A sea ice mapping method

Abstract

The present invention relates to a sea ice mapping method, which belongs to the field of Arctic sea ice mapping and monitoring technology, and solves the problem of low sea ice detection accuracy in existing ice-water mixed areas and ice or water areas covered by clouds. The method is applied to the Arctic, and includes the following steps: obtaining multiple MODIS L1B image data collected daily in the study area and preprocessing; classifying each preprocessed image data to obtain a classification result matrix for each image data; the classification categories include sea ice, water and cloud; each pixel in the classification result matrix stores the category label corresponding to the pixel; obtaining the total number of non-cloud categories and the mode of non-cloud categories of pixels corresponding to the same position in multiple classification result matrices, and obtaining a daily water and ice distribution map based on the total number of non-cloud categories and the mode of non-cloud categories; fusing the daily water and ice distribution maps to obtain a daily synthetic ice map.

Description

Technical Field

[0001] The present invention relates to the field of Arctic sea ice mapping and monitoring technology, and in particular to a sea ice mapping method. Background Art

[0002] As an important part of the cryosphere, sea ice has a profound impact on the study of global climate change, biodiversity, etc. In particular, with the global climate change, polar warming and the opening of the Arctic route, the demand for sea ice information has increased rapidly.

[0003] Traditional sea ice products mostly rely on passive microwave radiometer data, which is not affected by clouds and can form daily global coverage observations, but its spatial resolution is low, mostly data at the level of more than ten kilometers, and it is difficult to use these data for detailed analysis and guidance of shipping. SAR data images have the characteristics of all-weather and high spatial resolution, and can provide high-quality sea ice information, but the revisit period is long and the coverage is small, making it difficult to provide large-scale, high-temporal resolution sea ice products. Although optical satellite sensors are affected by cloud cover and light, their high temporal resolution and spatial resolution can still obtain rich sea ice information. For example, the MODIS sensors of the US TERRA and AQUA satellites can make multiple observations in polar regions every day, and the spatial resolution of the data obtained is much higher than that of passive microwave data. By comprehensively analyzing multi-phase data, all or part of the polar sea ice information can be obtained. The MOD29 single-view sea ice product provided by NASA based on MODIS data can help users obtain polar sea ice conditions to a certain extent, but this product often misjudges ice when detecting ice in ice-water mixed areas and ice or water areas covered by clouds. This is mainly because the cloud mask data used in the sea ice product provided by NASA is the MOD35 product, which is a global cloud product with poor recognition accuracy in polar regions, especially in ice areas covered by clouds, thin clouds or fog. This product tends to underestimate the amount of clouds over sea ice and overestimate the amount of clouds over open water.

[0004] Therefore, how to improve the accuracy of sea ice detection in ice-water mixed areas and ice or water areas covered by clouds is an issue that needs to be addressed urgently. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a sea ice mapping method to solve the problem of low sea ice detection accuracy in existing ice-water mixed areas and ice or water areas covered by clouds.

[0006] The present invention provides a sea ice mapping method, which is applied to the Arctic and comprises the following steps:

[0007] Obtain and preprocess multiple MODIS L1B image data collected daily in the study area;

[0008] Classifying each preprocessed image data to obtain a classification result matrix for each image data; the classification categories include sea ice, water and cloud; each pixel in the classification result matrix stores a category label corresponding to the pixel;

[0009] Obtaining the total number of non-cloud categories and the mode of non-cloud categories of pixels corresponding to the same position in multiple classification result matrices, and obtaining a daily water and ice distribution map based on the total number of non-cloud categories and the mode of non-cloud categories;

[0010] The daily water and ice distribution maps are integrated to obtain a daily synthetic ice map.

[0011] On the basis of the above scheme, the present invention also makes the following improvements:

[0012] Further, the daily water distribution map is obtained by performing the following operations:

[0013] For each pixel in the classification result matrix, if the total number of non-cloud categories is greater than a first threshold, the value of the pixel is the mode of the non-cloud categories;

[0014] Pixels with the category label of water are extracted, and other pixels are set to empty to obtain the daily water distribution map.

[0015] Further, the daily ice distribution map is obtained by performing the following operations:

[0016] For each pixel in the classification result matrix, if the total number of non-cloud categories is greater than a second threshold, the value of the pixel is the mode of the non-cloud categories;

[0017] Pixels with the category label of sea ice are extracted, and other pixels are left blank to obtain the daily ice distribution map;

[0018] The second threshold is greater than the first threshold.

[0019] Further, the fusion of the daily water and ice distribution maps to obtain a daily synthetic ice map includes:

[0020] The values ​​of the pixels at each corresponding position in the daily water and ice distribution maps are compared respectively. If they are consistent, they remain unchanged; if they are inconsistent, the value of the pixel is the non-empty category mode of the neighborhood set for the current pixel in the daily water and ice distribution maps; the values ​​of the remaining pixels are the category labels corresponding to the clouds, thereby forming a daily synthetic ice map.

[0021] Furthermore, the method further comprises: obtaining a weekly synthetic ice map based on the sea ice map for seven consecutive days:

[0022] Obtain the total number and mode of non-cloud categories of pixels corresponding to the same position in the ice map for seven consecutive days;

[0023] If the total number of non-cloud categories of the current pixel is not 0, its value is the mode of the non-cloud category of the current pixel. If the number of sea ice and water in the non-cloud category is equal, the value of the current pixel is the category label corresponding to water;

[0024] If the total number of non-cloud categories of the current pixel is 0, its value is the category label corresponding to the cloud;

[0025] This forms a weekly composite ice map.

[0026] Further, the classification result matrix of each image data is obtained by performing the following operations:

[0027] The characteristic bands corresponding to each image data after preprocessing are extracted, and the verified classification model is classified based on the extracted characteristic bands to obtain the classification result matrix of each image data.

[0028] Further, the verified classification model is obtained by performing the following operations:

[0029] Obtain MODIS L1B historical image data of multiple time periods in the study area and preprocess them in batches;

[0030] Extracting a plurality of sample data corresponding to each classification category from the batch preprocessed historical image data, and obtaining a category label corresponding to each sample data;

[0031] Based on the sample data and its corresponding category labels, a priori pixel sample library is constructed;

[0032] A training sample and a verification sample are selected from the prior pixel sample library, a classification model is trained based on the characteristic band of each sample data in the training sample and its corresponding category label, and the classification accuracy of the trained classification model is verified based on the characteristic band of each sample data in the verification sample and its corresponding category label. If the classification accuracy exceeds the classification accuracy threshold, the verification is passed, thereby obtaining a verified classification model.

[0033] Furthermore, the characteristic bands include: data corresponding to MODIS reflectivity bands 7, 2, and 1; NDSI, and MRELBP.

[0034] Furthermore, the classification category corresponding to the cloud further includes blue cloud, white cloud and red cloud subcategories; wherein each subcategory corresponds to a corresponding subcategory label.

[0035] Furthermore, the preprocessing includes geometric correction, radiation calibration, solar zenith angle correction and land mask.

[0036] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0037] The present invention provides a sea ice mapping method, which has the following beneficial effects:

[0038] First, on the basis of realizing automatic classification of images in the study area, the present invention uses the cloud motion characteristics in the Arctic region and the multi-temporal determination method of ground objects to merge the classification results of multiple images obtained every day, and synthesizes the daily ice map on this basis, thereby effectively solving the problem of easy confusion between ice and clouds;

[0039] Second, the present invention also forms a weekly synthetic ice map based on the daily synthetic ice map;

[0040] Third, in response to the low spatial resolution of traditional sea ice products, the present invention constructs a priori pixel sample library for polar feature classification based on the constructed multi-temporal remote sensing images of the Arctic, and uses a random forest algorithm with multi-feature band fusion to automatically classify the images of the study area. The classification results have greatly improved the spatial resolution and accuracy of traditional MODIS sea ice products. At the same time, the present invention also improves the feature bands selected in the classification process and the classification categories, so that the classification accuracy is effectively improved.

[0041] Other characteristic bands and advantages of the present invention will be described in the subsequent description, and some advantages may become apparent from the description, or may be understood by practicing the present invention. The purpose and other advantages of the present invention may be realized and obtained through the contents particularly pointed out in the description and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0043] Figure 1 A flow chart of the sea ice mapping method provided in Example 1 of the present invention;

[0044] Figure 2 A schematic diagram of the synthesis process of the daily synthetic ice map provided in Example 1 of the present invention;

[0045] Figure 3 (a) and (b) are both false color images of the B721 band combination provided by Example 2 of the present invention, where: Figure 3 (a) and (b) were collected on July 3, 2018 and August 3, 2019, respectively;

[0046] Figure 3(c) and (d) are respectively provided in Example 2 of the present invention Figure 3 (a), (b) Classification results of images using MFLFRF algorithm;

[0047] Figure 3 (e) and (f) are respectively provided in Example 2 of the present invention. Figure 3 (a) and (b) correspond to MOD29 sea ice products.

[0048] Figure 4 (a) is a daily synthetic ice map obtained based on Terra on August 1, 2019, provided in Example 2 of the present invention;

[0049] Figure 4 (b) Weekly composite ice map from August 1 to 7, 2019, obtained from the Terra and Aqua stars. DETAILED DESCRIPTION

[0050] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0051] Example 1

[0052] A specific embodiment of the present invention discloses a method for sea ice mapping, the flow chart of which is as follows: Figure 1 As shown, the method is applied to the Arctic and includes the following steps:

[0053] Step S1: Obtain and pre-process multiple MODIS L1B (L1B data is generated after decompression, positioning and calibration of MODIS raw data, and is MODIS level 1 product data) image data collected daily in the study area;

[0054] Preferably, the preprocessing includes geometric correction, radiation calibration, solar zenith angle correction and land mask, etc.; it should be emphasized that by performing land mask processing, the preprocessed image data only contains ice, water and clouds; therefore, the categories classified in step S2 include sea ice, water and clouds.

[0055] In this embodiment, the land mask processing process is: extract the land mask of the study area based on the MODIS land cover type annual global 500-meter product (MCD12Q1 product); in addition, the processes of geometric correction, radiation calibration, and solar zenith angle correction can all be achieved using existing methods and will not be repeated here.

[0056] Step S2: classifying each preprocessed image data to obtain a classification result matrix for each image data; the classification categories include sea ice, water and cloud; each pixel in the classification result matrix stores a category label corresponding to the pixel;

[0057] Preferably, the specific process of this step is: extracting the characteristic band corresponding to each image data after preprocessing, and classifying the verified classification model based on the extracted characteristic band to obtain the classification result matrix of each image data;

[0058] Taking into account the differences between sea ice and clouds in spectral and texture characteristic bands, the characteristic bands selected in this embodiment include: data corresponding to MODIS reflectivity bands 7, 2, and 1 (i.e., short-wave infrared band, near-infrared band, and red band of MODIS), and the data of these three bands can be directly obtained based on image data; NDSI (Normalized Difference Snow and Ice Index), and MRELBP (Median Robust Extended Local Binary Pattern), NDSI and MRELBP data need to be obtained by processing image data, and the processing process can be obtained by using the existing method, which will not be repeated here.

[0059] Before executing step S2, it is necessary to obtain a verified classification model in advance. The specific process is as follows:

[0060] (1) Obtain MODIS L1B historical image data of multiple time stages in the study area and preprocess them in batches; preferably, in the Arctic Ocean, the types of ground objects are relatively simple, mainly including sea ice, water and clouds; the features of ground objects are relatively fixed and are relatively less affected by factors such as seasons, which is suitable for identifying ground objects by sample migration and other methods. Therefore, this embodiment selects multiple images (more than 100 images) of different time stages such as the ice formation period and the ice melting period for batch preprocessing, and selects sample data from the preprocessed images.

[0061] (2) extracting a plurality of sample data corresponding to each classification category from the batch preprocessed historical image data, and obtaining a category label corresponding to each sample data;

[0062] (3) Based on the sample data and its corresponding category labels, a priori pixel sample library is constructed;

[0063] (4) Selecting training samples and verification samples from the prior pixel sample library, training a classification model based on the characteristic band of each sample data in the training samples and its corresponding category label, and verifying the classification accuracy of the trained classification model based on the characteristic band of each sample data in the verification samples and its corresponding category label. If the classification accuracy exceeds the classification accuracy threshold, the verification is passed, thereby obtaining a verified classification model.

[0064] Considering that the colors of clouds in the study area are blue, white, and red, etc., in order to better achieve classification, this embodiment further divides the classification categories corresponding to clouds into three subcategories: blue clouds, white clouds, and red clouds according to the color characteristics of clouds; wherein each subcategory corresponds to a corresponding subcategory label. Accordingly, the selected training samples and verification samples should also include sample data corresponding to each subcategory.

[0065] In the process of selecting training samples and verification samples, according to the principle of uniform distribution of samples, training samples and verification samples are selected from multiple images acquired at different times in the study area. For example, 80% of the samples in the constructed prior pixel sample library can be used as training samples, and the remaining 20% ​​of the samples can be used as verification samples, and the random forest (MFLFRF) algorithm can be used for model training. The number of decision trees ntree contained in the random forest is set to 400-500, and the number of variables at each level is set to 4-6.

[0066] After the above process, the classification result matrix of each image data can be obtained.

[0067] Considering that there may be misclassification in the classification process, for example, a small number of blue clouds and cloud shadows are mistakenly classified as sea ice. However, compared with sea ice, the clouds and cloud shadows covering it are moving, and in the same area of ​​the Arctic, multiple MODIS images can be obtained in one day (generally at least 6 observations, most of the time 9 observations, and at least 4 observations in special cases in the Arctic edge area). Therefore, the classification result matrix corresponding to the multiple image data observed on the same day can be further processed to reduce the impact of misclassification. The specific process is described in steps S3 and S4.

[0068] Step S3: obtaining the total number of non-cloud categories (i.e., the number of times the pixels are classified as water or sea ice, represented by N) and the mode of non-cloud categories of the pixels corresponding to the same position in the multiple classification result matrices, and obtaining a daily distribution map of water and ice based on the total number of non-cloud categories and the mode of non-cloud categories;

[0069] A large number of tests have shown that in the classification process of this embodiment, the classification accuracy of water is the highest. Therefore, the following steps are formed:

[0070] Step S31: For each pixel in the classification result matrix, if the total number of non-cloud categories is greater than the first threshold, the value of the pixel is the mode of the non-cloud category; if the total number of non-cloud categories is 0, the value of the pixel is the category label corresponding to the cloud; thereby obtaining a daily composite classification map M1; only extracting pixels with a category label of water in the daily composite classification map M1, and leaving other pixels empty, to obtain a daily water distribution map.

[0071] Step S32: For each pixel in the classification result matrix, if the total number of non-cloud categories is greater than the second threshold, the value of the pixel is the mode of the non-cloud category; if the total number of non-cloud categories is 0, the value of the pixel is the category label corresponding to the cloud; thereby obtaining a daily composite classification map M2, only extracting pixels with a category label of ice in the daily composite classification map M2, and leaving other pixels blank, to obtain a daily ice distribution map;

[0072] Among them, the second threshold is greater than the first threshold. This is because there is a phenomenon that a small number of cloud areas are easily misclassified as ice in the classification results. The misclassified cloud areas are filtered out by increasing the threshold. According to experience, the second threshold is set to 3; since the recognition accuracy of water is higher than that of sea ice, the threshold can be slightly lowered. According to experience, the first threshold is set to 1.

[0073] It should be noted that, considering that water has the highest classification accuracy, in the process of obtaining the daily composite classification maps M1 and M2, if the number of sea ice and water in the non-cloud category is equal, the value of the current pixel is the category label corresponding to water.

[0074] Step S4: Fusing the daily water and ice distribution maps to obtain a daily synthetic sea ice map.

[0075] Preferably, the values ​​of the pixels at each corresponding position in the daily water and ice distribution map are compared respectively. If they are consistent, they remain unchanged; if they are inconsistent, the value of the pixel is the mode of the non-empty categories of the neighborhood set for the current pixel in the daily water and ice distribution map. If the number of ice and water in the non-empty categories of the set neighborhood is equal, the value of the current pixel is the category label corresponding to water. (That is, when the number of ice and water in the neighborhood is equal, the judgment result is the label corresponding to water, because the classification accuracy of water is higher than that of sea ice); the values ​​of the remaining pixels are the category labels corresponding to clouds, thereby forming a daily synthetic ice map. The schematic diagram of the synthesis process of the daily synthetic ice map is shown below. Figure 2 shown.

[0076] On the basis of obtaining the daily synthetic ice map, this embodiment may further include the following steps:

[0077] Step S5: Based on the sea ice mapping for seven consecutive days, a weekly synthetic ice map is obtained:

[0078] The total number of non-cloud categories and the majority of non-cloud categories for the pixels corresponding to the same position in the ice map for seven consecutive days are obtained; if the total number of non-cloud categories for the current pixel is not 0, its value is the majority of non-cloud categories for the current pixel; if the number of ice and water in the non-cloud category is the same, the value of the current pixel is the category label corresponding to water; if the total number of non-cloud categories for the current pixel is 0, its value is the category label corresponding to cloud; this forms a weekly synthetic ice map.

[0079] In summary, compared with the prior art, the sea ice mapping method provided in this embodiment has the following beneficial effects:

[0080] First, on the basis of realizing automatic classification of images in the study area, the present invention uses the cloud motion characteristics in the Arctic region and the multi-temporal determination method of ground objects to merge the classification results of multiple images obtained every day, and synthesizes the daily ice map on this basis, thereby effectively solving the problem of easy confusion between ice and clouds;

[0081] Second, the present invention also forms a weekly synthetic ice map based on the daily synthetic ice map;

[0082] Third, in response to the low spatial resolution of traditional sea ice products, the present invention constructs a priori pixel sample library for polar feature classification based on the constructed multi-temporal remote sensing images of the Arctic, and uses a random forest algorithm with multi-feature band fusion to automatically classify the images of the study area. The classification results have greatly improved the spatial resolution and accuracy of traditional MODIS sea ice products. At the same time, the present invention also improves the feature bands selected in the classification process and the classification categories, so that the classification accuracy is effectively improved.

[0083] Example 2

[0084] Specific embodiment 2 of the present invention is used to verify the accuracy of the sea ice mapping method in embodiment 1 of the present invention; specifically,

[0085] In this embodiment, the classification model is trained on the constructed prior sample library based on the multi-feature band fusion random forest (MFLFRF) algorithm for the preprocessed original image, and the verified classification model is used to automatically classify the single-view images in batches. Taking two single-view images as an example, the image classification results are shown. Figure 3 As shown; among them, Figure 3 (a) and (b) are false color images of the B721 band combination, collected on July 3, 2018 and August 3, 2019, respectively; Figure 3 (c) and (d) are Figure 3 (a), (b) Classification results of images using MFLFRF classification algorithm; Figure 3 (e) and (f) are Figure 3 (a) and (b) correspond to MOD29 sea ice products.

[0086] In the B721 band combination mode, it is displayed as the identification of blue clouds (such as Figure 3 (a) The area marked by the box is mistakenly identified as sea ice in the MOD29 ice map. The classification model in the above embodiment can accurately identify the blue clouds in the area, but there are still a few cloud shadow areas that are misidentified. Figure 3 (b) The position marked by the box), the MOD29 ice map product cannot accurately identify the water area in the sea ice, while the classification model in the above embodiment (for example, based on the MFLFRF algorithm) can accurately identify the sea ice and the water area. In summary, the MFLFRF classification algorithm has a high recognition accuracy in the process of identifying sea ice, water and clouds.

[0087] The daily synthetic ice graph and weekly synthetic ice graph obtained based on the steps in Example 1 are as follows Figure 4 As shown; among them, Figure 4 (a) is the daily composite ice map obtained based on Terra on August 1, 2019; Figure 4 (b) is a weekly composite ice map based on Terra and Aqua from August 1 to 7, 2019. Figure 4 As can be seen from (a) and (b), the Arctic region is affected by clouds. Most areas in the daily synthetic ice map are covered by clouds, and only a small part is clear sky area. Most areas in the weekly synthetic ice map are clear sky areas, but there are still some areas that are covered by clouds for 7 consecutive days, showing cloud areas.

[0088] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0089] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for sea ice mapping, It is characterized in that The method is applied in the Arctic and comprises the following steps: Obtain and preprocess multiple MODIS L1B image data collected daily in the study area; Each image data after preprocessing is classified to obtain a classification result matrix of each image data; the classification categories include sea ice, water and cloud; each pixel in the classification result matrix stores the category label corresponding to the pixel; the classification result matrix of each image data is obtained by performing the following operations: Extract the characteristic bands corresponding to each image data after preprocessing, and classify the verified classification model based on the extracted characteristic bands to obtain the classification result matrix of each image data; The characteristic bands include: data corresponding to MODIS reflectivity bands 7, 2, and 1; NDSI, and MRELBP; Obtaining the total number of non-cloud categories and the mode of non-cloud categories of pixels corresponding to the same position in multiple classification result matrices, and obtaining a daily water and ice distribution map based on the total number of non-cloud categories and the mode of non-cloud categories; fusing the daily water and ice distribution maps to obtain a daily synthetic ice map; The daily water distribution map is obtained by performing the following operations: For each pixel in the classification result matrix, if the total number of non-cloud categories is greater than a first threshold, the value of the pixel is the mode of the non-cloud categories; Pixels with the category label of water are extracted, and other pixels are set to be empty, thus obtaining the daily water distribution map; The daily ice distribution map is obtained by performing the following operations: For each pixel in the classification result matrix, if the total number of non-cloud categories is greater than a second threshold, the value of the pixel is the mode of the non-cloud categories; Pixels with the category label of sea ice are extracted, and other pixels are left blank to obtain the daily ice distribution map; wherein the second threshold is greater than the first threshold; The fusing of the daily water and ice distribution maps to obtain a daily synthetic ice map includes: The values ​​of the pixels at each corresponding position in the daily water and ice distribution maps are compared respectively. If they are consistent, they remain unchanged; if they are inconsistent, the value of the pixel is the non-empty category mode of the neighborhood set for the current pixel in the daily water and ice distribution maps; the values ​​of the remaining pixels are the category labels corresponding to the clouds, thereby forming a daily synthetic ice map.

2. The sea ice mapping method according to claim 1, It is characterized in that The method further comprises: obtaining a weekly synthetic ice map based on the sea ice map for seven consecutive days: Obtain the total number and mode of non-cloud categories of pixels corresponding to the same location in the sea ice map for seven consecutive days; If the total number of non-cloud categories of the current pixel is not 0, its value is the mode of the non-cloud category of the current pixel. If the number of sea ice and water in the non-cloud category is equal, the value of the current pixel is the category label corresponding to water; If the total number of non-cloud categories of the current pixel is 0, its value is the category label corresponding to the cloud; This forms a weekly composite ice map.

3. The sea ice mapping method according to claim 1, It is characterized in that The verified classification model is obtained by performing the following operations: Obtain MODIS L1B historical image data of multiple time periods in the study area and preprocess them in batches; Extracting a plurality of sample data corresponding to each classification category from the batch preprocessed historical image data, and obtaining a category label corresponding to each sample data; Based on the sample data and its corresponding category labels, a priori pixel sample library is constructed; A training sample and a verification sample are selected from the prior pixel sample library, a classification model is trained based on the characteristic band of each sample data in the training sample and its corresponding category label, and the classification accuracy of the trained classification model is verified based on the characteristic band of each sample data in the verification sample and its corresponding category label. If the classification accuracy exceeds the classification accuracy threshold, the verification is passed, thereby obtaining a verified classification model.

4. The sea ice mapping method according to claim 1 or 3, It is characterized in that The classification category corresponding to the cloud further includes blue cloud, white cloud and red cloud subcategories; wherein each subcategory corresponds to a corresponding subcategory label.

5. The sea ice mapping method according to claim 1, It is characterized in that The preprocessing includes geometric correction, radiation calibration, solar zenith angle correction and land mask.

Images