Method and device for acquiring remote sensing clear sky data set, electronic equipment and storage medium

Abstract

The invention relates to the technical field of remote sensing data processing, and particularly discloses a remote sensing clear sky data set acquisition method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the preprocessing of remote sensing image data collected in a preset time period, and obtaining a global remote sensing image time sequence; selecting a target remote sensing image sequential sequence of the target area from the global remote sensing image sequential sequence, and extracting the reflectivity spectral characteristics of each pixel in the target wave band and the NDVI of each pixel from the target remote sensing image sequential sequence; aiming at a plurality of pixels at the same target position but with different time sequences in the time sequence of the target remote sensing image, clustering the plurality of pixels by using a K-means algorithm to obtain a plurality of clusters; identifying clear sky clustering clusters in the clustering clusters; selecting a clear sky pixel of the target position from the clear sky cluster; and synthesizing a clear sky image of the target area based on the clear sky pixels of the target positions in the target area. According to the technical scheme, clear sky data can be simply and quickly obtained.

Description

Technical Field

[0001] The present disclosure relates to the technical field of remote sensing data processing, and in particular to a method, device, electronic device and storage medium for acquiring a remote sensing clear sky data set. Background Art

[0002] Remote sensing clear sky dataset processing is a process that uses remote sensing technology to obtain image data from satellites or aerial platforms, and uses specific algorithms and analysis methods to distinguish cloud-free areas with clear atmospheric conditions (i.e., "clear sky") from multi-temporal data and perform clear sky synthesis. Its purpose is to improve the availability and quality of remote sensing data, ensure that subsequent applications such as surface feature extraction, environmental monitoring, and resource assessment can be based on accurate and clear data, thereby enhancing the practical application effect and scientific research value of remote sensing technology.

[0003] Traditional remote sensing clear sky dataset processing methods mainly rely on the identification and removal of cloud information in remote sensing images to extract clear sky areas. Clouds usually have high reflectivity and low transmittance, which is particularly evident in the visible and infrared bands. Traditional solutions mostly analyze the spectral characteristics (such as reflectivity, brightness temperature, etc.) of a large number of bands of remote sensing images such as visible light, infrared and microwave bands, as well as relevant data from a variety of external data sources (such as meteorological data, terrain data, etc.), identify and calculate the physical properties of clouds such as temperature, height, thickness, etc., and then distinguish clear sky areas to obtain clear sky data. However, this type of method usually requires comprehensive analysis and judgment based on a large amount of band information of remote sensing images and external data sources, and seasonal factors need to be considered during the analysis process, making the solution more complicated. Summary of the invention

[0004] In order to solve the problems in the related art, the embodiments of the present disclosure provide a method, device, electronic device and storage medium for acquiring a remote sensing clear sky dataset.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for acquiring a remote sensing clear sky dataset, comprising: Acquiring remote sensing image data collected within a predetermined time period; Preprocessing the remote sensing image data to obtain a global remote sensing image time series sequence within a predetermined time period, wherein the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order; Selecting a target remote sensing image time series sequence of a target area from the global remote sensing image time series sequence; Extracting the reflectance spectral characteristics of each pixel in the target band and the normalized difference vegetation index NDVI of each pixel from the target remote sensing image time series; For a plurality of pixels at the same target position but with different time series in the time series of the target remote sensing image, clustering the plurality of pixels using a K-means algorithm to obtain a plurality of cluster clusters; identifying a clear sky cluster among the clusters; Selecting a clear sky pixel at the target location from the clear sky cluster; A clear sky image of the target area is synthesized based on clear sky pixels at each target position in the target area.

[0006] In one possible implementation, The remote sensing image data is L1 level data of the MERSI payload of the Fengyun-3 satellite; the preprocessing of the remote sensing image data set to obtain a global remote sensing image time series within a predetermined time period includes: Perform atmospheric correction, equal longitude and latitude projection and data splicing on the collected remote sensing image data to obtain multiple global remote sensing images; The plurality of global remote sensing images within the predetermined time period are sorted in chronological order to form a global remote sensing image time series sequence.

[0007] In a possible implementation, the atmospheric correction process includes: Acquire daytime remote sensing image data from the remote sensing image data; Based on the daytime remote sensing image data, the original reflectivity and angle information of each pixel in the daytime remote sensing image data in each target band are obtained, wherein the angle information includes the solar altitude angle, the solar zenith angle, the satellite altitude angle, the satellite zenith angle, and the elevation data; When the solar zenith angle satisfies the correction condition, the original reflectivity of each pixel in the daytime remote sensing image data is corrected according to the angle information to obtain the remote sensing image data after atmospheric correction.

[0008] In a possible implementation, the equal longitude and latitude projection processing includes: Perform equal longitude and latitude projection on the remote sensing image data after atmospheric correction to obtain initial map projection data of a predetermined resolution; The initial map projection data is interpolated to obtain map projection data.

[0009] In a possible implementation, the data splicing process includes: Acquire the target band of the map projection data; For each target band, obtaining the equal longitude and latitude information of the map projection data of the target band; Based on the equal longitude and latitude information of the map projection data of the target band, multiple blocks of map projection data of the target band are synthesized into block data of a predetermined longitude and latitude span; Based on the satellite zenith angle of the block data of the predetermined longitude and latitude span, the block data of the predetermined longitude and latitude span are synthesized into whole block data, and then the whole block data are spliced ​​to obtain the global remote sensing image of the target band.

[0010] In a possible implementation, the target waveband includes a blue light waveband, a green light waveband, a red light waveband, and a near infrared waveband.

[0011] In a possible implementation, clustering the multiple pixels using a K-means algorithm to obtain multiple clusters includes: Using K-means algorithm to cluster the multiple pixels to obtain two clusters; identifying a suspected clear sky cluster among the two clusters; Using the K-means algorithm to cluster the pixels in the suspected clear sky cluster again, obtaining three clusters; A clear sky cluster is identified among the three clusters.

[0012] In a possible implementation manner, selecting the clear sky pixel of the target location from the clear sky cluster includes: A pixel with the largest NDVI is selected from the clear sky cluster as the clear sky pixel of the target location.

[0013] In a second aspect, an embodiment of the present disclosure provides a device for acquiring a remote sensing clear sky dataset, comprising: An acquisition module is configured to acquire remote sensing image data collected within a predetermined time period; A preprocessing module is configured to preprocess the remote sensing image data to obtain a global remote sensing image time series sequence within a predetermined time period, wherein the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order; A region selection module is configured to select a target remote sensing image time series sequence of a target region from a global remote sensing image time series sequence; A feature extraction module is configured to extract the reflectance spectral characteristics of each pixel in the target band and the normalized difference vegetation index NDVI of each pixel from the target remote sensing image time series; A clustering module is configured to cluster multiple pixels at the same target position but in different time series in the time series of the target remote sensing image using a K-means algorithm to obtain multiple clusters; an identification module, configured to identify a clear sky cluster among the clusters; A pixel selection module is configured to select a clear sky pixel of the target location from the clear sky cluster; The image synthesis module is configured to synthesize the clear sky image of the target area based on the clear sky pixels at each target position in the target area.

[0014] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer instructions, and wherein the one or more computer instructions are executed by the processor to implement a method as described in any one of the first aspects.

[0015] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium on which computer instructions are stored. When the computer instructions are executed by a processor, a method as described in any one of the first aspects is implemented.

[0016] According to the technical solution provided by the embodiment of the present disclosure, a high-quality global remote sensing image of a global scope can be constructed by preprocessing the remote sensing image data collected within a predetermined time period, and then a target remote sensing image time series sequence of the target area is selected from the global remote sensing image time series sequence; the reflectance spectral characteristics of each pixel in the target band and the NDVI of each pixel are extracted from the target remote sensing image time series sequence; then, for multiple pixels at the same target position but different time series in the target remote sensing image time series sequence, the multiple pixels are clustered using the K-means algorithm to obtain multiple cluster clusters; clear sky cluster clusters in the cluster clusters are identified; clear sky pixels at the target position are selected from the clear sky cluster clusters; and a clear sky image of the target area is synthesized based on the clear sky pixels at each target position in the target area, so that a high-quality clear sky image can be obtained. This scheme adopts the K-means algorithm, which only requires the reflectance spectral characteristics of several target bands and the NDVI of each pixel to cluster the clear sky pixels at the target location. The clear sky pixels at each target location in the target area can be synthesized into a clear sky image of the target area. Less data is required to identify clear sky pixels and it is not affected by other external factors. The analysis scheme is simple and easy to implement, and it maintains high efficiency and stability when processing large-scale remote sensing image data.

[0017] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objectives and advantages of the present disclosure will become more apparent through the following detailed description of non-limiting embodiments in conjunction with the accompanying drawings. In the accompanying drawings:

[0019] Figure 1A flowchart of a method for acquiring a remote sensing clear sky dataset according to an embodiment of the present disclosure is shown.

[0020] Figure 2 A schematic diagram showing the reflectance spectral characteristics and NDVI of pixels at the same target position extracted from a time series of target remote sensing images within a predetermined time period provided by an embodiment of the present disclosure in a target band.

[0021] Figure 3 A structural block diagram of a device for acquiring a remote sensing clear sky dataset according to an embodiment of the present disclosure is shown.

[0022] FIG. 4 shows a structural block diagram of an electronic device according to an embodiment of the present disclosure.

[0023] FIG5 is a schematic diagram showing the structure of a computer system suitable for implementing the method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement them. In addition, for the sake of clarity, parts not related to the description of the exemplary embodiments are omitted in the accompanying drawings.

[0025] In the present disclosure, it should be understood that terms such as "include" or "have" are intended to indicate the presence of features, numbers, steps, behaviors, components, parts, or a combination thereof disclosed in the present specification, and are not intended to exclude the possibility that one or more other features, numbers, steps, behaviors, components, parts, or a combination thereof exist or are added.

[0026] It should also be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0027] Figure 1 A flowchart of a method for acquiring a remote sensing clear sky dataset according to an embodiment of the present disclosure is shown. Figure 1 As shown, the method includes the following steps S101-S108:

[0028] In step S101, remote sensing image data collected within a predetermined time period is obtained; In step S102, the remote sensing image data is preprocessed to obtain a global remote sensing image time series sequence within a predetermined time period, wherein the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order; In step S103, a target remote sensing image time series sequence of a target area is selected from the global remote sensing image time series sequence; In step S104, the reflectance spectral characteristics of each pixel in the target band and the normalized difference vegetation index NDVI of each pixel are extracted from the target remote sensing image time series; In step S105, for a plurality of pixels at the same target position but in different time series in the time series of the target remote sensing image, the K-means algorithm is used to cluster the plurality of pixels to obtain a plurality of clusters; In step S106, a clear sky cluster is identified in the clusters; In step S107, a clear sky pixel of the target location is selected from the clear sky cluster; In step S108, a clear sky image of the target area is synthesized based on the clear sky pixels at each target position in the target area.

[0029] In a possible implementation, the method for acquiring a remote sensing clear sky dataset is applicable to a computer, computing device, server, server cluster, or other device that can execute acquisition of a remote sensing clear sky dataset.

[0030] In a possible implementation, remote sensing image data collected by a remote sensing device on a satellite payload within a predetermined time period may be acquired. The predetermined time period is a continuous time period, such as one month.

[0031] In a possible implementation, the remote sensing image data may be preprocessed, and different remote sensing image data may be preprocessed differently, as long as a global remote sensing image time series sequence within a predetermined time period can be obtained, and the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order. For example, the preprocessing may include correction processing, projection processing, and stitching processing. The correction processing is mainly to restore the true surface reflectivity of each pixel in the remote sensing image data. The projection processing is mainly to project the remote sensing image data into the same map coordinate system for subsequent stitching; the stitching processing is mainly to stitch together the remote sensing image data of the same time period to form a complete global image. Of course, in other implementations, some remote sensing image data may not need correction processing. In this case, the preprocessing may include projection processing and stitching processing.

[0032] In a possible implementation, a target area can be selected globally as needed, and for each global remote sensing image in a global remote sensing image time series sequence, a target remote sensing image corresponding to the target area in the global remote sensing image can be selected. These target remote sensing images can be arranged in time to form a target remote sensing image time series sequence.

[0033] In a possible implementation, in order to obtain the clear sky pixels in these target remote sensing images, it is necessary to extract features from the target remote sensing images that can characterize whether the pixel is a clear sky pixel or a non-clear sky pixel. The feature can be the reflectance spectral feature of each pixel in the target band. Using reflectance spectral feature analysis to distinguish clouds from other objects is a method based on the reflectance difference of objects in different spectral bands. Reflectance spectral features refer to the reflectance variation characteristics of objects in different wavelength ranges. These characteristics are mainly determined by the physical, chemical and biological properties of the objects, and are affected by multiple factors such as the wavelength of the incident light, the angle of incidence, the observation angle and atmospheric conditions. Therefore, a small number of target bands that can best distinguish the clear sky characteristics and non-clear sky characteristics (such as cloud characteristics) of the pixel can be selected, and the reflectance spectral features of the target bands can be used for subsequent cluster analysis. The reflectance spectral features include features such as reflectance and / or brightness temperature.

[0034] In a possible implementation, the information may also include the NDVI (Normalized Difference Vegetation Index) of each pixel. The NDVI is a commonly used vegetation index used to evaluate and quantify vegetation conditions in remote sensing images. The calculation formula is: NDVI=(NIR-Red) / (NIR+Red), where NIR represents the reflectance value of the near-infrared band and Red represents the reflectance value of the visible red band. This index estimates the lushness of vegetation by comparing the reflectance difference between the near-infrared band and the red light band. The NDVI can be used to indicate non-vegetated areas and may indirectly reflect cloud information.

[0035] In a possible implementation, the K-means algorithm is an unsupervised learning method in machine learning, and is a distance-based clustering algorithm. Its core idea is to divide the data into K clusters in an iterative manner, so that the data points in each cluster are as close to the cluster center as possible, and the data points between different clusters are as far away as possible. The K-means algorithm used in the present disclosure is mainly to iteratively cluster multiple pixels in the same target position but different time series in the time series of the target remote sensing image, and divide these pixels into different cluster clusters based on the reflectance spectral characteristics and NDVI of these pixels in several target bands. When clustering, the K value can be determined according to the research purpose and the characteristics of the target remote sensing image. If only clear sky and non-clear sky pixels need to be distinguished, K=2 can be set. If further subdivision of the regional category is required, a larger K value needs to be set. For multiple pixels at the same target position in the time series of the target remote sensing image, the distance from each pixel to the K cluster centers can be calculated, and it can be assigned to the cluster corresponding to the cluster center closest to it. For each cluster, its cluster center is recalculated (usually the mean of all pixels in the cluster). Repeat the above steps until a termination condition is met (such as the change in the cluster center is less than the specified threshold, the maximum number of iterations is reached, etc.). After the iteration is completed, K clusters can be obtained.

[0036] In a possible implementation, for the obtained K clusters, the type of cluster, such as cloud cluster or clear sky cluster, can be identified by analyzing the reflectance spectral characteristics of the cluster center in the cluster or analyzing the reflectance spectral characteristics of each pixel in the cluster.

[0037] In one possible implementation, the pixels in the clear sky cluster are all clear sky pixels at the target location, so a pixel can be randomly selected as the clear sky pixel at the target location. Of course, in order to obtain better clear sky data, the pixel with the most significant clear sky feature can be selected from the cluster as the clear sky pixel at the target location.

[0038] In a possible implementation, the above scheme can be used to obtain clear sky pixels at each target position in the target remote sensing image, and the clear sky pixels at each target position can be synthesized together to obtain a clear sky image of the target area.

[0039] This implementation can construct a high-quality global remote sensing image on a global scale by preprocessing remote sensing image data collected within a predetermined time period, and then select a target remote sensing image time series sequence of the target area from the global remote sensing image time series sequence; extract the reflectance spectral characteristics of each pixel in the target band and the NDVI of each pixel from the target remote sensing image time series sequence; then, for multiple pixels at the same target position but different time series in the target remote sensing image time series sequence, use the K-means algorithm to cluster the multiple pixels to obtain multiple cluster clusters; identify clear sky cluster clusters in the cluster clusters; select clear sky pixels at the target position from the clear sky cluster clusters; and synthesize a clear sky image of the target area based on the clear sky pixels at each target position in the target area, so as to obtain a high-quality clear sky image. This scheme adopts the K-means algorithm, which only requires the reflectance spectral characteristics of several target bands and the NDVI of each pixel to cluster the clear sky pixels at the target location. The clear sky pixels at each target location in the target area can be synthesized into a clear sky image of the target area. Less data is required to identify clear sky pixels and it is not affected by other external factors. The analysis scheme is simple and easy to implement, and it maintains high efficiency and stability when processing large-scale remote sensing image data.

[0040] In a possible implementation, the remote sensing image data is L1 data of the MERSI payload of the Fengyun-3 satellite; the preprocessing of the remote sensing image data set to obtain a global remote sensing image time series within a predetermined time period includes: Perform atmospheric correction, equal longitude and latitude projection and data splicing on the remote sensing image data collected every day to obtain the global remote sensing image every day; The global remote sensing images of each day within the predetermined time period are sorted to form a global remote sensing image time series sequence.

[0041] In this embodiment, the remote sensing image data is the L1 data of the MERSI (Medium Resolution Spectral Imager) payload of the Fengyun-3 satellite, which records multi-band and multi-spectral remote sensing data, providing a rich data source for the acquisition of clear sky data disclosed in the present invention. These data sources not only contain basic information such as surface reflectivity and brightness temperature, but also contain rich spectral features, providing an important basis for the acquisition of clear sky data.

[0042] In this embodiment, for the remote sensing image data of the MERSI payload of the Fengyun-3 satellite, it is necessary to perform atmospheric correction processing, equal longitude and latitude projection processing and data splicing processing on these remote sensing image data to obtain daily global remote sensing images. The atmospheric correction processing is mainly to eliminate the influence of the atmosphere on radiation transmission and restore the true reflectivity of the surface; equal longitude and latitude projection is an azimuthal projection, in which the meridians are projected as straight lines originating from the poles in the polar azimuth, and the latitudes are displayed as equally spaced concentric arcs; the equal longitude and latitude projection technology can be used to project the remote sensing image data into the same longitude and latitude coordinate system. Data splicing processing refers to splicing the remote sensing image data of the same time period together according to the longitude and latitude coordinates to form a complete global remote sensing image of the global scope. After the splicing is completed, the global remote sensing images within the predetermined time period can be sorted in chronological order to form a global remote sensing image time series sequence.

[0043] In a possible implementation, the atmospheric correction process includes: Acquire daytime remote sensing image data from the remote sensing image data; Based on the daytime remote sensing image data, the original reflectivity and angle information of each pixel in the daytime remote sensing image data in each target band are obtained, wherein the angle information includes the solar altitude angle, the solar zenith angle, the satellite altitude angle, the satellite zenith angle, and the elevation data; When the solar zenith angle satisfies the correction condition, the original reflectivity of each pixel in the daytime remote sensing image data is corrected according to the angle information to obtain the remote sensing image data after atmospheric correction.

[0044] In this embodiment, when performing atmospheric correction, it is first determined whether the remote sensing image data is daytime remote sensing image data. If not, it is directly discarded. If so, for each daytime remote sensing image data, the original reflectivity and angle information of each pixel in the daytime remote sensing image data in each target band are obtained, and then for each pixel, it is determined whether the solar zenith angle of the pixel meets the correction conditions. If not, the remote sensing image data where the pixel is located is directly discarded. If it meets the conditions, the original reflectivity of the pixel is corrected according to the angle information of the pixel to obtain the true surface reflectivity of the pixel. The specific atmospheric correction calculation is clearly understood by those skilled in the art and will not be described in detail here.

[0045] In a possible implementation, the equal longitude and latitude projection processing includes: Perform equal longitude and latitude projection on the remote sensing image data after atmospheric correction to obtain initial map projection data of a predetermined resolution; The initial map projection data is interpolated to obtain map projection data.

[0046] In this embodiment, for the atmospherically corrected remote sensing image data of each day, 5 minutes of atmospherically corrected remote sensing image data can be obtained, and then the atmospherically corrected remote sensing image data is projected with equal longitude and latitude to obtain initial map projection data of a predetermined resolution. Generally, the predetermined resolution is 1 km.

[0047] In this embodiment, some pixels in the initial map projection data obtained after projection may be blank. In this case, an interpolation algorithm can be used to interpolate the initial map projection data to obtain map projection data. For example, 1 km of initial map projection data can be read. For this 1 km of initial map projection data, block interpolation calculation can be performed based on the target band and satellite zenith angle. Interpolation can be performed along the X / Y axis first, and then the data corner points can be interpolated. In this way, interpolation is completed to obtain map projection data.

[0048] In a possible implementation, the data splicing process includes: Acquire the target band of the map projection data; For each target band, obtaining the equal longitude and latitude information of the map projection data of the target band; Based on the equal longitude and latitude information of the map projection data of the target band, multiple blocks of map projection data of the target band are synthesized into block data of a predetermined longitude and latitude span; Based on the satellite zenith angle of the block data of the predetermined longitude and latitude span, the block data of the predetermined longitude and latitude span are synthesized into whole block data, and then the whole block data are spliced ​​to obtain the global remote sensing image of the target band.

[0049] In this embodiment, when data splicing is performed, for each day's map projection data, the target band of the map projection data can be first determined, and the map projection data of the same target band can be spliced. For each target band, the equal longitude and latitude information (such as longitude and latitude information, resolution information, etc.) of the map projection data of the target band is obtained, and then, based on the equal longitude and latitude information and satellite zenith angle of the map projection data of the target band, multiple blocks of map projection data of the target band are first synthesized into multiple blocks of data with a predetermined longitude and latitude span (such as 10 degrees or 11.25 degrees), and these blocks of data are located at different positions of the predetermined longitude and latitude span. These blocks of data of the predetermined longitude and latitude span can be synthesized into a whole block of data of the predetermined longitude and latitude span, and then all the whole blocks of data are seamlessly spliced ​​to obtain the global remote sensing image of the target band, so that the global remote sensing image of each target band can be obtained.

[0050] In a possible implementation manner, the target wavebands include four wavebands, namely a blue light waveband, a green light waveband, a red light waveband, and a near infrared waveband.

[0051] In this embodiment, clouds generally have a high reflectivity in the visible light band (such as 0.4-0.76μm) because they reflect most of the incident visible light. In the infrared band, especially the short-wave infrared and medium-wave infrared (such as 1-3μm), the water vapor and ice crystals inside the cloud absorb and scatter the light of these wavelengths, resulting in a significant decrease in the reflectivity of the cloud. The spectral curve of the cloud is generally flatter and has a higher reflectivity in the visible light band; while in the infrared band, the reflectivity drops sharply. This difference gives the cloud a unique shape on the spectral curve graph, making it easy to identify. The reflectivity spectral characteristics of vegetation show obvious double peaks and double valleys in the visible light band. There is a reflection peak in the green light band (about 0.55μm), while there are two absorption bands in the blue light band (about 0.45μm) and the red light band (about 0.67μm). In the near-infrared band (such as 0.7-1.1μm), the reflectivity of vegetation rises sharply, forming the so-called "red edge" feature. This is a unique spectral feature of vegetation and can be used to distinguish vegetation from other ground objects. After research, this paper uses the reflectance spectral features of the blue light band, green light band, red light band, near-infrared band and NDVI in the L1 data as feature data for identifying clear skies.

[0052] Preferably, the target band can be band1 (0.47μm), band2 (0.55μm), band3 (0.65μm), and band4 (0.865μm). Of course, the value of the target band can fluctuate by 0.01μm around the values ​​of the above four bands, and there is no limitation here.

[0053] In a possible implementation, clustering the multiple pixels using a K-means algorithm to obtain multiple clusters includes: Using K-means algorithm to cluster the multiple pixels to obtain two clusters; identifying a suspected clear sky cluster among the two clusters; Using the K-means algorithm to cluster the pixels in the suspected clear sky cluster again, obtaining three clusters; A clear sky cluster is identified among the three clusters.

[0054] In this embodiment, the present disclosure can perform two clusterings when clustering the multiple pixels using the K-means algorithm. In the first clustering, the K value is 2, and two initial cluster centers are first selected. Then, the multiple pixels are iteratively assigned to the nearest cluster center according to their reflectance spectral characteristics and NDVI, and the cluster center is updated to the mean of the reflectance spectral characteristics and NDVI of the pixels in each category after each iteration. This process is repeated until the cluster center is stable or the preset number of iterations is reached. After the clustering is completed, by analyzing the reflectance spectral characteristics of each cluster center, one of the two cluster clusters is identified as a cloud cluster cluster, and the other cluster cluster is identified as a suspected clear sky cluster cluster. Then, the pixels in the suspected clear sky cluster cluster with lower reflectance are clustered for the second time, and the K value is 3 to obtain three cluster clusters, and the clear sky cluster cluster among the three cluster clusters can be identified. For example, the mean of the NDVI of each pixel in each cluster cluster can be calculated, and the cluster cluster with the largest mean value is taken as the clear sky cluster cluster.

[0055] This implementation method can more accurately determine clear sky pixels through two clustering operations, thereby obtaining a high-quality clear sky image.

[0056] In a possible implementation manner, selecting the clear sky pixel of the target location from the clear sky cluster includes: A pixel with the largest NDVI is selected from the clear sky cluster as the clear sky pixel of the target location.

[0057] In this implementation, the pixels in the clear sky cluster are all pixels at the same target location, and the pixel with the largest NDVI in the clear sky cluster is the pixel with the most significant clear sky feature at the target location, and this pixel can be used as the clear sky pixel of the target location.

[0058] In order to prove the effectiveness and reliability of the disclosed scheme, the following verification can be performed: Figure 2 A schematic diagram showing the reflectance spectral characteristics and NDVI of pixels at the same target location extracted from a time series of target remote sensing images within a predetermined time period provided by an embodiment of the present disclosure in a target band, Figure 2The reflectance and NDVI values ​​of four different target bands (band1 (0.47μm), band2 (0.55μm), band3 (0.65μm), band4 (0.865μm)) of a target position in a target remote sensing image every day from January 1 to January 31, 2023, that is, the reflectance and NDVI of 30 pixels in four different target bands. In order to distinguish different weather or cloud cover conditions from these data, the k-means clustering algorithm is used to cluster these 30 pixels twice, and these 30 pixels are roughly divided into four categories: cloud, thin cloud, suspected cloud, and clear sky. Figure 2 As shown, label=1 can be identified as cloud, label=3 as thin cloud, label=4 as clear sky, label=5 as suspected cloud, and finally the pixel of the second day with the largest NDVI in the cluster of label=4, i.e. the clear sky cluster, is selected as the clear sky pixel of the target location. In order to further verify its accuracy, the visual interpretation method is adopted, and the experts can directly identify and interpret the remote sensing images. By checking the classification results of the pixels at the target location day by day and comparing them with the actual true color images, it is found that the classification results of the k-means method are basically consistent with the visual interpretation results, which shows the effectiveness and reliability of the k-means algorithm in the present disclosure.

[0059] The present disclosure also provides a device for acquiring a remote sensing clear sky data set, Figure 3 The structure block diagram of the device for acquiring the remote sensing clear sky data set according to the embodiment of the present disclosure is shown. The device can be implemented as part or all of the electronic device through software, hardware or a combination of both. Figure 3 As shown, the device for acquiring the remote sensing clear sky data set includes:

[0060] An acquisition module 301 is configured to acquire remote sensing image data collected within a predetermined time period; A preprocessing module 302 is configured to preprocess the remote sensing image data to obtain a global remote sensing image time series sequence within a predetermined time period, wherein the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order; The region selection module 303 is configured to select a target remote sensing image time series sequence of a target region from the global remote sensing image time series sequence; The feature extraction module 304 is configured to extract the reflectance spectral feature of each pixel in the target band and the normalized difference vegetation index NDVI of each pixel from the target remote sensing image time series; The clustering module 305 is configured to cluster multiple pixels at the same target position but in different time series in the time series of the target remote sensing image using a K-means algorithm to obtain multiple clusters; An identification module 306 is configured to identify a clear sky cluster among the clusters; A pixel selection module 307 is configured to select a clear sky pixel of the target location from the clear sky cluster; The image synthesis module 308 is configured to synthesize a clear sky image of the target area based on the clear sky pixels at each target position in the target area.

[0061] In a possible implementation, the remote sensing image data is L1 level data of the MERSI payload of the Fengyun-3 satellite; the preprocessing module is configured as follows: Perform atmospheric correction, equal longitude and latitude projection and data splicing on the collected remote sensing image data to obtain multiple global remote sensing images; The plurality of global remote sensing images within the predetermined time period are sorted in chronological order to form a global remote sensing image time series sequence.

[0062] In a possible implementation, the atmospheric correction process in the preprocessing module is configured as follows: Acquire daytime remote sensing image data from the remote sensing image data; Based on the daytime remote sensing image data, the original reflectivity and angle information of each pixel in the daytime remote sensing image data in each target band are obtained, wherein the angle information includes the solar altitude angle, the solar zenith angle, the satellite altitude angle, the satellite zenith angle, and the elevation data; When the solar zenith angle satisfies the correction condition, the original reflectivity of each pixel in the daytime remote sensing image data is corrected according to the angle information to obtain the remote sensing image data after atmospheric correction.

[0063] In a possible implementation, the equal longitude and latitude projection processing in the preprocessing module is configured as follows: Perform equal longitude and latitude projection on the remote sensing image data after atmospheric correction to obtain initial map projection data of a predetermined resolution; The initial map projection data is interpolated to obtain map projection data.

[0064] In a possible implementation, the data splicing process in the preprocessing module is configured as follows: Acquire the target band of the map projection data; For each target band, obtaining the equal longitude and latitude information of the map projection data of the target band; Combining the multiple blocks of map projection data of the target band into block data of a predetermined longitude and latitude span; Based on the satellite zenith angle of the block data of the predetermined longitude and latitude span, the block data of the predetermined longitude and latitude span are synthesized into whole block data, and then the whole block data are spliced ​​to obtain the global remote sensing image of the target band.

[0065] In a possible implementation, the target waveband includes a blue light waveband, a green light waveband, a red light waveband, and a near infrared waveband.

[0066] In a possible implementation, the clustering module is configured as follows: Using K-means algorithm to cluster the multiple pixels to obtain two clusters; identifying a suspected clear sky cluster among the two clusters; Using the K-means algorithm to cluster the pixels in the suspected clear sky cluster again, obtaining three clusters; A clear sky cluster is identified among the three clusters.

[0067] In a possible implementation manner, the pixel selection module is configured as follows: A pixel with the largest NDVI is selected from the clear sky cluster as the clear sky pixel of the target location.

[0068] The technical terms and technical features mentioned in the implementation manner of the present device are the same as or similar to those mentioned in the implementation manner of the above method. For the explanation and description of the technical terms and technical features involved in the present device, reference may be made to the explanation and description of the implementation manner of the above method, and they will not be repeated here.

[0069] The present disclosure also discloses an electronic device. FIG4 shows a structural block diagram of the electronic device according to an embodiment of the present disclosure.

[0070] As shown in FIG. 4 , the electronic device 400 includes a memory 401 and a processor 402 , wherein the memory 401 is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor 402 to implement the method according to an embodiment of the present disclosure.

[0071] FIG5 is a schematic diagram showing the structure of a computer system suitable for implementing the method of the embodiment of the present disclosure.

[0072] As shown in FIG5 , the computer system 500 includes a processing unit 501, which can perform various processes in the above-mentioned embodiments according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the computer system 500 are also stored. The processing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0073] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage section 508 as needed. Among them, the processing unit 501 can be implemented as a processing unit such as a CPU, a GPU, a TPU, an FPGA, an NPU, etc.

[0074] In particular, according to an embodiment of the present disclosure, the method described above can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes computer instructions, and the computer instructions are executed by a processor to implement the method steps described above. In such an embodiment, the computer program product can be downloaded and installed from a network through the communication part 509, and / or installed from a removable medium 511.

[0075] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0076] The units or modules involved in the embodiments described in the present disclosure may be implemented by software or programmable hardware. The units or modules described may also be set in a processor, and the names of these units or modules do not constitute limitations on the units or modules themselves in some cases.

[0077] As another aspect, the present disclosure further provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system in the above embodiment; or a computer-readable storage medium that exists independently and is not assembled into a device. The computer-readable storage medium stores one or more programs, and the programs are used by one or more processors to execute the method described in the present disclosure.

[0078] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present disclosure is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other.

Claims

1. A method for acquiring a remote sensing clear sky dataset, characterized in that: include: Acquiring remote sensing image data collected within a predetermined time period; Preprocessing the remote sensing image data to obtain a global remote sensing image time series sequence within a predetermined time period, wherein the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order; Selecting a target remote sensing image time series sequence of a target area from the global remote sensing image time series sequence; Extracting the reflectance spectral characteristics of each pixel in the target band and the normalized difference vegetation index NDVI of each pixel from the target remote sensing image time series; For a plurality of pixels at the same target position but with different time series in the time series of the target remote sensing image, clustering the plurality of pixels using a K-means algorithm to obtain a plurality of cluster clusters; identifying a clear sky cluster among the clusters; Selecting a clear sky pixel at the target location from the clear sky cluster; A clear sky image of the target area is synthesized based on clear sky pixels at each target position in the target area.

2. The acquisition method according to claim 1, characterized in that: The remote sensing image data is L1 level data of the MERSI payload of the Fengyun-3 satellite; the preprocessing of the remote sensing image data set to obtain a global remote sensing image time series within a predetermined time period includes: Perform atmospheric correction, equal longitude and latitude projection and data splicing on the collected remote sensing image data to obtain multiple global remote sensing images; The plurality of global remote sensing images within the predetermined time period are sorted in chronological order to form a global remote sensing image time series sequence.

3. The method according to claim 2, characterized in that The atmospheric correction process includes: Acquire daytime remote sensing image data from the remote sensing image data; Based on the daytime remote sensing image data, the original reflectivity and angle information of each pixel in the daytime remote sensing image data in each target band are obtained, wherein the angle information includes the solar altitude angle, the solar zenith angle, the satellite altitude angle, the satellite zenith angle, and the elevation data; When the solar zenith angle satisfies the correction condition, the original reflectivity of each pixel in the daytime remote sensing image data is corrected according to the angle information to obtain the remote sensing image data after atmospheric correction.

4. The method according to claim 2, characterized in that: The equal longitude and latitude projection processing includes: Perform equal longitude and latitude projection on the remote sensing image data after atmospheric correction to obtain initial map projection data of a predetermined resolution; The initial map projection data is interpolated to obtain map projection data.

5. The method according to claim 4, characterized in that The data splicing process includes: Acquire the target band of the map projection data; For each target band, obtaining the equal longitude and latitude information of the map projection data of the target band; Based on the equal longitude and latitude information of the map projection data of the target band, multiple blocks of map projection data of the target band are synthesized into block data of a predetermined longitude and latitude span; Based on the satellite zenith angle of the block data of the predetermined longitude and latitude span, the block data of the predetermined longitude and latitude span are synthesized into whole block data, and then the whole block data are spliced ​​to obtain the global remote sensing image of the target band.

6. The method according to any one of claims 2 to 5, characterized in that: The target wavebands include a blue light waveband, a green light waveband, a red light waveband and a near infrared waveband.

7. The method according to claim 1, characterized in that The K-means algorithm is used to cluster the multiple pixels to obtain multiple clusters, including: Using K-means algorithm to cluster the multiple pixels to obtain two clusters; identifying a suspected clear sky cluster among the two clusters; Using the K-means algorithm to cluster the pixels in the suspected clear sky cluster again, obtaining three clusters; A clear sky cluster is identified among the three clusters.

8. The method according to claim 1, characterized in that: The selecting the clear sky pixel of the target location from the clear sky cluster comprises: A pixel with the largest NDVI is selected from the clear sky cluster as the clear sky pixel of the target location.

9. A device for acquiring a remote sensing clear sky data set, characterized in that: include: An acquisition module is configured to acquire remote sensing image data collected within a predetermined time period; A preprocessing module is configured to preprocess the remote sensing image data to obtain a global remote sensing image time series sequence within a predetermined time period, wherein the global remote sensing image time series sequence is a sequence of global remote sensing images arranged in chronological order; A region selection module is configured to select a target remote sensing image time series sequence of a target region from a global remote sensing image time series sequence; A feature extraction module is configured to extract the reflectance spectral characteristics of each pixel in the target band and the normalized difference vegetation index NDVI of each pixel from the target remote sensing image time series; A clustering module is configured to cluster multiple pixels at the same target position but in different time series in the time series of the target remote sensing image using a K-means algorithm to obtain multiple clusters; an identification module, configured to identify a clear sky cluster among the clusters; A pixel selection module is configured to select a clear sky pixel of the target location from the clear sky cluster; The image synthesis module is configured to synthesize the clear sky image of the target area based on the clear sky pixels at each target position in the target area.

10. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method according to any one of claims 1 to 8.

11. A readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed by a processor, the method steps described in any one of claims 1 to 8 are implemented.

Images