing cloud region matching processing on the current cloud region set and the next cloud region set using the area overlap method based on the first feature parameter set and the second feature parameter set to obtain a preliminary set of matched cloud pairs includes: Calculate multiple first areas in the current cloud region set that correspond one-to-one with multiple first cloud regions based on the first set of feature parameters; Calculate the multiple second areas in the cloud region set at the next moment, which correspond one-to-one with the multiple second cloud regions, based on the second set of feature parameters; Multiple first cloud cluster regions in the current cloud cluster region set are combined with multiple second cloud cluster regions in the next cloud cluster region set in pairs to obtain multiple cloud cluster region combination pairs; Based on the first set of feature parameters and the second set of feature parameters, calculate multiple area overlap rates that correspond one-to-one with the multiple cloud region combination pairs; wherein, the preliminary matching cloud pair set includes all cloud region combination pairs whose area overlap rate exceeds a preset overlap rate threshold.

[0011] In the above technical solution, the method can filter out non-homogeneous clouds with large spatial differences by quantitatively calculating the area overlap rate between cloud clusters and setting a threshold. This replaces manual experience judgment with objective quantitative indicators, thereby ensuring that the preliminary matching results have a high degree of confidence.

[0012] In some implementations, the step of performing matching processing based on individual feature similarity on the first set of remaining cloud regions and the second set of remaining cloud regions to obtain a supplementary set of matching cloud pairs includes: Based on the first set of feature parameters, obtain multiple first feature matching factors in the first set of remaining cloud regions that correspond one-to-one with multiple first remaining cloud regions; Calculate multiple second feature matching factors in the second set of remaining cloud regions that correspond one-to-one with multiple second remaining cloud regions based on the second set of feature parameters; Based on the plurality of first feature matching factors and the plurality of second feature matching factors, the individual feature similarity matching method is used to match the first set of remaining cloud regions with the second set of remaining cloud regions to obtain a set of supplementary matching cloud pairs.

[0013] In the above technical solution, the method can use multi-dimensional feature factors to quantify the individual attributes of cloud clusters, realize refined similarity matching of the remaining cloud clusters, effectively identify the homologous cloud clusters with insufficient spatial overlap due to splitting, merging or rapid deformation, thereby greatly reducing matching omissions and further improving the continuity of cloud cluster tracking.

[0014] In some implementations, the step of integrating and optimizing the total set of cloud cluster pairs, which consists of the initial set of matched cloud cluster pairs and the supplementary set of matched cloud cluster pairs, to obtain the target set of matched cloud cluster pairs includes: Statistical analysis is performed on the total set of cloud cluster pairs consisting of the preliminary matching cloud cluster pair set and the supplementary matching cloud cluster pair set to obtain the statistical analysis results. Based on the statistical analysis results, split cloud cluster regions exhibiting splitting phenomena and merged cloud cluster regions exhibiting merging phenomena are selected from multiple current-time cloud clusters in the total cloud cluster pair set. The next-time cloud cluster region paired with the split cloud cluster region and the next-time paired cloud cluster region paired with the merged cloud cluster region are obtained based on the total cloud cluster pair set. Based on the first set of feature parameters, the second set of feature parameters, and the cloud region at the next moment paired with the split cloud region, calculate the first area parameter corresponding to the split cloud region; Based on the first set of feature parameters, the second set of feature parameters, and the next time pairing cloud region with the merged cloud region, calculate the second area parameter corresponding to the merged cloud region; The pairing relationship of the split cloud region in the total cloud pair set is optimized according to the first area parameter to obtain an optimized cloud pair set. The pairing relationship of the merged cloud region in the optimized cloud pair set is optimized based on the second area parameter to obtain the target matching cloud pair set.

[0015] In the above technical solution, the method can accurately identify the splitting and merging of cloud clusters through statistical analysis, and combine area parameters to quantitatively optimize the pairing relationship, effectively resolve the matching conflict between one-to-many and many-to-one, thereby correcting the pairing deviation caused by complex morphological evolution, and outputting a logically consistent target matching result that fits the actual evolution process of cloud clusters.

[0016] In some implementations, the step of filtering split cloud regions exhibiting splitting and merged cloud regions exhibiting merging from multiple current-time cloud regions in the total cloud cluster set based on the statistical analysis results includes: Based on the statistical analysis results, obtain the number of times each of the total cloud cluster pairs corresponds to a cloud cluster region at a current time, and the number of times each of the total cloud cluster pairs corresponds to a cloud cluster region at a next time. The cloud regions in the total cloud pair set whose number of matches is greater than the threshold of the first match are identified as split cloud regions with splitting phenomena. The cloud region in the next moment in the total cloud region pair set that has been matched more than the second threshold number is determined as the target cloud region; The cloud cluster regions in the total cloud cluster pair set that match the target cloud cluster region at the current moment are identified as merged cloud cluster regions that exhibit merging phenomena.

[0017] In the above technical solution, the method can establish an objective basis for judging splitting and merging phenomena by quantitatively statistically analyzing the number of matching and being matched, thereby achieving automated and standardized identification of splitting and merging cloud clusters and avoiding misjudgments caused by experience.

[0018] Secondly, this application provides a matching and tracking device for phase-temporal satellite cloud clusters, comprising: The acquisition unit is used to acquire the first observation cloud map and the first brightness temperature background map at the current time, and the second observation cloud map and the second brightness temperature background map at the next time; wherein, the first observation cloud map and the first brightness temperature background map correspond spatially, and the second observation cloud map and the second brightness temperature background map correspond spatially. The identification unit is used to identify cloud regions in the first observed cloud map based on the first brightness temperature background map to obtain the current set of cloud regions, and to identify cloud regions in the second observed cloud map based on the second brightness temperature background map to obtain the next set of cloud regions. The feature extraction unit is used to extract a first set of feature parameters corresponding to the set of cloud regions at the current time based on the first observed cloud map, and to extract a second set of feature parameters corresponding to the set of cloud regions at the next time based on the second observed cloud map. The matching unit is used to perform cloud region matching processing on the current cloud region set and the next cloud region set according to the first feature parameter set and the second feature parameter set to obtain a target matching cloud pair set; The generation unit is used to generate cloud tracking results based on the target matching cloud pair set.

[0019] In the above technical solution, the device can determine the cloud cluster identification benchmark at different times through the brightness temperature base map and the observed cloud map, ensuring the accuracy and spatial consistency of the extraction of each cloud cluster region; at the same time, based on the feature parameter set, quantitative matching can replace manual comparison to realize the automation of cloud cluster tracking, thereby effectively solving the matching problem caused by cloud cluster splitting and merging, and thus significantly improving the accuracy and reliability of cloud cluster tracking results.

[0020] Thirdly, this application provides an electronic device including a memory and a processor, the memory storing a computer program, and the processor running the computer program to cause the electronic device to perform the matching and tracking method for a contemporary satellite cloud as described in the first aspect.

[0021] Fourthly, this application provides a readable storage medium storing a computer program, which, when executed by a processor, performs the matching and tracking method for a contemporary satellite cloud cluster as described in any one of the first aspects.

[0022] Fifthly, this application provides a computer program product comprising a computer program that, when executed by a processor, performs the matching and tracking method for a contemporary satellite cloud cluster as described in the first aspect.

[0023] The beneficial effects of this application are as follows: By organically integrating the area overlap method and the individual feature similarity matching method, this application endows the algorithm with adaptive matching capabilities for cloud images at different time intervals. This enables the algorithm to efficiently track short-distance moving cloud clusters and accurately identify long-distance moving cloud clusters. Simultaneously, the algorithm can reasonably handle complex morphological changes such as cloud cluster merging and splitting, and effectively lock onto the main individuals of the cloud cluster, avoiding deviation from the tracking line, thereby significantly improving the overall accuracy and robustness of cloud cluster matching and tracking in satellite cloud images. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating a matching and tracking method for a phase-temporal satellite cloud cluster in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a matching and tracking device for a phase-temporal satellite cloud in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0026] The embodiments of the technical solution of this application will now be described in detail with reference to the accompanying drawings. These embodiments are only used to more clearly illustrate the technical solution of this application and are therefore merely examples, and should not be used to limit the scope of protection of this application.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0028] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more (including two), similarly, "multiple sets" refers to two or more sets (including two sets), and "multiple pieces" refers to two or more pieces (including two pieces) unless otherwise explicitly defined.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0031] Currently, forecasters mainly rely on manual analysis of geostationary satellite cloud images from adjacent time periods to identify cloud clusters by visually comparing their position, shape, and intensity characteristics. However, this manual method is highly subjective and inefficient, making it difficult to accurately capture the complex evolutionary processes of cloud clusters, such as their formation, development, merging, splitting, and dissipation. It also fails to provide automated, high-precision data support for statistical analysis of cloud origins, life cycles, and movement paths, thus failing to meet operational application needs.

[0032] To address the aforementioned technical problems, this application provides a matching and tracking method for satellite clouds in a given time period. This method first uses the area overlap method to quickly match the main cloud cluster, and then uses the individual feature similarity matching method to supplement the matching of unmatched cloud clusters, forming a complete matching result of "matching the main body first, and then supplementing the details". Finally, by statistically analyzing the matching relationships and calculating the area parameters, the method specifically identifies and optimizes the splitting and merging phenomena of cloud clusters, thereby achieving automated and high-precision tracking of cloud clusters.

[0033] like Figure 1 As shown, some embodiments of this application provide a matching and tracking method for phase-temporal satellite clouds, which includes: S100. Obtain the first observation cloud map and the first brightness temperature background map at the current moment, and the second observation cloud map and the second brightness temperature background map at the next moment; wherein, the first observation cloud map and the first brightness temperature background map correspond spatially, and the second observation cloud map and the second brightness temperature background map correspond spatially.

[0034] In this embodiment, the first observation cloud map and the first brightness temperature base map are strictly geometrically registered and have the same geographic projection method, spatial resolution and image size; at the same time, the second observation cloud map and the second brightness temperature base map are also identical.

[0035] In this embodiment, the geographical locations (latitude and longitude) of the Earth's surface represented by the same pixel coordinates (row, column) in the first observation cloud map and the first brightness temperature base map are exactly the same; at the same time, the geographical coverage of the first observation cloud map and the first brightness temperature base map completely overlap, enabling point-to-point comparison of the observed brightness temperature with the reference brightness temperature of the same area. Similarly, the second observation cloud map and the second brightness temperature base map are also like this.

[0036] For example, the first and second observation cloud images can be data observed by the long-wave infrared channel (10.3~11.3μm) of the domestic geostationary meteorological satellite FY-4A. The observations in both the first and second observation cloud images are brightness temperature.

[0037] In this embodiment, the observation time interval of the long-wave infrared channel is 15 minutes, and it supports continuous observation day and night. Therefore, this method can simultaneously acquire observation results during the day and at night.

[0038] In this embodiment, the brightness temperature base map is a benchmark reference map that is spatially matched with the observed cloud map. It is used to unify the benchmark for cloud cluster identification and provide a reference for subsequent cloud detection and region identification.

[0039] S200. Based on the first brightness temperature background map, cloud cluster regions are identified in the first observed cloud map to obtain the current set of cloud cluster regions. Based on the second brightness temperature background map, cloud cluster regions are identified in the second observed cloud map to obtain the next set of cloud cluster regions.

[0040] In this embodiment, cloud region identification is based on cloud detection results based on brightness temperature base map. It can extract independent cloud regions from the observed cloud map through binarization processing and connected component identification, thereby achieving clear definition and effective differentiation of cloud boundaries.

[0041] As an optional implementation, after obtaining the set of cloud regions at the current time and the set of cloud regions at the next time, the method further includes a preprocessing step for the cloud regions: Small-scale cloud clusters were removed.

[0042] For example, the method can set an area threshold (e.g., less than 100 pixels) and remove tiny connected regions with an area smaller than the threshold from the current cloud region set and the next cloud region set, retaining only the core area of ​​the cloud, so as to eliminate irrelevant interference such as broken clouds and noise, thereby reducing the amount of subsequent matching calculations and improving matching accuracy.

[0043] S300. Based on the first observed cloud map, extract the first set of feature parameters corresponding to the cloud cluster region set at the current time, and based on the second observed cloud map, extract the second set of feature parameters corresponding to the cloud cluster region set at the next time.

[0044] In this embodiment, both the first and second feature parameter sets include parameters such as cloud area, center position, eccentricity, equivalent diameter, actual volume, orientation, average brightness temperature, and centroid position. These parameters are used to provide quantitative basis for subsequent cloud matching.

[0045] S400. Based on the first set of feature parameters and the second set of feature parameters, perform cloud region matching processing on the current cloud region set and the next cloud region set to obtain the target matching cloud pair set.

[0046] In this embodiment, cloud region matching can adopt a hierarchical strategy of "area overlap method + individual feature similarity matching method". First, the main cloud cluster is quickly matched by the area overlap method. Then, the cloud clusters that were not matched are supplemented by the individual feature similarity matching method. Finally, the splitting and merging phenomena are integrated and optimized to ensure that the matching results are comprehensive and accurate.

[0047] S500: Generate cloud tracking results based on the target matching cloud pair set.

[0048] In this embodiment, the cloud tracking results can intuitively present the correspondence between cloud clusters at the current moment and the next moment. This correspondence includes successfully matched cloud cluster pairs, tracking links of split cloud clusters, and source associations of merged cloud clusters.

[0049] In this embodiment, cloud tracking results can provide data support for cloud life history analysis and movement path statistics.

[0050] In the above embodiments, the method can unify the cloud identification benchmark between the current and next time moment through the brightness temperature background map, ensuring the accuracy and spatial consistency of cloud region extraction; at the same time, quantitative matching based on feature parameter set can replace manual comparison to realize the automation of cloud tracking, thereby effectively solving the matching problem caused by cloud splitting and merging, and thus significantly improving the accuracy and reliability of cloud tracking results.

[0051] In some embodiments, step S200 may include: S210. Perform cloud detection processing on the first observed cloud map based on the first brightness temperature base map to obtain the first cloud detection result, and perform binarization processing on the first cloud detection result to obtain the first binarized result map.

[0052] In this embodiment, cloud detection can employ the brightness temperature background map method. Cloud identification is achieved by calculating the difference between the brightness temperature value of each pixel in the first observed cloud image and the brightness temperature value of the corresponding pixel in the first brightness temperature background map. Specifically, when the difference is greater than a preset threshold (which can be determined based on historical data statistics), the pixel is determined to be a cloud; otherwise, it is determined to be clear sky.

[0053] In this embodiment, the binarization process can be based on the cloud identification results described above, marking cloud pixels as value 1 and clear sky pixels as value 0. At the same time, by retaining these 0 / 1 identifiers and eliminating redundant information, a first binarized result image that distinguishes only clouds and clear sky is finally formed.

[0054] S220. Perform cloud detection processing on the second observed cloud map based on the second brightness temperature base map to obtain the second cloud detection result, and perform binarization processing on the second cloud detection result to obtain the second binarized result map.

[0055] In this embodiment, the logic of cloud detection and binarization processing is consistent with step S210, that is, cloud pixels are determined by brightness temperature difference, and a second binarized result map is generated with 0 / 1 to identify clouds and clear sky, so as to ensure that the processing standard is consistent with the first observed cloud map.

[0056] S230. Perform connected component identification processing on the first binarized result map to obtain the set of cloud regions at the current time.

[0057] In this embodiment, the method can use the 8-connectivity algorithm to perform connected component identification processing on the first binarized result graph.

[0058] In this embodiment, the 8-connectivity algorithm can mark adjacent cloud pixels (value 1) in the first binarized result image (including horizontal, vertical, and diagonal directions) as the same connected region. Each connected region is an independent cloud cluster, and all independent cloud clusters together constitute the cloud cluster region set at the current time.

[0059] S240. Perform connected component identification processing on the second binarized result map to obtain the set of cloud regions at the next time step.

[0060] In this embodiment, the method can use the same 8-connectivity algorithm as step S230 to identify connected regions in the second binarized result map, and mark adjacent cloud pixels as independent cloud clusters to form a set of cloud cluster regions at the next time step, thereby ensuring the consistency of cloud cluster identification logic at previous and subsequent time steps.

[0061] In the above embodiments, the method can perform accurate cloud detection and binarization segmentation based on the brightness temperature base map, and extract independent cloud cluster regions through connected component identification, effectively eliminating non-cloud interference and ensuring the integrity and independence of the cloud cluster region set at the current and next time moments.

[0062] In some embodiments, step S400 may include: S410. Based on the first set of feature parameters and the second set of feature parameters, the area overlap method is used to perform cloud region matching processing on the current cloud region set and the next cloud region set to obtain a preliminary set of matched cloud pairs.

[0063] In this embodiment, the method can calculate the area overlap rate of cloud clusters at the current time and the next time by using the area overlap method, thereby filtering out cloud cluster pairs with strong spatial correlation, and using the filtered cloud cluster pairs as preliminary matching results.

[0064] In this embodiment, the area overlap method described above can be applied to cloud matching scenarios with small movement distances.

[0065] To ensure that the preliminary matching results have a high degree of confidence, step S410 may also include: S411. Calculate the multiple first areas in the current cloud region set that correspond one-to-one with the multiple first cloud regions based on the first feature parameter set.

[0066] In this embodiment, "one-to-one correspondence" means that each independent cloud cluster identified at the current moment is individually calculated and assigned a unique area value. That is, a first cloud cluster region corresponds to a unique first area, thus forming a one-to-one mapping relationship.

[0067] In this embodiment, the first area is the total number of pixels corresponding to each first cloud region at the current moment. The first area can be directly obtained from the area parameters extracted from the first feature parameter set.

[0068] In this embodiment, the first area is used to reflect the spatial scale of the cloud.

[0069] S412. Calculate the multiple second areas in the cloud region set at the next moment that correspond one-to-one with the multiple second cloud regions based on the second feature parameter set.

[0070] In this embodiment, "one-to-one correspondence" means that each independent second cloud region identified in the next moment has its area value calculated separately based on the second feature parameter set. That is, one second cloud region corresponds to one second area, thus forming a one-to-one mapping relationship.

[0071] In this embodiment, the second area is the total number of pixels corresponding to each second cloud region at the next time step. The second area can be directly obtained from the area parameters extracted from the second feature parameter set.

[0072] In this embodiment, the calculation standard for the second area is consistent with that for the first area.

[0073] S413. Combine multiple first cloud cluster regions in the current cloud cluster region set with multiple second cloud cluster regions in the next cloud cluster region set in pairs to obtain multiple cloud cluster region combination pairs.

[0074] In this embodiment, pairwise combination refers to pairing each first cloud cluster region at the current time with each second cloud cluster region at the next time time to construct a full set of candidate matches.

[0075] S414. Based on the first set of feature parameters and the second set of feature parameters, calculate multiple area overlap rates that correspond one-to-one with multiple cloud region combination pairs; wherein, the preliminary matching set of cloud region pairs includes all cloud region combination pairs whose area overlap rate exceeds a preset overlap rate threshold.

[0076] In this embodiment, "one-to-one correspondence" refers to pairing each cloud cluster at the current moment with each cloud cluster at the next moment. For each pair of clouds generated, the area overlap rate of that pair is calculated. That is, one cloud cluster region pair corresponds to one area overlap rate, thus forming a one-to-one mapping relationship.

[0077] In this embodiment, the area overlap rate can be obtained by calculating the ratio of the area of ​​the overlapping part between the first cloud region and the second cloud region in each cloud region combination pair to the total coverage area of ​​the two.

[0078] In this embodiment, the area overlap rate is used to quantify the degree of spatial overlap of cloud clusters.

[0079] In this embodiment, the preset overlap rate threshold is set to 0.3 (30%), and its value range is 0~1 (100%).

[0080] In this embodiment, if the area overlap rate of a cloud region pair is greater than or equal to the threshold, the pair is considered a successful match. All successfully matched pairs constitute a preliminary set of matched cloud pairs.

[0081] S420. Based on the preliminary set of matched cloud cluster pairs, determine the first set of unmatched cloud cluster regions remaining in the current cloud cluster region set and the second set of unmatched cloud cluster regions remaining in the next cloud cluster region set.

[0082] In this embodiment, the method can remove the first cloud cluster region that has been included in the preliminary matching cloud cluster pair set from the current cloud cluster region set, so that the remaining part constitutes the first remaining cloud cluster region set.

[0083] In this embodiment, the method can remove the matched second cloud region from the set of cloud regions at the next moment, or the remaining part can form a second set of remaining cloud regions. These types of clouds are mostly those that have moved a large distance or have undergone significant morphological changes.

[0084] S430. Perform matching processing based on individual feature similarity on the first set of remaining cloud regions and the second set of remaining cloud regions to obtain a set of supplementary matching cloud pairs.

[0085] In this embodiment, the method can use individual feature similarity matching to calculate similarity based on the core feature parameters of the cloud cluster, and perform accurate matching on the first and second remaining cloud cluster regions, thereby making up for the shortcomings of the area overlap method in matching long-distance moving clouds.

[0086] To effectively identify homologous cloud clusters with insufficient spatial overlap due to splitting, merging, or rapid deformation, step S430 may further include: S431. Obtain multiple first feature matching factors in the first remaining cloud region set that correspond one-to-one with multiple first remaining cloud regions based on the first feature parameter set.

[0087] In this embodiment, "one-to-one correspondence" means that for each remaining cloud cluster in the first set of remaining cloud cluster regions, a set of exclusive feature matching factors (such as area, center position, and average brightness temperature) are extracted from the first set of feature parameters. That is, one remaining cloud cluster region corresponds to one first feature matching factor, thus forming a one-to-one mapping relationship.

[0088] In this embodiment, the first feature matching factor selects core parameters such as the area, center position, and average brightness temperature of the first remaining cloud region. These parameters can stably reflect the individual attributes of the cloud and have strong resistance to morphological change interference.

[0089] S432. Calculate multiple second feature matching factors in the second remaining cloud region set that correspond one-to-one with multiple second remaining cloud regions based on the second feature parameter set.

[0090] In this embodiment, "one-to-one correspondence" means that, following the same rules as for the first remaining cloud cluster, a unique set of second feature matching factors is calculated and extracted from the second feature parameter set for each cloud cluster in the second remaining cloud cluster region set. That is, one second remaining cloud cluster region corresponds to one second feature matching factor, thus forming a one-to-one mapping relationship.

[0091] In this embodiment, the second feature matching factor is selected in the same way as the first feature matching factor, which are parameters such as area, center position, and average brightness temperature, to ensure the comparability of matching factors at different times.

[0092] S433. Based on multiple first feature matching factors and multiple second feature matching factors, the individual feature similarity matching method is used to match the first set of remaining cloud regions with the second set of remaining cloud regions to obtain a set of supplementary matching cloud pairs.

[0093] In this embodiment, the method can calculate the similarity distance between the first feature matching factor and the second feature matching factor using the individual feature similarity matching method.

[0094] In this embodiment, when calculating the similarity distance using the individual feature similarity matching method, the area, center position, and average brightness temperature in the first and second feature matching factors can be normalized to eliminate the influence of different dimensions. Then, the weighted Euclidean distance formula is used to calculate the distance between the first remaining cloud cluster at the current time and the second remaining cloud cluster at the next time in the multidimensional feature space (where the center position has the highest weight, followed by the area and average brightness temperature; the smaller the distance value, the higher the similarity of the individual features of the two cloud clusters).

[0095] In this embodiment, the preset maximum similarity threshold is 80 (the smaller the value, the more similar the matched objects). Based on this preset threshold, if the calculated similarity distance is less than 80, it is determined that the match is successful, and a supplementary matching cloud pair set is formed.

[0096] S440. The total set of cloud cluster pairs, consisting of the initial set of matching cloud cluster pairs and the supplementary set of matching cloud cluster pairs, is integrated and optimized to obtain the target set of matching cloud cluster pairs.

[0097] In this embodiment, the integration optimization process can quantify the splitting and merging phenomena in the total cloud cluster set by using area parameters, thereby resolving matching conflicts, optimizing pairing relationships, and ensuring the continuity of the cloud cluster tracking link.

[0098] To correct pairing deviations caused by complex morphological evolution, step S440 may further include: S441. Perform statistical analysis on the total cloud cluster set to obtain the statistical analysis results.

[0099] In this embodiment, the statistical analysis mainly focuses on the number of times the cloud cluster region is matched at each current moment in the total cloud cluster set, and the number of times the cloud cluster region is matched at each next moment.

[0100] S442. Based on the statistical analysis results, select the split cloud regions that exhibit splitting and the merged cloud regions that exhibit merging from the multiple current-time cloud regions in the total cloud cluster set.

[0101] In this embodiment, the method can make a judgment by using the quantization threshold of the number of matching and the number of times matched, thereby accurately locating split and merged cloud clusters.

[0102] As an optional implementation, based on statistical analysis results, split cloud regions exhibiting splitting phenomena and merged cloud regions exhibiting merging phenomena are selected from multiple current-time cloud regions in the total cloud cluster set, including: Based on the statistical analysis results, obtain the number of times the total cloud cluster pairs correspond one-to-one with the cloud cluster regions at multiple current times, and the number of times the total cloud cluster pairs correspond one-to-one with the cloud cluster regions at multiple next times. The cloud regions in the total set of cloud pairs whose number of matches is greater than the threshold of the first match are identified as split cloud regions with splitting phenomena. The cloud region in the next moment whose number of matching times in the total cloud pair set is greater than the second threshold is determined as the target cloud region; The cloud cluster regions in the total cloud cluster set that match the target cloud cluster region at the current moment are identified as merged cloud cluster regions that exhibit merging behavior.

[0103] In this embodiment, the first "one-to-one correspondence" refers to: for each cloud cluster in the set at the current moment, counting and recording the number of cloud clusters at the next moment that it successfully matches (i.e., the number of matches). That is, one current moment cloud cluster region corresponds to one number of matches.

[0104] In this embodiment, the second "one-to-one correspondence" refers to: for each cloud cluster in the next time step, counting and recording how many current time step cloud clusters matched it (i.e., the number of times it was matched). That is, one cloud cluster region in the next time step corresponds to one number of matches.

[0105] In this embodiment, both the first and second number thresholds can be set to 1. That is, when the number of matches in the cloud region at the current moment is ≥2, it is determined to be a split cloud region; when the number of matches in the cloud region at the next moment is ≥2, the corresponding cloud region at the current moment is determined to be a merged cloud region.

[0106] By implementing this method, an objective basis for judging splitting and merging phenomena can be established through quantitative statistics of the number of matching and the number of times matched, thereby achieving automated and standardized identification of splitting and merging clouds and avoiding misjudgments caused by experience.

[0107] S443. Obtain the next-time cloud cluster region paired with the split cloud cluster region and the next-time paired cloud cluster region paired with the merged cloud cluster region based on the total cloud cluster pair set.

[0108] In this embodiment, the method can extract all next-time cloud regions that have a pairing relationship with the split cloud region from the total cloud region set, as well as the next-time cloud regions that have a pairing relationship with the merged cloud region (i.e., the next-time paired cloud regions), thereby clarifying the associated objects of split and merged clouds.

[0109] S444. Based on the first set of feature parameters, the second set of feature parameters, and the cloud region at the next moment paired with the split cloud region, calculate the first area parameter corresponding to the split cloud region.

[0110] In this embodiment, the first area parameter is the ratio of the area of ​​the split cloud region to the area of ​​the cloud region at the next time step (the ratio of the cloud area at time t to the tracked cloud area at time t+1).

[0111] In this embodiment, the first area parameter is used to quantify the degree to which the split cloud inherits the characteristics of the original cloud.

[0112] S445. Based on the first set of feature parameters, the second set of feature parameters, and the next time pairing cloud region with the merged cloud region, calculate the second area parameter corresponding to the merged cloud region.

[0113] In this embodiment, the second area parameter is the ratio of the area of ​​the merged cloud region to the area of ​​the paired cloud region at the next time step (the ratio of the cloud area at time t to the tracked cloud area at time t+1).

[0114] In this embodiment, the second area parameter is used to quantify the contribution of the cloud cluster before merging to the target cloud cluster.

[0115] S446. Optimize the pairing relationship of split cloud regions in the total cloud pair set according to the first area parameter to obtain the optimized cloud pair set.

[0116] In this embodiment, when a split cloud region corresponds to multiple cloud regions in the next time step, if there exists one and only one cloud region with a first area parameter P < 2, then the pairing relationship is retained, and tracking of other related cloud regions is stopped. At this point, the method can determine that the cloud region in the next time step inherits the main characteristics of the split cloud and completes the pairing optimization of the split cloud.

[0117] In this embodiment, the first area parameter P < 2, which means that the area of ​​the cloud at time t+1 is at least 1 / 2 of the area of ​​the cloud at time t.

[0118] S447. Based on the second area parameter, optimize the pairing relationship of the merged cloud region in the optimized cloud pair set to obtain the target matching cloud pair set.

[0119] In this embodiment, when multiple merged cloud cluster regions correspond to the same paired cloud cluster region in the next time step, if there exists one and only one cloud cluster region with a second area parameter P > 0.5, the pairing relationship is retained, and tracking of other related cloud clusters is stopped. At this time, the method can determine that the merged cloud cluster region is the main source of the target cloud cluster, complete the merged cloud cluster pairing optimization, and finally form a set of target matching cloud cluster pairs.

[0120] In this embodiment, the second area parameter P > 0.5, which means that the area of ​​the cloud at time t accounts for more than half (>50%) of the area of ​​the cloud at time t+1.

[0121] In the above embodiments, the method can effectively solve the matching ambiguity problem caused by cloud splitting, merging and edge blurring through hierarchical matching and integration optimization strategy. In this way, while ensuring the high efficiency of matching the main cloud, it can achieve accurate matching of weakly related cloud, thereby significantly reducing the omission and mismatch rate, and finally obtaining cloud matching results with both integrity and consistency.

[0122] This embodiment verifies the effectiveness of the method through specific experiments, the details of which are as follows: First, the experiment set the time interval between the current time (time t) and the next time (time t+1) of the cloud image to be 1 hour. Specifically, the FY-4A long-wave infrared channel satellite cloud images at 4:00 AM (UTC, time t) and 5:00 AM (time t+1) on June 8, 2023 were selected as the experimental data.

[0123] Secondly, the two sets of cloud images were preprocessed to remove small-scale cloud clusters and retain only the core area of ​​the cloud clusters in order to eliminate irrelevant interference.

[0124] Then, using the method of this application, cloud cluster matching is carried out through the area overlap method and the individual feature similarity matching method; at the same time, a comparative experiment is set up to compare the effect of using three types of features, namely “area, center position, and average brightness temperature” and multiple types of features, namely “area, center position, eccentricity, equivalent diameter, realization, direction, average brightness temperature, and centroid position”, as matching factors.

[0125] The experimental results show that a total of 22 cloud clusters were identified in the cloud map at time t, and 19 of them were successfully matched, with an overall matching success rate of 86%. Among them, 12 were matched by the area overlap method and 7 were matched by the individual feature similarity method. The latter effectively solved the problem of matching cloud clusters with large positional movements.

[0126] Furthermore, comparative experiments also show that selecting area, center position, and average brightness temperature as three features as matching factors is basically equivalent to the matching results of multi-feature combinations.

[0127] like Figure 2 As shown, some embodiments of this application provide a schematic diagram of a matching and tracking device for phase-temporal satellite clouds. It should be understood that this device is related to... Figure 1 The method executed in the middle corresponds to the steps involved in the aforementioned method. The specific functions and effects of the device can be found in the description above. To avoid repetition, detailed descriptions are omitted here.

[0128] The matching and tracking device for phase-temporal satellite cloud clusters includes: The acquisition unit 610 is used to acquire the first observation cloud map and the first brightness temperature background map at the current time, and the second observation cloud map and the second brightness temperature background map at the next time; wherein the first observation cloud map and the first brightness temperature background map correspond to each other in space, and the second observation cloud map and the second brightness temperature background map correspond to each other in space. The identification unit 620 is used to identify cloud regions in the first observed cloud map based on the first brightness temperature background map to obtain the current set of cloud regions, and to identify cloud regions in the second observed cloud map based on the second brightness temperature background map to obtain the next set of cloud regions. The feature extraction unit 630 is used to extract the first set of feature parameters corresponding to the set of cloud regions at the current time based on the first observed cloud map, and to extract the second set of feature parameters corresponding to the set of cloud regions at the next time based on the second observed cloud map. The matching unit 640 is used to perform cloud region matching processing on the current cloud region set and the next cloud region set according to the first feature parameter set and the second feature parameter set to obtain a target matching cloud pair set. The generation unit 650 is used to generate cloud tracking results based on the target matching cloud pair set.

[0129] In some embodiments, the identification unit 620 includes: The cloud detection subunit 621 is used to perform cloud detection processing on the first observed cloud map based on the first brightness temperature background map to obtain the first cloud detection result, and to perform binarization processing on the first cloud detection result to obtain the first binarized result map. Binarization subunit 622 is used to perform cloud detection processing on the second observed cloud map based on the second brightness temperature base map, obtain the second cloud detection result, and perform binarization processing on the second cloud detection result to obtain the second binarized result map; The identification subunit 623 is used to perform connected component identification processing on the first binarized result map to obtain the set of cloud regions at the current time. The identification subunit 624 is also used to perform connected region identification processing on the second binarized result map to obtain the cloud region set at the next time step.

[0130] In some embodiments, the matching unit 640 includes: The matching subunit 641 is used to perform cloud region matching processing on the current cloud region set and the next cloud region set according to the first feature parameter set and the second feature parameter set, using the area overlap method, to obtain a preliminary set of matched cloud pairs. The subunit 642 is used to determine, based on the preliminary set of matched cloud pairs, the first set of unmatched cloud regions remaining in the current set of cloud regions and the second set of unmatched cloud regions remaining in the next set of cloud regions; The matching subunit 641 is also used to perform matching processing based on individual feature similarity on the first set of remaining cloud regions and the second set of remaining cloud regions to obtain a supplementary set of matching cloud pairs. The optimization subunit 643 is used to integrate and optimize the total set of cloud cluster pairs consisting of the initial matching cloud cluster pair set and the supplementary matching cloud cluster pair set to obtain the target matching cloud cluster pair set.

[0131] In some embodiments, the matching subunit 641 includes: The first calculation module is used to calculate, based on the first set of feature parameters, multiple first areas that correspond one-to-one with multiple first cloud regions in the current cloud region set; The first calculation module is also used to calculate, based on the second feature parameter set, multiple second areas in the cloud region set at the next moment that correspond one-to-one with multiple second cloud regions; The combination module is used to combine multiple first cloud cluster regions in the current cloud cluster region set with multiple second cloud cluster regions in the next cloud cluster region set in pairs to obtain multiple cloud cluster region combination pairs. The first calculation module is also used to calculate multiple area overlap rates corresponding one-to-one with multiple cloud region combination pairs based on the first feature parameter set and the second feature parameter set; wherein, the preliminary matching cloud pair set includes all cloud region combination pairs whose area overlap rate exceeds a preset overlap rate threshold.

[0132] In some embodiments, the matching subunit 641 includes: The first acquisition module is used to acquire multiple first feature matching factors in the first set of remaining cloud regions that correspond one-to-one with multiple first remaining cloud regions, based on the first set of feature parameters. The first calculation module is also used to calculate, based on the second feature parameter set, multiple second feature matching factors that correspond one-to-one with multiple second remaining cloud regions in the second remaining cloud region set; The matching module is used to match the first set of remaining cloud regions with the second set of remaining cloud regions using an individual feature similarity matching method based on multiple first feature matching factors and multiple second feature matching factors, so as to obtain a set of supplementary matching cloud pairs.

[0133] In some embodiments, the optimization subunit 643 includes: The statistics module is used to perform statistical analysis on the total set of cloud cluster pairs, which consists of the initial set of matched cloud cluster pairs and the supplementary set of matched cloud cluster pairs, and to obtain the statistical analysis results. The filtering module is used to filter out split cloud regions with splitting phenomena and merged cloud regions with merging phenomena from multiple current-time cloud regions in the total cloud cluster set based on statistical analysis results. The second acquisition module is used to acquire the next-time cloud cluster region paired with the split cloud cluster region and the next-time paired cloud cluster region paired with the merged cloud cluster region based on the total cloud cluster pair set. The second calculation module is used to calculate the first area parameter corresponding to the split cloud region based on the first feature parameter set, the second feature parameter set, and the next time-step cloud region paired with the split cloud region. The second calculation module is also used to calculate the second area parameter corresponding to the merged cloud region based on the first feature parameter set, the second feature parameter set, and the next time pairing cloud region with the merged cloud region; The optimization module is used to optimize the pairing relationship of split cloud regions in the total cloud pair set according to the first area parameter, so as to obtain the optimized cloud pair set. The optimization module is also used to optimize the pairing relationship of the merged cloud regions in the optimized cloud pair set according to the second area parameter, so as to obtain the target matching cloud pair set.

[0134] In some embodiments, the filtering module is specifically used to obtain, based on statistical analysis results, multiple matching counts corresponding one-to-one with multiple current-time cloud cluster regions in the total cloud cluster pair set, and multiple matching counts corresponding one-to-one with multiple next-time cloud cluster regions in the total cloud cluster pair set; determine the current-time cloud cluster regions in the total cloud cluster pair set whose matching count is greater than the first threshold as split cloud cluster regions with splitting phenomena; determine the next-time cloud cluster regions in the total cloud cluster pair set whose matching count is greater than the second threshold as target cloud cluster regions; and determine the current-time cloud cluster regions in the total cloud cluster pair set that match the target cloud cluster regions as merged cloud cluster regions with merging phenomena.

[0135] like Figure 3 As shown, this application provides an electronic device 700, which includes a processor 701 and a memory 702. The processor 701 and the memory 702 are interconnected and communicate with each other through a communication bus 703 and / or other forms of connection mechanism (not shown). The memory 702 stores a computer program that can be executed by the processor 701. When the computing device is running, the processor 701 executes the computer program to perform the method in any of the aforementioned optional implementations.

[0136] This application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the method in any of the aforementioned optional implementations.

[0137] The computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0138] This application provides a computer program product, which includes a computer program that, when run by a processor, executes the method in any of the aforementioned optional implementations.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and they should all be covered within the scope of the claims and specification of this application. In particular, as long as there is no conflict, the various technical features mentioned in the embodiments can be combined in any way. This application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A matching and tracking method for phase-temporal satellite cloud clusters, characterized in that, include: Obtain the first observation cloud image and the first brightness temperature background image at the current moment, and the second observation cloud image and the second brightness temperature background image at the next moment; wherein, the first observation cloud image and the first brightness temperature background image correspond spatially, and the second observation cloud image and the second brightness temperature background image correspond spatially. Based on the first brightness temperature background map, cloud cluster regions are identified in the first observed cloud map to obtain the current set of cloud cluster regions. Based on the second brightness temperature background map, cloud cluster regions are identified in the second observed cloud map to obtain the next set of cloud cluster regions. Based on the first observed cloud map, extract the first set of feature parameters corresponding to the cloud cluster region set at the current time, and based on the second observed cloud map, extract the second set of feature parameters corresponding to the cloud cluster region set at the next time. Based on the first set of feature parameters and the second set of feature parameters, cloud cluster region matching processing is performed on the current cloud cluster region set and the next cloud cluster region set to obtain a target matching cloud cluster pair set. Cloud cluster tracking results are generated based on the target matching cloud cluster pair set.

2. The matching and tracking method for phase-temporal satellite clouds according to claim 1, characterized in that, The step of identifying cloud regions in the first observed cloud map based on the first brightness temperature background map to obtain the current set of cloud regions, and identifying cloud regions in the second observed cloud map based on the second brightness temperature background map to obtain the next set of cloud regions, includes: Based on the first brightness temperature background map, cloud detection processing is performed on the first observed cloud map to obtain the first cloud detection result, and the first cloud detection result is binarized to obtain the first binarized result map; Based on the second brightness temperature base map, cloud detection processing is performed on the second observed cloud map to obtain the second cloud detection result, and the second cloud detection result is binarized to obtain the second binarized result map; Perform connected component identification processing on the first binarized result image to obtain the set of cloud regions at the current time; The second binarized result image is processed for connected component identification to obtain the set of cloud regions at the next time step.

3. The matching and tracking method for phase-temporal satellite clouds according to claim 1, characterized in that, The step of performing cloud region matching processing on the current cloud region set and the next cloud region set based on the first feature parameter set and the second feature parameter set to obtain a target matching cloud pair set includes: Based on the first set of feature parameters and the second set of feature parameters, the area overlap method is used to perform cloud region matching processing on the current cloud region set and the next cloud region set to obtain a preliminary set of matched cloud pairs. Based on the preliminary matched cloud cluster pair set, determine the first set of unmatched cloud cluster regions remaining in the current cloud cluster region set and the second set of unmatched cloud cluster regions remaining in the next cloud cluster region set; The first set of remaining cloud regions and the second set of remaining cloud regions are matched based on individual feature similarity to obtain a set of supplementary matching cloud pairs. The total set of cloud cluster pairs, consisting of the initial set of matched cloud cluster pairs and the supplementary set of matched cloud cluster pairs, is integrated and optimized to obtain the target set of matched cloud cluster pairs.

4. The matching and tracking method for phase-temporal satellite clouds according to claim 3, characterized in that, The step involves matching the current cloud region set with the next cloud region set using the area overlap method based on the first feature parameter set and the second feature parameter set, to obtain a preliminary set of matched cloud pairs, including: Calculate multiple first areas in the current cloud region set that correspond one-to-one with multiple first cloud regions based on the first set of feature parameters; Calculate the multiple second areas in the cloud region set at the next moment, which correspond one-to-one with the multiple second cloud regions, based on the second set of feature parameters; Multiple first cloud cluster regions in the current cloud cluster region set are combined with multiple second cloud cluster regions in the next cloud cluster region set in pairs to obtain multiple cloud cluster region combination pairs; Based on the first set of feature parameters and the second set of feature parameters, calculate multiple area overlap rates that correspond one-to-one with the multiple cloud region combination pairs; wherein, the preliminary matching cloud pair set includes all cloud region combination pairs whose area overlap rate exceeds a preset overlap rate threshold.

5. The matching and tracking method for phase-temporal satellite clouds according to claim 3, characterized in that, The step of performing matching processing based on individual feature similarity on the first set of remaining cloud regions and the second set of remaining cloud regions to obtain a supplementary set of matching cloud pairs includes: Based on the first set of feature parameters, obtain multiple first feature matching factors in the first set of remaining cloud regions that correspond one-to-one with multiple first remaining cloud regions; Calculate multiple second feature matching factors in the second set of remaining cloud regions that correspond one-to-one with multiple second remaining cloud regions based on the second set of feature parameters; Based on the plurality of first feature matching factors and the plurality of second feature matching factors, the individual feature similarity matching method is used to match the first set of remaining cloud regions with the second set of remaining cloud regions to obtain a set of supplementary matching cloud pairs.

6. The matching and tracking method for phase-temporal satellite clouds according to claim 3, characterized in that, The process of integrating and optimizing the total set of cloud cluster pairs, which consists of the initial set of matched cloud cluster pairs and the supplementary set of matched cloud cluster pairs, to obtain the target set of matched cloud cluster pairs includes: Statistical analysis is performed on the total set of cloud cluster pairs consisting of the preliminary matching cloud cluster pair set and the supplementary matching cloud cluster pair set to obtain the statistical analysis results. Based on the statistical analysis results, split cloud cluster regions exhibiting splitting phenomena and merged cloud cluster regions exhibiting merging phenomena are selected from multiple current-time cloud clusters in the total cloud cluster pair set. The next-time cloud cluster region paired with the split cloud cluster region and the next-time paired cloud cluster region paired with the merged cloud cluster region are obtained based on the total cloud cluster pair set. Based on the first set of feature parameters, the second set of feature parameters, and the cloud region at the next moment paired with the split cloud region, calculate the first area parameter corresponding to the split cloud region; Based on the first set of feature parameters, the second set of feature parameters, and the next time pairing cloud region with the merged cloud region, calculate the second area parameter corresponding to the merged cloud region; The pairing relationship of the split cloud region in the total cloud pair set is optimized according to the first area parameter to obtain an optimized cloud pair set. The pairing relationship of the merged cloud region in the optimized cloud pair set is optimized based on the second area parameter to obtain the target matching cloud pair set.

7. The matching and tracking method for phase-temporal satellite clouds according to claim 6, characterized in that, The step of filtering split cloud regions exhibiting splitting and merged cloud regions exhibiting merging from multiple current-time cloud regions in the total cloud cluster set based on the statistical analysis results includes: Based on the statistical analysis results, obtain the number of times each of the total cloud cluster pairs corresponds to a cloud cluster region at a current time, and the number of times each of the total cloud cluster pairs corresponds to a cloud cluster region at a next time. The cloud regions in the total cloud pair set whose number of matches is greater than the threshold of the first match are identified as split cloud regions with splitting phenomena. The cloud region in the next moment in the total cloud region pair set that has been matched more than the second threshold number is determined as the target cloud region; The cloud cluster regions in the total cloud cluster pair set that match the target cloud cluster region at the current moment are identified as merged cloud cluster regions that exhibit merging phenomena.

8. A matching and tracking device for phase-temporal satellite cloud clusters, characterized in that, The matching and tracking device for phase-temporal satellite clouds includes: The acquisition unit is used to acquire the first observation cloud map and the first brightness temperature background map at the current time, and the second observation cloud map and the second brightness temperature background map at the next time; wherein, the first observation cloud map and the first brightness temperature background map correspond spatially, and the second observation cloud map and the second brightness temperature background map correspond spatially. The identification unit is used to identify cloud regions in the first observed cloud map based on the first brightness temperature background map to obtain the current set of cloud regions, and to identify cloud regions in the second observed cloud map based on the second brightness temperature background map to obtain the next set of cloud regions. The feature extraction unit is used to extract a first set of feature parameters corresponding to the set of cloud regions at the current time based on the first observed cloud map, and to extract a second set of feature parameters corresponding to the set of cloud regions at the next time based on the second observed cloud map. The matching unit is used to perform cloud region matching processing on the current cloud region set and the next cloud region set according to the first feature parameter set and the second feature parameter set to obtain a target matching cloud pair set; The generation unit is used to generate cloud tracking results based on the target matching cloud pair set.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor running the computer program to cause the electronic device to perform the matching and tracking method for a phase-specific satellite cloud as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that, The readable storage medium stores a computer program, which, when executed by a processor, performs the matching and tracking method for a phase-temporal satellite cloud cluster as described in any one of claims 1 to 7.

Images