A method, apparatus, equipment and medium for filling normalized vegetation index (NWRI)

Abstract

This invention discloses a method, apparatus, device, and medium for imputing the Normalized Difference Vegetation Index (NDV), relating to the field of satellite or aerial image data processing. The NDC imputation method includes: segmenting the target region into multiple temporal phases based on a second time-series dataset of the target region to determine each homogeneous region of the target region; numerically correcting the first time-series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time-series dataset of the target region and the second time-series dataset; wherein each pixel in the image data of the first and second time-series datasets corresponds to a NDC; the spatial resolution of the first time-series dataset is smaller than the spatial resolution of the second time-series dataset; and imputing missing values ​​in the missing image data of the second time-series dataset based on the corrected first time-series dataset to obtain a second time-series dataset with imputed missing values.

Description

Technical Field

[0001] This invention relates to the field of satellite or aerial image data processing, and more particularly to a method, apparatus, device, and medium for filling in the normalized vegetation index. Background Technology

[0002] High-resolution Normalized Difference Vegetation Index (NDVI) time-series data are crucial for supporting decision-making in areas such as surface vegetation monitoring, agricultural management practices, climate change research, disaster early warning, and land use change mapping. Particularly in agricultural engineering, the spatiotemporal continuity of NDVI data significantly impacts crop yield prediction and farmland resource management efficiency. While existing NDVI products generated by Moderate Resolution Imaging Spectroradiometers (MODIS) offer an 8-day temporal resolution, their low spatial resolution makes it difficult to capture surface details, thus limiting their application in heterogeneous regions. Although medium-resolution satellite data (such as NDVI data generated by Sentinel-2) improves the spatiotemporal resolution of surface observation data, significant data gaps remain in heterogeneous landscape areas with abundant cloud cover and complex topography.

[0003] Therefore, how to fill in the missing values ​​in high-resolution normalized vegetation index time series data has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to provide a method, apparatus, device, and medium for filling missing values ​​in normalized vegetation index (NWDI) time series data.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for imputing normalized vegetation index (NVI) includes: segmenting the target region into multiple temporal phases based on a second time-series dataset of the target region to determine homogeneous regions within the target region; numerically correcting the first time-series dataset corresponding to each homogeneous region based on image data of each homogeneous region in a first time-series dataset of the target region and the second time-series dataset, to obtain a numerically corrected first time-series dataset; wherein each pixel in the image data of the first and second time-series datasets corresponds to a normalized vegetation index; the spatial resolution of each image data in the first time-series dataset is greater than the spatial resolution of each image data in the second time-series dataset; and imputing missing values ​​in the missing image data in the second time-series dataset based on the corrected first time-series dataset, to obtain a second time-series dataset with imputed missing values.

[0006] In one optional embodiment of this application, the step of performing multi-temporal segmentation of the target region based on the second time-series dataset of the target region to determine each homogeneous region of the target region includes: using a simple non-iterative clustering algorithm, combining the spatial distribution of image data in the target region and the temporal distribution between different image data in the second time-series dataset, to perform multi-temporal segmentation of the target region and determine each homogeneous region of the target region.

[0007] In one optional embodiment of this application, the step of numerically correcting the first time series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset to obtain a numerically corrected first time series dataset includes: updating the image data in the first time series dataset based on the median of the differences between the image data of each image data in the homogeneous region to obtain an updated first time series dataset; and numerically correcting the updated first time series dataset based on the second time series dataset to obtain a numerically corrected first time series dataset.

[0008] In one optional embodiment of this application, updating the image data in the first time-series dataset based on the median of the differences between the various image data in the homogeneous region to obtain an updated first time-series dataset includes: for any homogeneous region, taking the image data in the first time-series dataset corresponding to the homogeneous region as the reference data; determining first update data based on the median of the differences between the image data in the second time-series dataset and the reference data; updating the image data in the second time-series dataset based on the first update data and the reference data; determining second update data based on the median of the differences between the image data in the third time-series dataset and the image data in the second position; updating the image data in the third position in the first time-series dataset based on the first update data, the second update data, and the reference data; and so on, until all image data in the first time-series dataset are traversed to obtain the updated first time-series dataset.

[0009] In one optional embodiment of this application, the image data in the first time-series dataset is updated using the following formula to obtain the updated first time-series dataset: ; ; in, This represents the nth image data in the updated first time series dataset, where n is an integer greater than or equal to 2; This represents the image data that is the first in the time series of the first time series dataset; This represents the image data sorted at position i in the first time series dataset; This represents the median of the difference between the normalized vegetation index of each image data ranked at position i and each image data ranked at position i-1 in the first time series dataset in a homogeneous region. This indicates the coordinates of each image in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

[0010] In one optional embodiment of this application, the step of numerically correcting the updated first time series dataset according to the second time series dataset to obtain a numerically corrected first time series dataset includes: taking each image data in the second time series dataset as the target, using a linear transfer function to numerically correct each image data in the first time series dataset to obtain a corrected first time series dataset.

[0011] In one optional embodiment of this application, the updated first time-series data is numerically corrected using the following formula to obtain the numerically corrected first time-series dataset: ; in, This indicates the coordinates of each image in the second time series dataset. The normalized vegetation index sequence corresponding to the pixels; and Indicates the coordinates in the updated first time series dataset. The correction factor for the normalized vegetation index of the pixels; and Minimize the corrected first time series dataset using the least squares method. The differences between them were obtained; This indicates the coordinates of each image in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

[0012] Compared with existing technologies, the normalized vegetation index (NVI) imputation method provided by this invention performs multi-temporal segmentation on a second time-series dataset of the target region to obtain homogeneous regions within the target region. Using homogeneous units as correction units, it numerically corrects a first time-series dataset with a spatial resolution lower than the second time-series dataset, and then imputes missing image data in the second time-series dataset based on the numerically corrected first time-series dataset. This method, by combining the second time-series dataset with multi-temporal segmentation of the target region, considers both the spatial and temporal distributions of the NVI, thus helping to improve the data quality of the corrected first time-series dataset and the data quality of the second time-series dataset after missing value imputation during the numerical correction process.

[0013] The present invention also provides a device for filling normalized vegetation index, comprising: The temporal segmentation unit is used to perform multi-temporal segmentation of the target region based on the second temporal dataset of the target region, and to determine each homogeneous region of the target region.

[0014] The numerical correction unit is used to perform numerical correction on the image data of each homogeneous region in the first time series dataset of the target area to obtain the numerically corrected first time series dataset; wherein, each pixel in the image data of the first time series dataset and the second time series dataset corresponds to a normalized vegetation index; the spatial resolution of each image data in the first time series dataset is greater than the spatial resolution of each image data in the second time series dataset.

[0015] The numerical imputation unit is used to impute missing values ​​in the missing image data in the second time series dataset based on the corrected first time series dataset, so as to obtain the second time series dataset after missing value imputation.

[0016] Compared with the prior art, the beneficial effects of the normalized vegetation index filling device provided by the present invention are the same as those of the normalized vegetation index filling method described in the above technical solutions, and will not be repeated here.

[0017] The present invention also provides an electronic device, comprising: processor.

[0018] Memory used to store executable instructions of the processor.

[0019] The processor is configured to execute the above-described method for filling in the normalized vegetation index by running instructions in the memory.

[0020] Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as those of the normalized vegetation index filling method described in the above technical solution, and will not be repeated here.

[0021] The present invention also provides a computer storage medium storing instructions that, when executed, implement the above-described method for filling the normalized vegetation index.

[0022] Compared with the prior art, the beneficial effects of the computer storage medium provided by the present invention are the same as the beneficial effects of the normalized vegetation index filling method described in the above technical solutions, and will not be repeated here. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1 A flowchart illustrating the method for filling in the normalized vegetation index (NWRI) provided in this application embodiment.

[0024] Figure 2 This is a schematic diagram of a homogeneous region of the target area provided in an embodiment of this application.

[0025] Figure 3 Data curves provided for embodiments of this application Figure 1 .

[0026] Figure 4 Data curves provided for embodiments of this application Figure 2 .

[0027] Figure 5 Data curves provided for embodiments of this application Figure 3 .

[0028] Figure 6 A structural diagram of the normalized vegetation index filling device provided in the embodiments of this application.

[0029] Figure 7 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0030] To facilitate a clear description of the technical solutions in the embodiments of the present invention, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect. For example, the first threshold and the second threshold are merely used to distinguish different thresholds and do not limit their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" are not necessarily different.

[0031] It should be noted that in this invention, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0032] In this invention, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, a combination of a and b, a combination of a and c, a combination of b and c, or a, b, and c, where a, b, and c can be single or multiple.

[0033] High-resolution Normalized Difference Vegetation Index (NDVI) time-series data are crucial for supporting decision-making in areas such as surface vegetation monitoring, agricultural management practices, climate change research, disaster early warning, and land use change mapping. Particularly in agricultural engineering, the spatiotemporal continuity of NDVI data significantly impacts crop yield prediction and farmland resource management efficiency. While existing NDVI products generated by Moderate Resolution Imaging Spectroradiometers (MODIS) offer an 8-day temporal resolution, their low spatial resolution makes it difficult to capture surface details, thus limiting their application in heterogeneous regions. Although medium-resolution satellite data (such as NDVI data generated by Sentinel-2) improves the spatiotemporal resolution of surface observation data, significant data gaps remain in heterogeneous landscape areas with abundant cloud cover and complex topography.

[0034] How to fill in the missing values ​​in high-resolution normalized vegetation index time series data has become a technical problem that urgently needs to be solved by those skilled in the art.

[0035] To address the aforementioned technical problems, this application provides a method, apparatus, device, and medium for filling the normalized vegetation index, which will be described in detail in the following embodiments.

[0036] This application first provides a method for filling in the normalized vegetation index (NVI). Please refer to [reference needed]. Figure 1 , Figure 1 A flowchart illustrating the method for filling in the normalized vegetation index (NWRI) provided in this application embodiment.

[0037] like Figure 1 As shown, the method for filling the normalized vegetation index includes the following steps S101 to S103.

[0038] S101, based on the second time-series dataset of the target region, perform multi-temporal segmentation on the target region to determine each homogeneous region of the target region.

[0039] The second time-series dataset refers to the preprocessed Normalized Difference Vegetation Index (NDVI) covering the target area. In practical applications, the second time-series dataset includes multiple image data sorted by shooting time, and each pixel in the image data corresponds to an NDVI at that pixel location.

[0040] NDVI is a widely used indicator in remote sensing to assess vegetation growth status and vegetation cover. NDVI is calculated using reflectance data in the near-infrared (NIR) and red bands of satellite or aerial imagery.

[0041] The purpose of the normalized vegetation index filling method provided in this application embodiment is to solve the problem of discontinuous NDVI data caused by the spatiotemporal discontinuity of optical images under cloud and fog weather, and to improve the accuracy and efficiency of NDVI reconstruction in complex terrain and heterogeneous landscape areas. Its core idea is to fill in NDVI data with high spatial resolution but spatiotemporal discontinuity of optical images, thereby solving the above-mentioned technical problems.

[0042] In practical applications, the good spatial resolution of medium-resolution satellites is a significant advantage. Therefore, in one optional embodiment of this application, the second time-series dataset can utilize a series of image data acquired by the Sentinel-2 satellite. Band features are extracted from the near-infrared and red bands of the image data to obtain NDVI data. Specifically, the Sentinel-2 satellite has a spatial resolution of 10 meters and a temporal resolution of 5 to 10 days. However, considering the presence of numerous clouds and complex terrain in heterogeneous landscape areas, the NDVI data acquired by the Sentinel-2 satellite still suffers from missing data. Therefore, NDVI data augmentation is necessary.

[0043] Considering that the image data acquired by Sentinel-2 satellite may have different time intervals, before performing multi-temporal segmentation on the target area, the second time-series dataset acquired by Sentinel-2 satellite can be declouded and repaired to improve the quality of the second time-series dataset.

[0044] Specifically, the process of removing clouds and repairing the second time-series dataset includes: performing cloud removal processing on each image data in the second time-series dataset based on a pre-trained machine learning model and combining multispectral features to obtain the cloud-removed second time-series dataset; and performing low-quality pixel removal processing on the cloud-removed second time-series dataset based on the maximum value synthesis method to obtain the low-quality pixel removal processed second time-series dataset.

[0045] That is, a machine learning model is used to identify and extract cloud-polluted pixels in each image data of the second time series dataset, and the maximum NDVI value of each pixel is selected from the image data of the second time series dataset to generate synthetic data at 16-day intervals, so as to reduce the impact of cloud cover.

[0046] Furthermore, regarding the above S101, the purpose of performing multi-temporal segmentation on the target area is to obtain homogeneous areas in the target area with uniform changes in the normalized vegetation index by performing multi-temporal segmentation on the target area. In the subsequent data imputation process of the image data of the second time series dataset, the homogeneous objects are used as the basic units to improve the structural correlation between the imputed data and the target area and the imputation efficiency.

[0047] Specifically, S101 includes: using a simple non-iterative clustering algorithm, combining the spatial distribution of image data in the second time series dataset in the target region and the temporal distribution between different image data, to perform multi-temporal segmentation of the target region and determine each homogeneous region of the target region.

[0048] Please refer to Figure 2 , Figure 2 This is a schematic diagram of a homogeneous region of the target area provided in an embodiment of this application.

[0049] like Figure 2 As shown, Figure 2 The target region includes a first homogeneous region 201, a second homogeneous region 202, a third homogeneous region 203, and a fourth homogeneous region 204. These homogeneous regions together constitute the target region. Each pixel in a homogeneous region corresponds to an NDVI value. The NDVI values ​​of each pixel in each homogeneous region are similar in time and space. That is, the NDVI values ​​of each pixel in the same homogeneous region are similar in an image data of the second time series dataset, and the differences in images at different time series are uniform.

[0050] In practical applications, the simple non-iterative clustering algorithm can be used to perform multi-temporal segmentation of the target region based on the Google Earth Engine platform. Specifically, in the application process, the compactness of the simple non-iterative clustering algorithm can be set to 0, the seed size to 15, the neighborhood range to 40, and the connectivity to 8.

[0051] S102, based on the image data of each homogeneous region in the first time series dataset of the target area and the second time series dataset, the first time series dataset corresponding to each homogeneous region is numerically corrected to obtain the numerically corrected first time series dataset.

[0052] Similar to the second time-series dataset mentioned above, the image data of the first time-series dataset also refers to the preprocessed Normalized Difference Vegetation Index (NDVI) covering the target area.

[0053] In this embodiment of the application, the first time series dataset is the basic condition for filling in the second time series dataset.

[0054] In practical applications, the first time-series dataset can be acquired by the Moderate Resolution Imaging Spectroradiometer (MODIS) mounted on the Terra and Aqua satellites, with a spatial resolution of 250 meters and a temporal resolution of 16 days. MODIS typically includes MOD13Q1 and MYD13Q1. In practical applications, to further improve the temporal resolution of the NDVI data acquired by MODIS and to provide higher reconstruction accuracy for filling in the NDVI data acquired by MODIS, the best values ​​can be selected from all 16 days of NDVI data based on the coverage, clouds, and aerosols of the image data acquired by MOD13Q1 and MYD13Q1 to synthesize the NDVI data acquired by MOD13Q1 and MYD13Q1, so as to generate NDVI data with a higher temporal resolution (8 days).

[0055] Considering that the image data acquired by MODIS and Sentinel-2 satellites may have different time intervals, cloud removal and repair can be performed on the image data acquired by MODIS to improve the quality of the first time-series dataset.

[0056] To improve the data quality of the first time-series dataset, the method further includes: employing a cloud removal algorithm to perform cloud removal processing on each image data in the first time-series dataset to obtain a cloud-removed first time-series dataset; using time harmonic analysis to remove noise from the cloud-removed first time-series dataset and to impute temporal missing values ​​in the first time-series dataset to obtain a noise-removed and time-missing value-impacted first time-series dataset; and employing bicubic convolution interpolation to impute spatial missing values ​​in the noise-removed and time-missing value-impacted first time-series dataset to obtain a spatially imputed first time-series dataset.

[0057] Furthermore, the above S102 includes the following S1 and S2.

[0058] S1, based on the median of the differences between the image data in the first time series dataset in the homogeneous region, update the image data in the first time series dataset to obtain the updated first time series dataset.

[0059] The purpose of updating the image data in the first time series dataset is to eliminate the difference in NDVI changes between image data at the same time in the first and second time series datasets.

[0060] For easier understanding, please refer to Figure 3 , Figure 3 Data curves provided for embodiments of this application Figure 1 .

[0061] like Figure 3 As shown, Figure 3 This includes: the updated MODIS NDVI data curve (i.e., the data curve corresponding to the updated first time series dataset), and discrete Sentinel-2 NDVI data (i.e., data points of the second time series dataset).

[0062] Specifically, the updated MODIS NDVI data curve refers to the change curve of the MODIS NDVI data of a pixel at a certain position in a homogeneous region within a preset time range. Discrete Sentinel-2 NDVI data refers to the NDVI data of the pixel at that position in the Sentinel-2 image data within the homogeneous region.

[0063] based on Figure 3 It can be seen that the changing trend of the updated MODIS NDVI data curve is roughly the same as that of the discrete Sentinel-2 NDVI data.

[0064] In the prior art, the similar pixel search method (GF-SG) is usually used to search for similar pixels near the target pixel in order to eliminate the difference in NDVI variation between the first time series dataset and the second time series dataset. The principle is shown in the following formula (1): (1); in, Reference timing data representing the target pixel; The position of the target pixel in the image data of the first time-series dataset and the second time-series dataset is indicated, wherein the target pixel has the same position in the image data of the first time-series dataset and the second time-series dataset; j represents the j-th similar pixel of the target pixel; n represents the total number of similar pixels of the target pixel; This represents a data matrix composed of the NDVI data corresponding to the j-th similar pixel in the image data of the first time-series dataset; This indicates the position of the j-th similar pixel in the target region; This represents the correlation coefficient between the j-th similar pixel and the target pixel.

[0065] That is, the NDVI data of all similar pixels are added together according to their weights to generate reference time-series data of the target pixels, and the reference time-series data of each target pixel is used as the time-series dataset after the change difference is eliminated.

[0066] However, this method usually requires using a moving window to search for similar pixels of the target pixel, which is usually accompanied by a lot of computation. Moreover, this pixel-level data update method ignores the inherent shape information of ground objects and the regional similarity and temporal correlation of the normalized vegetation index, which can easily cause intra-class salt-and-pepper noise.

[0067] To address this issue, the normalized vegetation index (NDVI) filling method provided in this application update the first time-series dataset based on the homogeneous regions of the target area obtained in S101, using the homogeneous regions as objects, to eliminate the NDVI variation differences between the various image data in the first time-series dataset.

[0068] Specifically, S1 above includes the following S11 to S14.

[0069] S11, for any homogeneous region, the image data with the first time sequence in the first time series dataset corresponding to the homogeneous region is used as the reference data.

[0070] S12, determine the first update data based on the median of the difference between the second-ranked image data in the first time series dataset and the reference data; update the second-ranked image data in the first time series dataset based on the first update data and the reference data.

[0071] S13, determine the second update data based on the median of the difference between the third-ranked image data and the second-ranked image data in the first time series dataset; update the third-ranked image data in the first time series dataset based on the first update data, the second update data, and the reference data.

[0072] The purpose of using the median of the differences is to eliminate the influence of low-quality pixels in the image data on the normalized vegetation index of the homogeneous region.

[0073] S14, and so on, until all image data in the first time series dataset are traversed to obtain the updated first time series dataset.

[0074] Specifically, the process of updating each image data in the first time series dataset can be represented by the following formulas (2) and (3): (2); (3); in, This represents the nth image data in the updated first time series dataset, where n is an integer greater than or equal to 2; This represents the image data that is the first in the time series of the first time series dataset; This represents the image data sorted at position i in the first time series dataset; This represents the median of the difference between the normalized vegetation index of each image data ranked at position i and each image data ranked at position i-1 in the first time series dataset in a homogeneous region. This indicates the coordinates of each image in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

[0075] S2, based on the second time series dataset, perform numerical correction on the updated first time series dataset to obtain the numerically corrected first time series dataset.

[0076] The purpose of numerical correction to the updated first time series dataset is to eliminate the numerical differences between the first time series dataset and the second time series dataset.

[0077] Please refer to Figure 4 , Figure 4 Data curves provided for embodiments of this application Figure 2 .

[0078] Figure 4 This includes the updated MODIS NDVI data curve (i.e., the data curve corresponding to the updated first time series dataset), discrete Sentinel-2 NDVI data (i.e., the data points corresponding to the second time series dataset), and the numerically corrected MODIS NDVI data curve (i.e., the data curve corresponding to the numerically corrected first time series dataset).

[0079] according to Figure 4 It can be seen that, compared with the numerically corrected MODIS NDVI data curve, the numerically corrected MODIS NDVI data curve eliminates the amplitude difference between the updated MODIS NDVI data curve and the discrete Sentinel-2 NDVI data.

[0080] Specifically, S2 includes: taking each image data in the second time series dataset as the target, using a linear transfer function to numerically correct each image data in the first time series dataset, and obtaining the corrected first time series dataset.

[0081] The numerical correction of each image data in the first time series dataset using a linear transfer function can be expressed by the following formula (4): (4); in, express In the coordinates of each image data The normalized vegetation index sequence corresponding to the pixels; and Indicates the coordinates in the updated first time series dataset. The correction factor for the normalized vegetation index of the pixels; and Minimize the corrected first time series dataset using the least squares method. The differences between them were obtained; This indicates the coordinates of each image in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

[0082] S103, based on the corrected first time series dataset, missing image data in the second time series dataset are imputed to obtain the second time series dataset after missing value imputation.

[0083] S103 above refers to using the image data of the corrected first time series dataset as the missing image data in the second time series dataset, thereby completing the filling of missing values ​​in the second time series dataset.

[0084] In practical applications, considering the potential for low-quality values ​​in the second time-series dataset, weighted Savitzky-Golay (SG) filtering can be used to denoise the imputed second time-series dataset. This process preserves the NDVI variation trend in each image data within the second time-series dataset while removing cloud contamination and sensor errors.

[0085] Please refer to Figure 5 , Figure 5 Data curves provided for embodiments of this application Figure 3 .

[0086] like Figure 5 As shown in Figure 5, the Sentinel-2 NDVI data curves include those with missing values ​​filled in, as well as those with filtered Sentinel-2 NDVI data curves.

[0087] according to Figure 5 The filtered Sentinel-2 NDVI data curve shown is smoother than the Sentinel-2 NDVI data curve after missing value imputation.

[0088] In the weighted Savitzky-Golay (SG) filtering method, the weights are modified by two adjustment factors during the weighting process. Specifically, for each image data in the original second time series dataset, the weight is set to 1. For the infilled image data, the weight can be expressed by the following formula (5): (5); in, This represents the weight of the i-th infilled image data; This represents the NDVI standard deviation between adjacent pixels within a window centered on the target pixel in the i-th filled image data; express The maximum value in.

[0089] In practical applications, the half-width of the window can be set to 4, and the degree of the smoothing polynomial used in the weighted Savitzky-Golay (SG) filter can be set to 6.

[0090] Furthermore, in order to evaluate the normalized vegetation index filling method provided in the embodiments of this application, the root mean square error, average deviation, and edge information were used to quantitatively analyze the accuracy of the first time series dataset after filling.

[0091] Specifically, the root mean square error, mean deviation, and edge index can be calculated using the following formulas (6) to (8): (6); (7); (8); in, Indicates the root mean square error; This represents the NVDI value of the i-th pixel in the image data of the second time-series dataset after missing value imputation; This represents the NVDI value of the i-th pixel in the image data of the reference time-series dataset; N represents the total number of pixels in the image data. Indicates the average deviation; Represents edge information; This represents the NDVI value corresponding to the pixel in the x-th row and y-th column of the image data.

[0092] To quantitatively analyze the normalized vegetation index (NDVI) filling method provided in this application embodiment, the study selected the NDVI image of Erhai Lake acquired by Sentinel-2 on January 17, 2019 as image data, and used the NDVI image of the same area on January 14, 2019 as a reference for accuracy quantitative analysis. The results showed that RMSE, AD, and Edge were 0.0615, 0.0390, and -0.1297, respectively. These data indicate that the normalized vegetation index filling method has high accuracy in filling missing image data and can meet the production needs of NVDI products.

[0093] In addition, during the process of filling in the normalized vegetation index (NDVI) values ​​on the GEE platform, the above steps S101 to S103 took a total of about 15 minutes, indicating that the method has high feasibility and efficiency in large-scale NDVI reconstruction.

[0094] In summary, the normalized vegetation index (NVI) imputation method provided in this application involves multi-temporal segmentation of a second time-series dataset of a target region to obtain homogeneous regions within the target region. Using homogeneous units as correction units, the method performs numerical correction on a first time-series dataset whose spatial resolution is lower than that of the second time-series dataset, and then imputes missing image data in the second time-series dataset based on the numerically corrected first time-series dataset. This method, by combining the second time-series dataset with multi-temporal segmentation of the target region, considers both the spatial and temporal distributions of the NVI, thus helping to improve the data quality of the corrected first time-series dataset and the data quality of the second time-series dataset after missing value imputation during the numerical correction process.

[0095] This application also provides a device for filling the normalized vegetation index (NWRI), please refer to [reference needed]. Figure 6 , Figure 6 A structural diagram of the normalized vegetation index filling device provided in the embodiments of this application.

[0096] like Figure 6 As shown, the device for filling the normalized vegetation index includes: The temporal segmentation unit 601 is used to perform multi-temporal segmentation on the target region based on the second temporal dataset of the target region, and to determine each homogeneous region of the target region.

[0097] The numerical correction unit 602 is used to perform numerical correction on the first time series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset, to obtain the numerically corrected first time series dataset; wherein, each pixel in the image data of the first time series dataset and the second time series dataset corresponds to a normalized vegetation index; the spatial resolution of each image data in the first time series dataset is smaller than the spatial resolution of each image data in the second time series dataset.

[0098] The numerical imputation unit 603 is used to impute missing values ​​in the missing image data in the second time series dataset based on the corrected first time series dataset, so as to obtain the second time series dataset after missing value imputation.

[0099] In one optional embodiment of this application, the step of performing multi-temporal segmentation of the target region based on the second time-series dataset of the target region to determine each homogeneous region of the target region includes: using a simple non-iterative clustering algorithm, combining the spatial distribution of image data in the target region and the temporal distribution between different image data in the second time-series dataset, to perform multi-temporal segmentation of the target region and determine each homogeneous region of the target region.

[0100] In one optional embodiment of this application, the step of numerically correcting the first time series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset to obtain a numerically corrected first time series dataset includes: updating the image data in the first time series dataset based on the median of the differences between the image data of each image data in the homogeneous region to obtain an updated first time series dataset; and numerically correcting the updated first time series dataset based on the second time series dataset to obtain a numerically corrected first time series dataset.

[0101] In one optional embodiment of this application, updating the image data in the first time-series dataset based on the median of the differences between the various image data in the homogeneous region to obtain an updated first time-series dataset includes: for any homogeneous region, taking the image data in the first time-series dataset corresponding to the homogeneous region as the reference data; determining first update data based on the median of the differences between the image data in the second time-series dataset and the reference data; updating the image data in the second time-series dataset based on the first update data and the reference data; determining second update data based on the median of the differences between the image data in the third time-series dataset and the image data in the second position; updating the image data in the third position in the first time-series dataset based on the first update data, the second update data, and the reference data; and so on, until all image data in the first time-series dataset are traversed to obtain the updated first time-series dataset.

[0102] In one optional embodiment of this application, the image data in the first time-series dataset is updated using the following formula to obtain the updated first time-series dataset: ; ; in, This represents the nth image data in the updated first time series dataset, where n is an integer greater than or equal to 2; This represents the image data that is the first in the time series of the first time series dataset; This represents the image data sorted at position i in the first time series dataset; This represents the median of the difference between the normalized vegetation index of each image data ranked at position i and each image data ranked at position i-1 in the first time series dataset in a homogeneous region. This indicates the coordinates of each image in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

[0103] In one optional embodiment of this application, the step of numerically correcting the updated first time series dataset according to the second time series dataset to obtain a numerically corrected first time series dataset includes: taking each image data in the second time series dataset as the target, using a linear transfer function to numerically correct each image data in the first time series dataset to obtain a corrected first time series dataset.

[0104] In one optional embodiment of this application, the updated first time-series data is numerically corrected using the following formula to obtain the numerically corrected first time-series dataset: ; in, This indicates the coordinates of each image in the second time series dataset. The normalized vegetation index sequence corresponding to the pixels; and Indicates the coordinates in the updated first time series dataset. The correction factor for the normalized vegetation index of the pixels; and Minimize the corrected first time series dataset using the least squares method. The differences between them were obtained; This indicates the coordinates of each image in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

[0105] Compared with the prior art, the beneficial effects of the normalized vegetation index filling device provided by the present invention are the same as those of the normalized vegetation index filling method described in the above technical solutions, and will not be repeated here.

[0106] The method embodiments provided in this example and the system embodiments of this application belong to the same application concept. For technical details not described in detail in this example, please refer to the specific processing content of the normalized vegetation index filling method provided in the above embodiments of this application, which will not be repeated here.

[0107] This application also provides an electronic device, such as... Figure 7 As shown, Figure 7 This is a schematic diagram of an electronic device structure provided in an embodiment of this application.

[0108] like Figure 7 As shown, the electronic device includes: Processor 210.

[0109] Memory 200 for storing executable instructions of the processor 210.

[0110] The processor 210 is configured to execute the normalized vegetation index filling method disclosed in any of the above embodiments by running instructions in the memory 200.

[0111] The processor 210, memory 200, communication interface 220, input device 230 and output device 240 are interconnected via a bus.

[0112] A bus can include a pathway for transmitting information between various components of a computer system.

[0113] Processor 210 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0114] Processor 210 may include a main processor, as well as a baseband chip, modem, etc.

[0115] The memory 200 stores a program that executes the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 200 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0116] Input device 230 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, touch screen, etc.

[0117] Output device 240 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0118] The communication interface 220 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0119] The processor 210 executes the program stored in the memory 200 and calls other devices, and can be used to implement the various steps of any of the normalized vegetation index filling methods provided in the above embodiments of this application.

[0120] In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the normalized vegetation index filling methods of various embodiments of this application.

[0121] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0122] Furthermore, embodiments of this application may also be storage media storing a computer program, which is executed by a processor in the steps of the normalized vegetation index filling method of various embodiments of this application.

[0123] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0124] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0125] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0126] The modules and sub-modules in the apparatus and terminal in the various embodiments of this application can be merged, divided, and deleted according to actual needs.

[0127] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0128] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0129] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.

[0130] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0131] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0132] Finally, 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.

[0133] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for filling in the normalized vegetation index, characterized in that, include: Based on the second time-series dataset of the target region, the target region is segmented into multiple time phases to determine each homogeneous region of the target region. Based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset, the first time series dataset corresponding to each homogeneous region is numerically corrected to obtain the numerically corrected first time series dataset; wherein, each pixel in the image data of the first time series dataset and the second time series dataset corresponds to a normalized vegetation index; the spatial resolution of each image data in the first time series dataset is smaller than the spatial resolution of each image data in the second time series dataset; Based on the corrected first time series dataset, missing image data in the second time series dataset are imputed to obtain the second time series dataset after missing value imputation. The step of numerically correcting the first time series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset to obtain the numerically corrected first time series dataset includes: Based on the median of the differences between the image data of the first time series dataset in the homogeneous region, the image data in the first time series dataset is updated to obtain the updated first time series dataset. Based on the second time series dataset, the updated first time series dataset is numerically corrected to obtain the numerically corrected first time series dataset. Based on the median of the differences between the various image data in the first time-series dataset within the homogeneous region, the image data in the first time-series dataset are updated to obtain an updated first time-series dataset, including: For any homogeneous region, the image data with the first time sequence in the first time series dataset corresponding to the homogeneous region is used as the reference data; First update data is determined based on the median of the difference between the second-ranked image data in the first time series dataset and the reference data; the second-ranked image data in the first time series dataset is updated based on the first update data and the reference data. The second update data is determined based on the median of the difference between the third-ranked image data and the second-ranked image data in the first time series dataset; the third-ranked image data in the first time series dataset is updated based on the first update data, the second update data, and the baseline data. This process continues until all image data in the first time series dataset are traversed to obtain the updated first time series dataset.

2. The method for filling in the normalized vegetation index according to claim 1, characterized in that, The step of performing multi-temporal segmentation of the target region based on the second temporal dataset of the target region to determine each homogeneous region of the target region includes: A simple non-iterative clustering algorithm is used, combined with the spatial distribution of image data in the target region and the temporal distribution among different image data in the second time series dataset, to perform multi-temporal segmentation of the target region and determine each homogeneous region of the target region.

3. The method for filling in the normalized vegetation index according to claim 1, characterized in that, The image data in the first time-series dataset is updated using the following formula to obtain the updated first time-series dataset: ； ； in, This represents the nth image data in the updated first time series dataset, where n is an integer greater than or equal to 2; This represents the image data that is the first in the time series of the first time series dataset; This represents the image data sorted at position i in the first time series dataset; This represents the median of the difference between the normalized vegetation index of each image data ranked at position i and the image data ranked at position i-1 in the first time series dataset in a homogeneous region. This indicates the coordinates of each image data in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

4. The method for filling in the normalized vegetation index according to claim 1, characterized in that, The step of numerically correcting the updated first time series dataset based on the second time series dataset to obtain a numerically corrected first time series dataset includes: Using each image data in the second time series dataset as the target, a linear transfer function is used to numerically correct each image data in the first time series dataset to obtain the corrected first time series dataset.

5. The method for filling in the normalized vegetation index according to claim 1 or 4, characterized in that, The updated first time series data is numerically corrected using the following formula to obtain the numerically corrected first time series dataset: ； in, This indicates the coordinates of each image data in the second time series dataset. The normalized vegetation index sequence corresponding to the pixels; and Indicates the coordinates in the updated first time series dataset. The correction factor for the normalized vegetation index of the pixels; and Minimize the corrected first time series dataset using the least squares method. The differences between them were obtained; This indicates the coordinates of each image data in the updated first time series dataset. The normalized vegetation index sequence corresponding to the pixels.

6. A device for filling normalized vegetation index, characterized in that, include: The temporal segmentation unit is used to perform multi-temporal segmentation of the target region based on the second temporal dataset of the target region, and to determine each homogeneous region of the target region; The numerical correction unit is used to perform numerical correction on the first time series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset, to obtain the numerically corrected first time series dataset; wherein, each pixel in the image data of the first time series dataset and the second time series dataset corresponds to a normalized vegetation index; the spatial resolution of each image data in the first time series dataset is smaller than the spatial resolution of each image data in the second time series dataset; The numerical imputation unit is used to impute missing values ​​in the missing image data in the second time series dataset based on the corrected first time series dataset, so as to obtain the second time series dataset after missing value imputation. The step of numerically correcting the first time series dataset corresponding to each homogeneous region based on the image data of each homogeneous region in the first time series dataset of the target region and the second time series dataset to obtain the numerically corrected first time series dataset includes: Based on the median of the differences between the image data of the first time series dataset in the homogeneous region, the image data in the first time series dataset is updated to obtain the updated first time series dataset. Based on the second time series dataset, the updated first time series dataset is numerically corrected to obtain the numerically corrected first time series dataset. Based on the median of the differences between the various image data in the first time-series dataset within the homogeneous region, the image data in the first time-series dataset are updated to obtain an updated first time-series dataset, including: For any homogeneous region, the image data with the first time sequence in the first time series dataset corresponding to the homogeneous region is used as the reference data; First update data is determined based on the median of the difference between the second-ranked image data in the first time series dataset and the reference data; the second-ranked image data in the first time series dataset is updated based on the first update data and the reference data. The second update data is determined based on the median of the difference between the third-ranked image data and the second-ranked image data in the first time series dataset; the third-ranked image data in the first time series dataset is updated based on the first update data, the second update data, and the baseline data. This process continues until all image data in the first time series dataset are traversed to obtain the updated first time series dataset.

7. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the normalized vegetation index filling method according to any one of claims 1 to 5 by running instructions in the memory.

8. A computer storage medium, characterized in that, The computer storage medium stores instructions that, when executed, perform the normalized vegetation index filling method according to any one of claims 1 to 5.

Images