Satellite data-based sand-dust monitoring analysis method and system

Abstract

This invention relates to the field of satellite remote sensing monitoring and image processing technology, specifically to a method and system for monitoring and analyzing dust storms based on satellite data. Addressing the problems of existing methods being susceptible to interference from surface background differences and single-time anomalies, and the unclear correlation and easily broken processes of dust storm regions across different time periods, this invention acquires continuous time-series satellite observation data and divides it into blocks, constructing similar surface reference groups; extracts dust storm characteristics from the current time-series analysis block to identify blocks to be confirmed; searches for connecting blocks that meet the conditions in subsequent time periods to determine the starting block; assigns event chain identifiers and transmits them to subsequent blocks, forming a dust storm event chain; then performs anomaly removal and intermittent continuation correction, outputting dust storm index products, dust storm visibility products, dust storm propagation process information, and regional statistical results. It can be used for regional environmental monitoring and dust storm process analysis.

Description

Technical Field

[0001] This invention relates to the field of satellite remote sensing monitoring and image processing technology, specifically to a method and system for monitoring and analyzing sand and dust based on satellite data. Background Technology

[0002] With the development of satellite remote sensing technology, Earth observation technology, and image data processing technology, atmospheric environmental monitoring using multispectral satellite observation data has become an important technological direction. Early dust monitoring mainly relied on ground-based station observations and manual interpretation, resulting in limited monitoring range and poor continuity. Subsequently, dust identification, thematic map generation, and process tracking based on satellite imagery have gradually developed. Dust can be identified and analyzed using image features such as visible light, infrared brightness temperature, and brightness-temperature difference, and this technology has been widely applied in regional environmental monitoring, disaster prevention and mitigation, and remote sensing operational processing.

[0003] Existing dust monitoring methods based on satellite image data mostly focus on single-time image recognition or direct threshold determination, which are easily affected by differences in surface type, high background reflectivity interference, and short-term abnormal fluctuations, resulting in a high misclassification rate of the area to be identified. At the same time, the correlation rules between dust areas at different times are unclear, which can easily lead to process breaks, misconnections, or omissions, affecting the accuracy of continuous tracking and statistical analysis results of dust events. It is difficult to balance the accuracy of image recognition and the completeness of process analysis. Therefore, there is a need for dust monitoring and analysis methods and systems based on satellite data to solve the above problems. Summary of the Invention

[0004] (a) Technical problem to be solved: In view of the shortcomings of the existing technology, the present invention provides a method and system for monitoring and analyzing sand and dust based on satellite data, which solves the above-mentioned problems.

[0005] (II) Technical Solution: To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring and analyzing sandstorms based on satellite data, comprising: S1. Obtain continuous satellite observation data of the target area, divide the target area into blocks, and construct a corresponding similar surface reference group for each analysis block; S2. Extract the dust characterization features of each analysis block in the current time period, and determine the blocks to be confirmed based on the deviation of the dust characterization features from the corresponding similar surface reference group; S3. In subsequent iterations, search for a receiving block that meets the receiving conditions of the block to be confirmed, and if the receiving block exists, determine the block to be confirmed as the starting block; S4. Assign an event chain identifier to the starting block and pass the event chain identifier to the blocks that meet the acceptance conditions in subsequent times to form a sandstorm event chain; S5. Perform anomaly removal and intermittent connection correction on the dust event chain to obtain dust monitoring and analysis results.

[0006] Furthermore, step S1 specifically includes the following steps: S11. Acquire multispectral satellite observation data of the target area over consecutive time intervals; S12. Divide the target area into grids according to fixed angular intervals in the longitude direction and fixed angular intervals in the latitude direction to obtain multiple analysis blocks; S13. For each analysis block, select reference blocks with the same or similar surface type and spatial proximity to the analysis block from satellite observation data from historical dust-free periods that are in the same period as the current analysis period, and form a corresponding similar surface reference group.

[0007] Furthermore, step S2 specifically includes the following steps: S21. Extract at least two of the visible light reflectance, infrared brightness temperature, and brightness temperature difference of each analysis block in the current time as characterization features of sand and dust. S22. Calculate the deviation of the dust characterization characteristics of each analysis block from the mean of the corresponding similar land surface reference group; S23. When the deviation of the same analysis block in at least two dust characterization features exceeds the standard deviation of the corresponding similar surface reference group, the analysis block shall be identified as a block to be confirmed.

[0008] Furthermore, the acceptance conditions are specifically as follows: In subsequent time intervals, the target block and the previous time interval block used as the comparison benchmark are in the next or second time interval. The previous time interval block used as the comparison benchmark is the block to be confirmed, or a block carrying an event chain identifier. The target block and the previous time interval block used as the comparison benchmark are spatially connected, or the distance between their center points is no greater than twice the diagonal length of the previous time interval block used as the comparison benchmark. Furthermore, the target block and the previous time interval block used as the comparison benchmark show the same direction of change in dust characterization features, and the deviation of the target block relative to its corresponding similar surface reference group is no less than half of the corresponding deviation of the previous time interval block used as the comparison benchmark.

[0009] Furthermore, step S4 specifically includes the following steps: S41. Assign a unique event chain identifier to the determined starting block, wherein the event chain identifier includes at least the event number and the starting time. S42. In subsequent iterations, search for analysis blocks that satisfy the acceptance conditions and carry the event chain identifier. S43. Associate the analysis blocks that meet the acceptance conditions with the corresponding event chain identifier to form a sandstorm event chain.

[0010] Furthermore, the following steps are included after S4: Based on the temporal position of each block in the dust storm event chain and the changing state of dust storm characteristics, the blocks in the dust storm event chain are divided into roles. The first block identified is the starting block. Blocks located after the starting block and whose deviation is greater than the corresponding deviation of the previous block, or whose absolute value of the difference between their deviation and the corresponding deviation of the previous block is not greater than the standard deviation of the corresponding surface reference group, are transmission blocks. Blocks located after the transmission block and whose deviation continuously decreases are attenuation blocks.

[0011] Furthermore, the anomaly removal in S5 specifically includes the following steps: Perform consistency checks on adjacent time intervals for each block in the sandstorm event chain; When there is no block carrying the same event chain identifier and satisfying the acceptance condition in adjacent time intervals for the target block, and the deviation of the target block relative to the corresponding similar surface reference group falls back to the corresponding standard deviation range, the target block is removed from the corresponding dust event chain; wherein, for blocks located in the middle of the chain, the adjacent time intervals include the previous time interval and the next time interval; for blocks located at the beginning of the chain, the adjacent time intervals include at least the next time interval; for blocks located at the end of the chain, the adjacent time intervals include at least the previous time interval.

[0012] Furthermore, the discontinuity correction in S5 specifically includes the following steps: When no block carrying the target event chain identifier is formed at a certain time, and the block in the previous time and the block in the next time carry the same event chain identifier, the analysis block located near the center line connecting the previous time block and the next time block, and which satisfies the connection condition with the previous time block and the next time block respectively, is selected as the continuation block, and the continuation block is merged into the corresponding dust event chain.

[0013] Furthermore, the dust monitoring and analysis results are generated based on the corrected dust event chain. The dust monitoring and analysis results include at least one of the following: dust index product, dust visibility product, dust propagation process information, and regional statistical results.

[0014] This invention also provides a dust monitoring and analysis system based on satellite data, comprising: The data acquisition and partitioning module is used to acquire continuous satellite observation data of the target area, divide the target area into blocks, and construct a corresponding similar surface reference group for each analysis block. The module for determining unconfirmed blocks is used to extract the dust characterization features of each analysis block in the current time, and to determine the unconfirmed blocks based on the deviation of the dust characterization features from the corresponding similar land surface reference group. The starting block determination module is used to search for accepting blocks that meet the accepting conditions with the block to be confirmed in subsequent time steps, and to determine the block to be confirmed as the starting block when the accepting block exists; An event chain construction module is used to assign an event chain identifier to the starting block and pass the event chain identifier to blocks that meet the acceptance conditions in subsequent times to form a sandstorm event chain. The result correction and output module is used to perform anomaly removal and intermittent continuation correction on the dust event chain to obtain dust monitoring and analysis results.

[0015] (III) Beneficial Effects: Compared with the prior art, the present invention provides a method and system for monitoring and analyzing sand and dust based on satellite data, which has the following beneficial effects: 1. This method and system for monitoring and analyzing dust storms based on satellite data divides the target area into blocks and constructs corresponding similar surface reference groups for each analysis block. This allows the current dust storm characteristics to be determined by a single fixed threshold, instead of directly judging them by comparing differences with similar or identical surface backgrounds. This effectively reduces the interference of deserts, bare land, high reflectivity backgrounds, and local short-term abnormal fluctuations on the identification results, improves the targeting and accuracy of the screening of blocks to be confirmed, and solves the problem of easy misjudgment in single-time image recognition in existing technologies.

[0016] 2. This satellite-based dust monitoring and analysis method and system searches for connecting blocks that meet the connection conditions with the block to be confirmed in subsequent time periods. When a connection relationship exists, the starting block is determined, and an event chain identifier is assigned to the starting block and passed on to subsequent blocks. This establishes the connection between dust regions in adjacent time periods on the basis of temporal continuity, spatial proximity, and feature continuity. It can more accurately identify the initiation, continuation, and evolution relationship of the same dust process, and avoid misconnection, omission, and mixing between different dust regions. This solves the problems of unclear association rules and unstable process tracking between different time periods in the existing technology.

[0017] 3. This satellite-based dust monitoring and analysis method and system, by performing anomaly removal and intermittent continuation correction on the formed dust event chain, removes blocks that do not meet the consistency requirements of the preceding and following times and whose abnormal characteristics have declined from the event chain. On the other hand, it corrects the chain breaks caused by short-term missing measurements or identification interruptions. This can improve the integrity and reliability of the dust event chain, thereby improving the accuracy of dust index products, dust visibility products, propagation process information, and regional statistical results. This solves the problems of dust process fragility and insufficient completeness of statistical analysis results in existing technologies. Attached Figure Description

[0018] Figure 1 A schematic diagram illustrating the steps of the dust monitoring and analysis method based on satellite data provided by this invention; Figure 2 A flowchart illustrating the dust monitoring and analysis method based on satellite data provided by this invention; Figure 3 A schematic flowchart of step S1 of the dust monitoring and analysis method based on satellite data provided by the present invention; Figure 4 A schematic diagram of the S2 flow of the dust monitoring and analysis method based on satellite data provided by the present invention; Figure 5 A schematic diagram of the S4 flow of the dust monitoring and analysis method based on satellite data provided by the present invention; Figure 6 A schematic diagram of the structure of the dust monitoring and analysis system based on satellite data provided by the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To enable those skilled in the art to better understand the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] Please see Figure 1 and Figure 2 , Figure 1 A schematic diagram illustrating the steps of the dust monitoring and analysis method based on satellite data provided by this invention; Figure 2 This is a flowchart illustrating the dust monitoring and analysis method based on satellite data provided by the present invention; the present invention provides a dust monitoring and analysis method based on satellite data, comprising: S1. Obtain continuous satellite observation data of the target area, divide the target area into blocks, and construct a corresponding similar surface reference group for each analysis block; Specifically, S1 is used to establish the basis for block-level time series analysis, enabling each analysis block to perform feature extraction, succession judgment, and event chain association in consecutive time periods, and to reduce the interference of surface background differences on dust discrimination by using similar surface reference groups.

[0022] For further details, please refer to Figure 3 , Figure 3 This is a flowchart illustrating step S1 of the dust monitoring and analysis method based on satellite data provided by the present invention; in one embodiment provided by this application, step S1 specifically includes the following steps: S11. Acquire multispectral satellite observation data of the target area over consecutive time intervals; S12. Divide the target area into grids according to fixed angular intervals in the longitude direction and fixed angular intervals in the latitude direction to obtain multiple analysis blocks; S13. For each analysis block, select reference blocks with the same or similar surface type and spatial proximity to the analysis block from satellite observation data from historical dust-free periods that are in the same period as the current analysis period, and form a corresponding similar surface reference group.

[0023] Specifically, in S11, the acquired data is multispectral satellite observation data covering the target area. The data time-series includes at least the current analysis time and subsequent times adjacent to it. In terms of bands, it can include visible light, infrared, and relevant bands used for brightness-temperature difference calculation. To facilitate subsequent block-level comparisons, data from different times are preferably organized using unified projected coordinates and a unified spatial coverage area, ensuring that the same geographical location corresponds to the same spatial region in each time period.

[0024] Specifically, in S12, when dividing the target area into grids, a fixed angular interval is used in both the longitude and latitude directions, jointly defining the spatial range of a single analysis block. After division, each analysis block forms a regular arrangement within the entire target area, with each block corresponding to a unique spatial location identifier. Using a fixed angular interval for grid division facilitates maintaining consistent block boundaries across different time periods and allows for subsequent searches of successor blocks and event chain connections based on block center locations and boundary adjacency relationships. In practice, the fixed angular interval can be set according to satellite observation resolution, target area range, and operational processing granularity.

[0025] In one embodiment, a single analysis block may consist of 3×3 to 10×10 pixels; preferably, a single analysis block consists of 5×5 to 8×8 pixels. When local latitude and longitude grids are used for division, a single analysis block may correspond to 0.02°×0.02° to 0.2°×0.2°; preferably, a single analysis block corresponds to 0.05°×0.05° to 0.1°×0.1°. When the satellite observation resolution is high, the analysis block range can be appropriately reduced; when the satellite observation resolution is low, the analysis block range can be appropriately increased, but the same target area maintains a consistent division method in all time periods.

[0026] Specifically, in S13, the similar surface reference group is used to characterize the background features of the corresponding analysis block under dust-free conditions. For any analysis block, candidate reference blocks corresponding to dust-free periods are first screened from historical satellite observation data. The dust-free periods can be determined based on at least one of historical operational annotation results, meteorological observation records, or existing dust-free determination results. Subsequently, reference blocks in the same period as the current analysis are further screened from the candidate reference blocks. Then, blocks with the same or similar surface type and spatial proximity to the analysis block are selected to form the similar surface reference group for that analysis block.

[0027] The term "same period" is used to limit the candidate reference block to having similar seasonal background and solar radiation conditions as the current analysis block. In one embodiment, "same period" may refer to the same month as the current analysis time; in another embodiment, "same period" may refer to a date difference of no more than 15 to 45 days from the current analysis time; in yet another embodiment, "same period" may refer to the same season as the current analysis time and a local time difference of no more than 2 hours. Preferably, a similar surface reference group is constructed using reference blocks corresponding to dust-free times in the same month as the current analysis time and with a local time difference of no more than 1 hour.

[0028] Land surface types can be classified based on their spectral performance, brightness temperature response, and land cover attributes in historical observations. Land surface types must include at least one of the following: bare land, sandy land, urban land, vegetated land, and water bodies. When land surface types are completely identical, reference blocks of that type are prioritized. If there are not enough reference blocks of the same type, the selection can be broadened to include land surface types with similar spectral or brightness temperature responses. Spatial proximity refers to the distance between the center point of a candidate reference block and the center point of the current analysis block not exceeding a proximity distance threshold, which can be set to 3 to 8 times the diagonal length of the current analysis block. To ensure statistical stability, the number of reference blocks in the same type of land surface reference group is preferably no less than 5. When the number of reference blocks meeting the conditions of the same period, the same or similar land surface type, and spatial proximity is insufficient, the spatial search range can be expanded first, and the number of reference blocks can be gradually increased while maintaining the same period conditions.

[0029] Once constructed, each analysis block corresponds to a set of background reference blocks. Subsequently, the dust characterization features extracted from the current analysis block can be compared with the statistical features of the same type of land surface reference group to determine whether the current analysis block shows abnormal changes.

[0030] In one embodiment, the selection order of reference blocks is as follows: first, select reference blocks that are in the same period as the current analysis time; then, select reference blocks with the same or similar surface types; and finally, select a preset number of reference blocks in order of spatial distance from near to far.

[0031] S2. Extract the dust characterization features of each analysis block in the current time period, and determine the blocks to be confirmed based on the deviation of the dust characterization features from the corresponding similar surface reference group; Specifically, S2 is used to screen analysis blocks with anomalous dust response characteristics from the block-level observation information of the current time period. Based on the similar surface reference group constructed by S1, the spectral and thermal infrared characteristics of the analysis block of the current time period are extracted and compared with the corresponding background level; when the same analysis block deviates from the corresponding background level in multiple dust characterization characteristics at the same time, it can be identified as a candidate for subsequent judgment.

[0032] For further details, please refer to Figure 4 , Figure 4 This is a flowchart illustrating step S2 of the dust monitoring and analysis method based on satellite data provided by the present invention; in one embodiment provided by this application, step S2 specifically includes the following steps: S21. Extract at least two of the visible light reflectance, infrared brightness temperature, and brightness temperature difference of each analysis block in the current time as characterization features of sand and dust. S22. Calculate the deviation of the dust characterization characteristics of each analysis block from the mean of the corresponding similar land surface reference group; S23. When the deviation of the same analysis block in at least two dust characterization features exceeds the standard deviation of the corresponding similar surface reference group, the analysis block shall be identified as a block to be confirmed.

[0033] Specifically, in S21, the dust characterization features are selected from at least two of the following: visible light reflectance, infrared brightness temperature, and brightness-temperature difference. Visible light reflectance characterizes the visible light reflection response of the analysis block at the current time, suitable for reflecting the scattering effect of particles above the ground on incident radiation; infrared brightness temperature characterizes the radiation temperature response of the analysis block in the thermal infrared band, suitable for reflecting changes in the thermal radiation characteristics of the block under dust cover; brightness-temperature difference characterizes the response differences between different infrared bands, suitable for depicting the differences in thermal infrared characteristics between dust and ordinary surface background, clouds, or other targets. In practice, the visible light reflectance and infrared brightness temperature corresponding to a single band can be directly extracted, or the brightness-temperature difference can be obtained by subtracting the brightness temperature values ​​of two infrared bands. Using at least two features to jointly characterize the analysis block allows the determination of the block to be confirmed to be based on multi-dimensional remote sensing response, rather than relying solely on changes in a single band.

[0034] Specifically, in step S22, for each analysis block, the mean and standard deviation of each dust characterization feature are statistically analyzed within the corresponding similar land surface reference group. Then, the value of the current analysis block on the corresponding feature is compared with the mean to obtain the deviation of the analysis block on the corresponding feature. In one embodiment, the deviation can be defined as the absolute difference between the corresponding feature value of the current analysis block and the mean of the corresponding feature in the similar land surface reference group. Further, the ratio of the absolute difference to the corresponding standard deviation can be used as a normalized deviation value to characterize the degree of anomaly.

[0035] The deviation is used to characterize the degree of anomaly of the current analysis block relative to similar background blocks. If the deviation of the current analysis block from its similar surface reference group on a certain feature is small, it indicates that the analysis block is still close to the background state in that feature dimension; if the deviation is large, it indicates that the analysis block has shown a significantly different current temporal response from its similar background in that feature dimension. By calculating the deviation for each analysis block, the anomaly distribution results for the current time can be formed at the block scale.

[0036] Specifically, in S23, for the same analysis block, it is determined whether the deviation of each dust characterization feature exceeds the standard deviation of the corresponding similar land surface reference group. Since the deviation is expressed as an absolute difference in this embodiment, it can be directly compared with the corresponding standard deviation. When at least two dust characterization features simultaneously meet this condition, the analysis block is identified as a block to be confirmed. The standard deviation is used as the judgment criterion to match the screening of blocks to be confirmed with the dispersion of the corresponding similar land surface reference group itself. For similar land surface reference groups with relatively small background fluctuations, small feature changes can be identified as anomalies; for similar land surface reference groups with relatively large background fluctuations, only changes exceeding their normal fluctuation range will enter the scope of confirmation. By adopting the judgment method of "at least two features simultaneously meeting", the blocks to be confirmed can simultaneously possess multiple abnormal response features, thereby concentrating the blocks entering subsequent judgments on the candidate areas that are more worthy of attention in the current time period.

[0037] S3. In subsequent iterations, search for a receiving block that meets the receiving conditions of the block to be confirmed, and if the receiving block exists, determine the block to be confirmed as the starting block; Specifically, S3 is used to confirm whether the block to be confirmed has a subsequent continuation relationship. For the block to be confirmed selected in the current time step, it is not directly identified as the starting block. Instead, it is searched in subsequent time steps for a successor block that forms a continuous relationship with it in terms of time, space, and feature changes. When a successor block that meets the succession conditions exists, the block to be confirmed is identified as the starting block. Otherwise, it does not enter the subsequent event chain construction process.

[0038] Furthermore, in one embodiment provided in this application, the acceptance condition specifically includes: In subsequent time intervals, the target block and the previous time interval block used as the comparison benchmark are in the next or second time interval. The previous time interval block used as the comparison benchmark is the block to be confirmed, or a block carrying an event chain identifier. The target block and the previous time interval block used as the comparison benchmark are spatially connected, or the distance between their center points is no greater than twice the diagonal length of the previous time interval block used as the comparison benchmark. Furthermore, the target block and the previous time interval block used as the comparison benchmark show the same direction of change in dust characterization features, and the deviation of the target block relative to its corresponding similar surface reference group is no less than half of the corresponding deviation of the previous time interval block used as the comparison benchmark.

[0039] Specifically, the acceptance criteria define the correspondence between subsequent target blocks and previous blocks based on three aspects: temporal adjacency, spatial adjacency, and feature continuity. Temporal adjacency is limited to the next or second-to-last time period to restrict the search scope to a short time window following the current block to be confirmed, ensuring that the acceptance judgment revolves around a continuous evolution process. If the time span is too large, many intermediate processes are missing, making it difficult to distinguish between the continuity of the same dust storm process and accidental proximity between different processes. Limiting the search scope to the next or second-to-last time period maintains a close temporal correlation between the block to be confirmed and the candidate accepting blocks. When currently in the initial block determination stage, the benchmark block is the block to be confirmed; when entering the subsequent event chain extension stage, the benchmark block switches to a block already carrying an event chain identifier, allowing the same acceptance criteria to be used for both initial block confirmation and subsequent event chain extension.

[0040] Spatial adjacency is used to define the spatial continuity possibility between the target block and the benchmark block. Spatially adjacent upper boundaries indicate that the two blocks are directly adjacent at the grid boundary, suitable for characterizing the local expansion or translation of the same dust region in adjacent time intervals. The distance between their center points is no greater than twice the diagonal length of the benchmark block, indicating that even if the two blocks are not directly in contact at the boundary, as long as they are still within a close range of the benchmark block, they can be considered candidate contiguous blocks. Using the diagonal length as the spatial scale benchmark is directly related to the grid size of the analysis block itself, facilitating consistent contiguousness judgments across different grid granularities. When the target block exceeds this range, it indicates that the spatial interval between the target block and the benchmark block has deviated from the local continuity relationship, and it is generally not treated as a contiguous block.

[0041] Feature continuity is used to determine the continuity of the dust response between the target block and the benchmark block. The consistent direction of change can refer to the target block and the benchmark block having the same direction of deviation relative to the mean of their respective similar surface reference groups, or the same trend in the normalized deviation, across at least two commonly used dust characterization features. Specifically, when a feature is higher or lower than the mean of the corresponding reference group in both blocks, the direction of change of that feature can be determined to be consistent; when at least two features meet the above conditions, the two blocks can be determined to have the same direction of change in dust characterization features.

[0042] A deviation of no less than half the deviation of the benchmark block indicates that while the target block may show some reduction compared to the benchmark block, it still retains characteristic responses sufficient to reflect an abnormal state. This proportional limit allows for local diffusion, weakening, or boundary changes in the dust process in subsequent time intervals, while avoiding the incorrect inclusion of blocks that have clearly returned to their background state within the acceptance range. Only target blocks that simultaneously satisfy temporal adjacency, spatial adjacency, and characteristic continuity are recognized as acceptance blocks.

[0043] During the determination of the starting block, for each block to be confirmed, the analysis blocks in the next and second subsequent time periods can be searched sequentially, and compared one by one according to the acceptance criteria. If there is at least one target block that meets the acceptance criteria within the search range, the current block to be confirmed is determined as the starting block; if there are multiple target blocks that meet the acceptance criteria, these target blocks can all be used as candidates for the acceptance direction of the current starting block in subsequent time periods, for further processing in the subsequent event chain extension stage.

[0044] In one embodiment, if multiple target blocks satisfying the acceptance criteria exist in the same subsequent time interval, these multiple target blocks can be simultaneously used as candidate acceptance blocks for the current starting block and incorporated into the same event chain. If the same target block simultaneously satisfies the acceptance criteria of multiple different event chains, its assigned event chain can be determined according to at least one of the following rules: minimum center point distance priority, minimum deviation difference priority, or earlier starting time interval priority. If no target block satisfying the acceptance criteria is found in the next time interval or the second next time interval, the current block to be confirmed will not be processed as a starting block in the current analysis round.

[0045] S4. Assign an event chain identifier to the starting block and pass the event chain identifier to the blocks that meet the acceptance conditions in subsequent times to form a sandstorm event chain; Specifically, S4 is used to establish a unified event chain identifier around the confirmed starting block, and continuously associate the analysis blocks that meet the conditions in subsequent time periods according to the acceptance conditions, thereby forming a cross-time dust event chain.

[0046] For further details, please refer to Figure 5 , Figure 5 This is a flowchart illustrating step S4 of the dust monitoring and analysis method based on satellite data provided by the present invention; in one embodiment provided by this application, step S4 specifically includes the following steps: S41. Assign a unique event chain identifier to the determined starting block, wherein the event chain identifier includes at least the event number and the starting time. S42. In subsequent iterations, search for analysis blocks that satisfy the acceptance conditions and carry the event chain identifier. S43. Associate the analysis blocks that meet the acceptance conditions with the corresponding event chain identifier to form a sandstorm event chain.

[0047] Specifically, in S41, the event chain identifier is used to represent the unified identity of the same dust storm event. For each confirmed starting block, after assigning a unique event chain identifier, that starting block becomes the starting point of the corresponding dust storm event chain. The event number is used to distinguish different dust storm event chains, and the start time is used to record the starting position of the event chain in the time dimension. In actual processing, event numbers can be generated sequentially within the current processing batch according to the confirmation order of the starting blocks, and the event number and start time are written together into the event attributes of the starting block. After this processing is completed, each starting block is no longer just an anomalous block in space, but a chain starting point with event attribution information.

[0048] Specifically, in S42, the block already carrying the event chain identifier is used as the current comparison benchmark block. Within its subsequent time range, an analysis block is searched, and each block is evaluated according to the continuation condition to determine whether it can form a continuation relationship with the current comparison benchmark block. The search time range can use the next and second-to-last time range settings from S3, ensuring a continuous time window for the event chain as it extends. The current comparison benchmark block can be either the starting block or other blocks already merged into the event chain. In this way, the event chain search process can proceed progressively along the time direction: when the starting block finds an analysis block that meets the continuation condition, the newly merged analysis block continues as the comparison benchmark block for the next round of searching, thus continuously expanding the event chain from front to back.

[0049] Specifically, in S43, for analysis blocks that meet the continuation conditions, they are associated with the same event chain identifier as the current benchmark block, and their corresponding time position is recorded. If a benchmark block corresponds to multiple analysis blocks that meet the continuation conditions in a later time period or two later time periods, these analysis blocks can all be merged into the same event chain identifier, indicating that the dust storm process has expanded, branched, or continued in parallel with multiple blocks in subsequent time periods. If no analysis block that meets the continuation conditions is found in subsequent time periods, the event chain extension of the current benchmark block in that direction ends. Through the above association process, blocks of the same dust storm process in multiple consecutive time periods can be gradually linked together to form a dust storm event chain.

[0050] Furthermore, in one embodiment provided in this application, the following steps are included after step S4: Based on the temporal position of each block in the dust storm event chain and the changing state of dust storm characteristics, the blocks in the dust storm event chain are divided into roles. The first identified block is designated as the starting block; blocks following the starting block whose deviation is greater than the corresponding deviation of the preceding block, or whose absolute difference between their deviation and the corresponding deviation of the preceding block is not greater than the standard deviation of the corresponding surface reference group, are designated as transmission blocks; blocks following the transmission blocks whose deviation continuously decreases within a preset consecutive time window are designated as attenuation blocks. In one embodiment, the preset consecutive time window may be two consecutive time periods or three consecutive time periods.

[0051] Specifically, role classification is used to distinguish the functional position of each block in the dust storm event chain during the event's development, based on the established chain. The block that is first identified and assigned an event chain identifier is located at the very beginning of the entire event chain, corresponding to the starting position of the dust storm process, and is therefore designated as the starting block. In blocks following the starting block, if the deviation continues to increase, it indicates that the anomaly of the current block relative to the similar surface background is still increasing; if the absolute value of the difference between the deviation and the corresponding deviation of the previous block is not greater than the standard deviation of the corresponding similar surface reference group, it indicates that the current block maintains an approximately continuous state of anomaly with the previous block. Both types of blocks can characterize the continued spatial propagation and maintenance of the dust storm process, and are therefore designated as transmission blocks. Blocks located after the transmission blocks whose deviation shows a decreasing trend for at least two consecutive time intervals indicate that the anomaly of the block relative to the background is gradually decreasing, corresponding to the dust storm process entering a weakening phase, and are therefore designated as decay blocks.

[0052] Role classification can be performed segment by segment along the chronological order of the same event chain. First, the first block of the event chain is marked as the starting block. Then, the deviation changes of each subsequent block from the previous block are compared, and they are sequentially determined as either transmission blocks or attenuation blocks. Once a block in the event chain has entered the attenuation phase, if subsequent blocks continue to show decreasing deviations, they are further classified as attenuation blocks. Based on this role classification, the start, continuation, and weakening range of a dust storm process can be further expressed.

[0053] S5. Perform anomaly removal and intermittent connection correction on the dust event chain to obtain dust monitoring and analysis results.

[0054] Specifically, S5 is used to correct the continuity and effectiveness of established dust storm event chains, including anomaly removal and discontinuity correction, to improve the integrity and reliability of the event chains.

[0055] Furthermore, in one embodiment provided in this application, the anomaly removal in S5 specifically includes the following steps: Perform consistency checks on adjacent time intervals for each block in the sandstorm event chain; When there is no block carrying the same event chain identifier and satisfying the acceptance condition in adjacent time intervals for the target block, and the deviation of the target block relative to the corresponding similar surface reference group falls back to the corresponding standard deviation range, the target block is removed from the corresponding dust event chain; wherein, for blocks located in the middle of the chain, the adjacent time intervals include the previous time interval and the next time interval; for blocks located at the beginning of the chain, the adjacent time intervals include at least the next time interval; for blocks located at the end of the chain, the adjacent time intervals include at least the previous time interval.

[0056] Specifically, anomaly removal is used to identify and remove isolated blocks in the event chain that lack continuous support. For any block in the event chain, it is first checked in the previous and next time periods to see if there are any blocks carrying the same event chain identifier that can form a successor relationship with the current block. If no corresponding block meeting the conditions is found in either direction, it means that the block is not connected to other blocks in the chain in time. Then, it is judged based on whether the deviation of the block relative to the same type of surface reference group has fallen back to within the standard deviation range. If the deviation has returned to the normal fluctuation range of the same background, it indicates that the block no longer maintains a sufficiently clear anomalous state in the current time period, and the block is removed from the corresponding dust event chain. This avoids keeping short-term isolated responses lacking continuous support in the event chain.

[0057] Consistency checks are performed sequentially, one block at a time, within the event chain. For blocks in the middle of the chain, the correspondence between the preceding and following time events is checked simultaneously; for blocks at the beginning of the chain, the correspondence between the following time events is checked; and for blocks at the end of the chain, the correspondence between the preceding time events is checked. If a block cannot find a block that meets the continuity criteria on one side, but a block that can maintain continuity with it exists on the other side, then the block remains in the event chain and is not treated as an anomalous block. Anomaly removal targets blocks that lack continuity support and have fallen back into the background fluctuation range, rather than simply deleting all locally weakly connected blocks.

[0058] Furthermore, in one embodiment provided in this application, the discontinuity correction in S5 specifically includes the following steps: When no block carrying the target event chain identifier is formed at a certain time, and the block in the previous time and the block in the next time carry the same event chain identifier, the analysis block located near the center line connecting the previous time block and the next time block, and which satisfies the connection condition with the previous time block and the next time block respectively, is selected as the continuation block, and the continuation block is merged into the corresponding dust event chain.

[0059] Specifically, discontinuous continuation correction is used to handle short-term breaks in event chains. When a block carrying the target event chain identifier is not directly formed at a certain time point, but the blocks before and after that time point indicate that the same event chain exists continuously in time, the missing time point can be regarded as a candidate continuation time point. Subsequently, within the candidate continuation time point, an analysis block located near the center line connecting the blocks of the previous and subsequent time points is searched, and it is further determined whether the analysis block satisfies the continuation conditions with the blocks of the previous and subsequent time points respectively. If both conditions are met, it indicates that the analysis block is located near the continuous propagation paths at both ends in spatial location and can also form a continuous relationship with the blocks at both ends in terms of feature changes, so it can be added to the corresponding event chain as a continuation block.

[0060] The area near the center line is used to define the priority search region for continuation blocks. Using the line connecting the center of a previous block and the center of a subsequent block as a reference, if the vertical distance from the center point of a candidate analysis block to this line is not greater than a preset proximity threshold, the analysis block can be considered a candidate continuation block. This preset proximity threshold can be set to 0.5 to 1.5 times the diagonal length of a single analysis block. Further, blocks covered by the line, blocks directly intersecting the line, and several analysis blocks closest to the line can be considered as candidate continuation blocks, and then filtered according to continuation conditions. This ensures that continuation correction maintains both temporal continuity and consistent spatial direction. If multiple analysis blocks satisfying the conditions exist within a candidate time period, these blocks can be merged into the corresponding event chain, indicating that the event chain in the missing time period simultaneously covers multiple blocks within a local area; if no analysis block satisfying the conditions is found within a candidate time period, continuation correction is not performed for that time period.

[0061] Furthermore, in one embodiment provided in this application, the dust monitoring and analysis results are generated based on the modified dust event chain, and the dust monitoring and analysis results include at least one of dust index products, dust visibility products, dust propagation process information, and regional statistical results.

[0062] Specifically, the dust index product forms a thematic distribution result based on the dust characterization characteristics, deviation, and spatial location of each block in the corrected event chain at the current time, which is used to express the spatial distribution of dust response in the target area. In one embodiment, the deviation of visible light reflectance, infrared brightness temperature, and brightness temperature difference can be weighted and summed to obtain the block-level dust index, and the block-level dust index can be mapped back to the target area to form a dust index thematic map.

[0063] Dust visibility products can generate block-level visibility distribution results based on the feature combination relationships corresponding to the event chain coverage blocks, and map each block result back to the target area to obtain the visibility thematic map for the corresponding time period. In one embodiment, the block-level dust characterization feature combination values ​​can be converted into corresponding visibility values ​​based on the empirical mapping relationship, regression model, or lookup table obtained from sample calibration. Since both types of products are based on the corrected event chain, the blocks involved in the result generation have undergone anomaly removal and intermittent continuation correction, and the block distribution in the product is consistent with the evolution process of continuous time periods.

[0064] Information on dust propagation processes is used to express the start, continuation, and decay of the same event chain across multiple consecutive time periods. The start time, duration, propagation direction, and set of covered blocks for each time period can be generated based on the time position of each block in the event chain, the event chain identifier, and the role classification results. The propagation direction can be determined based on the displacement vector of the center point of the corresponding block in adjacent time periods. Regional statistical results can be compiled around administrative regions or custom regional boundaries. For example, it can count the number of blocks carrying the same event chain identifier or any event chain identifier, the corresponding coverage area, and the distribution of occurrence times in each region within a given time period. The coverage area can be calculated by multiplying the number of blocks by the area of ​​a single analyzed block.

[0065] This invention also provides a dust monitoring and analysis system based on satellite data, used to implement the above-mentioned dust monitoring and analysis method based on satellite data, including: The data acquisition and partitioning module is used to acquire continuous satellite observation data of the target area, divide the target area into blocks, and construct a corresponding similar surface reference group for each analysis block. The module for determining unconfirmed blocks is used to extract the dust characterization features of each analysis block in the current time, and to determine the unconfirmed blocks based on the deviation of the dust characterization features from the corresponding similar land surface reference group. The starting block determination module is used to search for accepting blocks that meet the accepting conditions with the block to be confirmed in subsequent time steps, and to determine the block to be confirmed as the starting block when the accepting block exists; An event chain construction module is used to assign an event chain identifier to the starting block and pass the event chain identifier to blocks that meet the acceptance conditions in subsequent times to form a sandstorm event chain. The result correction and output module is used to perform anomaly removal and intermittent continuation correction on the dust event chain to obtain dust monitoring and analysis results.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for sand and dust monitoring and analysis based on satellite data, characterized in that, include: S1. Obtain continuous satellite observation data of the target area, divide the target area into blocks, and construct a corresponding similar surface reference group for each analysis block; S2. Extract the dust characterization features of each analysis block in the current time period, and determine the blocks to be confirmed based on the deviation of the dust characterization features from the corresponding similar surface reference group; S3. In subsequent iterations, search for a receiving block that meets the receiving conditions of the block to be confirmed, and if the receiving block exists, determine the block to be confirmed as the starting block; S4. Assign an event chain identifier to the starting block and pass the event chain identifier to the blocks that meet the acceptance conditions in subsequent times to form a sandstorm event chain; S5. Perform anomaly removal and intermittent connection correction on the dust event chain to obtain dust monitoring and analysis results.

2. The satellite data based dust monitoring analysis method of claim 1, wherein, S1 specifically includes the following steps: S11. Acquire multispectral satellite observation data of the target area over consecutive time intervals; S12. Divide the target area into grids according to fixed angular intervals in the longitude direction and fixed angular intervals in the latitude direction to obtain multiple analysis blocks; S13. For each analysis block, select reference blocks with the same or similar surface type and spatial proximity to the analysis block from satellite observation data from historical dust-free periods that are in the same period as the current analysis period, and form a corresponding similar surface reference group. 3.The satellite data based dust monitoring analysis method of claim 1, wherein, S2 specifically includes the following steps: S21. Extract at least two of the visible light reflectance, infrared brightness temperature, and brightness temperature difference of each analysis block in the current time as characterization features of sand and dust. S22. Calculate the deviation of the dust characterization characteristics of each analysis block from the mean of the corresponding similar land surface reference group; S23. When the deviation of the same analysis block in at least two dust characterization features exceeds the standard deviation of the corresponding similar surface reference group, the analysis block shall be identified as a block to be confirmed. 4.The satellite data based dust monitoring analysis method of claim 1, wherein, The specific conditions for acceptance are as follows: In subsequent time intervals, the target block and the previous time interval block used as the comparison benchmark are in the next or second time interval. The previous time interval block used as the comparison benchmark is the block to be confirmed, or a block carrying an event chain identifier. The target block and the previous time interval block used as the comparison benchmark are spatially connected, or the distance between their center points is no greater than twice the diagonal length of the previous time interval block used as the comparison benchmark. Furthermore, the target block and the previous time interval block used as the comparison benchmark show the same direction of change in dust characterization features, and the deviation of the target block relative to its corresponding similar surface reference group is no less than half of the corresponding deviation of the previous time interval block used as the comparison benchmark.

5. The satellite data based dust monitoring analysis method of claim 4, wherein, S4 specifically includes the following steps: S41. Assign a unique event chain identifier to the determined starting block, wherein the event chain identifier includes at least the event number and the starting time. S42. In subsequent iterations, search for analysis blocks that satisfy the acceptance conditions and carry the event chain identifier. S43. Associate the analysis blocks that meet the acceptance conditions with the corresponding event chain identifier to form a sandstorm event chain.

6. The satellite data based dust monitoring analysis method of claim 5, wherein, The following steps are included after step S4: Based on the temporal position of each block in the dust storm event chain and the changing state of dust storm characteristics, the blocks in the dust storm event chain are divided into roles. The first block identified is the starting block. Blocks located after the starting block and whose deviation is greater than the corresponding deviation of the previous block, or whose absolute value of the difference between their deviation and the corresponding deviation of the previous block is not greater than the standard deviation of the corresponding surface reference group, are transmission blocks. Blocks located after the transmission block and whose deviation continuously decreases are attenuation blocks.

7. The method for monitoring and analyzing sandstorms based on satellite data according to claim 5, characterized in that, The anomaly removal in S5 specifically includes the following steps: Perform consistency checks on adjacent time intervals for each block in the sandstorm event chain; When there is no block carrying the same event chain identifier and satisfying the acceptance condition in adjacent time intervals for the target block, and the deviation of the target block relative to the corresponding similar surface reference group falls back to the corresponding standard deviation range, the target block is removed from the corresponding dust event chain; wherein, for blocks located in the middle of the chain, the adjacent time intervals include the previous time interval and the next time interval; for blocks located at the beginning of the chain, the adjacent time intervals include at least the next time interval; for blocks located at the end of the chain, the adjacent time intervals include at least the previous time interval.

8. The method for monitoring and analyzing sandstorms based on satellite data according to claim 5, characterized in that, The discontinuous connection correction in S5 specifically includes the following steps: When no block carrying the target event chain identifier is formed at a certain time, and the block in the previous time and the block in the next time carry the same event chain identifier, the analysis block located near the center line connecting the previous time block and the next time block, and which satisfies the connection condition with the previous time block and the next time block respectively, is selected as the continuation block, and the continuation block is merged into the corresponding dust event chain.

9. The method for monitoring and analyzing sandstorms based on satellite data according to claim 8, characterized in that: The dust monitoring and analysis results are generated based on the corrected dust event chain. The dust monitoring and analysis results include at least one of the following: dust index product, dust visibility product, dust propagation process information, and regional statistical results.

10. A dust monitoring and analysis system based on satellite data, characterized in that, include: The data acquisition and partitioning module is used to acquire continuous satellite observation data of the target area, divide the target area into blocks, and construct a corresponding similar surface reference group for each analysis block. The module for determining unconfirmed blocks is used to extract the dust characterization features of each analysis block in the current time, and to determine the unconfirmed blocks based on the deviation of the dust characterization features from the corresponding similar land surface reference group. The starting block determination module is used to search for accepting blocks that meet the accepting conditions with the block to be confirmed in subsequent time steps, and to determine the block to be confirmed as the starting block when the accepting block exists; An event chain construction module is used to assign an event chain identifier to the starting block and pass the event chain identifier to blocks that meet the acceptance conditions in subsequent times to form a sandstorm event chain. The result correction and output module is used to perform anomaly removal and intermittent continuation correction on the dust event chain to obtain dust monitoring and analysis results.

Images