A method, system, device and storage medium for reconstructing NDVI images

Abstract

The application discloses a reconstruction method, system and device of NDVI images and a storage medium, relates to the technical field of image reconstruction, and comprises the following steps: acquiring original normalized difference vegetation index time series images; determining normalized difference vegetation index images of pixels with cloud pollution in the original normalized difference vegetation index time series images as target images; obtaining normalized difference vegetation index prediction values of the pixels with cloud pollution according to the number of time periods corresponding to the target images and a cloud pollution prediction value calculation formula; replacing actual normalized difference vegetation index values of the pixels with cloud pollution with the normalized difference vegetation index prediction values to obtain reconstructed target images; and determining reconstructed normalized difference vegetation index time series images according to the reconstructed target images and the original normalized difference vegetation index time series images. The application improves the reconstruction accuracy of NDVI images.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image reconstruction, in particular to a NDVI image reconstruction method, system, device and storage medium. BACKGROUND

[0002] In order to realize the monitoring of the growth of vegetation (including crops), generally, according to remote sensing images, based on the growth cycle of vegetation, the normalized difference vegetation index (NDVI) time series image of the vegetation growing to the vigorous stage is obtained, and then the NDVI image is reconstructed. At present, in the image reconstruction, two aspects are mainly considered, one is to directly filter through all the pixel points to construct the image. This method adds abnormal pixel points, which affects the image information of normal pixel points, so that the difference between the final image and the real image is large. The second is based on the detection of abnormal points, and the abnormal pixel points are reconstructed according to the spatial and temporal positions of the abnormal pixel points. This method has a strong correlation with the accuracy of abnormal point detection, and the accuracy is often affected by the accuracy of the detected image, resulting in a large difference. Therefore, the existing method for reconstructing the NDVI image has the problem of low reconstruction accuracy. SUMMARY

[0003] The purpose of the present application is to provide a NDVI image reconstruction method, system, device and storage medium, which improves the reconstruction accuracy of the NDVI image.

[0004] To achieve the above purpose, the present application provides the following scheme:

[0005] A NDVI image reconstruction method, the method comprises:

[0006] Obtain an original normalized difference vegetation index time series image; the normalized difference vegetation index time series image comprises normalized difference vegetation index images of vegetation to be monitored in different periods; the normalized difference vegetation index time series image comprises a plurality of pixels, and each pixel has a corresponding normalized difference vegetation index actual value in different periods;

[0007] Determine the normalized difference vegetation index image of the pixel with cloud pollution in the original normalized difference vegetation index time series image as a target image;

[0008] According to the period number of the corresponding period of the target image and the cloud pollution prediction value calculation formula, the normalized difference vegetation index prediction value of the pixel with cloud pollution is obtained; the cloud pollution prediction value calculation formula is determined by using the least square method based on the historical normalized difference vegetation index time series image and the period number of the corresponding period;

[0009] Replace the actual normalized difference vegetation index value of the pixel with cloud pollution with the predicted normalized difference vegetation index value to obtain a reconstructed target image;

[0010] Determine a reconstructed normalized difference vegetation index time series image according to the reconstructed target image and the original normalized difference vegetation index time series image.

[0011] Optionally, the determining process of the cloud pollution prediction value calculation formula comprises:

[0012] Construct an initial expression of the cloud pollution prediction value calculation formula; the parameters prepared by the cloud pollution prediction value calculation formula comprise a first coefficient, a second coefficient, a predicted normalized difference vegetation index value, and a period number;

[0013] Obtain a historical normalized difference vegetation index time series image of pixels without cloud pollution; the historical normalized difference vegetation index time series image is a historical normalized difference vegetation index image of N periods of the entire growth cycle of the vegetation for training; N>1;

[0014] Determine all historical normalized difference vegetation indexes corresponding to each pixel based on the current image; wherein, the initial image is the historical normalized difference vegetation index time series image;

[0015] Determine a current maximum normalized difference vegetation index of each pixel and a current maximum period number of the corresponding period;

[0016] Image synthesize all current maximum normalized difference vegetation indexes to obtain a current maximum normalized difference vegetation index image, and image synthesize all current maximum period numbers to obtain a current maximum period number image;

[0017] Add the current maximum normalized difference vegetation index image to a first parameter determination data set, and add the current maximum period number image to a second parameter determination data set;

[0018] Delete all current maximum normalized difference vegetation indexes in the current image;

[0019] Determine whether there is a historical normalized difference vegetation index in the current image;

[0020] If yes, return to the step of determining all historical normalized difference vegetation indexes corresponding to each pixel based on the current image;

[0021] If no, determine the first coefficient and the second coefficient based on the first parameter determination data set and the second parameter determination data set by using the least square method, so as to determine the cloud pollution prediction value calculation formula.

[0022] Optionally, the expression of the cloud pollution prediction value calculation formula is:

[0023]

[0024] wherein NDVI is a normalized difference vegetation index predicted value, a is a first coefficient, b is a second coefficient, and T is a period number.

[0025] A reconstruction system of an NDVI image, the system comprising:

[0026] an original image acquisition module configured to acquire an original normalized difference vegetation index time series image; the normalized difference vegetation index time series image comprises normalized difference vegetation index images of vegetation to be monitored in different periods; the normalized difference vegetation index time series image comprises a plurality of pixels, and each pixel has a corresponding normalized difference vegetation index actual value in different periods;

[0027] a target image determination module configured to determine normalized difference vegetation index images of pixels with cloud pollution in the original normalized difference vegetation index time series image as target images;

[0028] a predicted value calculation module configured to obtain normalized difference vegetation index predicted values of the pixels with cloud pollution according to a period number of a corresponding period of the target images and a cloud pollution predicted value calculation formula; the cloud pollution predicted value calculation formula is determined based on historical normalized difference vegetation index time series images and the period number of the corresponding period by using a least square method;

[0029] a target image reconstruction module configured to replace normalized difference vegetation index actual values of the pixels with cloud pollution with normalized difference vegetation index predicted values to obtain a reconstructed target image;

[0030] a reconstructed image determination module configured to determine a reconstructed normalized difference vegetation index time series image according to the reconstructed target image and the original normalized difference vegetation index time series image.

[0031] Optionally, the predicted value calculation module comprises a cloud pollution predicted value calculation formula determination submodule, and the cloud pollution predicted value calculation formula determination submodule comprises:

[0032] an initial expression construction unit configured to construct an initial expression of the cloud pollution predicted value calculation formula; parameters of the cloud pollution predicted value calculation formula include a first coefficient, a second coefficient, a normalized difference vegetation index predicted value, and a period number;

[0033] a historical image acquisition unit configured to acquire historical normalized difference vegetation index time series images of pixels without cloud pollution, wherein the historical normalized difference vegetation index time series images are historical normalized difference vegetation index images of N periods of the whole growth cycle of vegetation for training, and N>1;

[0034] a historical index determination unit configured to determine all historical normalized difference vegetation indexes corresponding to each pixel based on the current image, wherein the initial image is the historical normalized difference vegetation index time series images;

[0035] a maximum index and period number determination unit configured to determine the current maximum normalized difference vegetation index of each pixel and the current maximum period number of the corresponding period;

[0036] a synthesis unit configured to synthesize all current maximum normalized difference vegetation indexes to obtain a current maximum normalized difference vegetation index image, and to synthesize all current maximum period numbers to obtain a current maximum period number image;

[0037] a data set determination unit configured to add the current maximum normalized difference vegetation index image to a first parameter determination data set, and to add the current maximum period number image to a second parameter determination data set;

[0038] a deletion unit configured to delete all current maximum normalized difference vegetation indexes in the current image;

[0039] a judgment unit configured to judge whether there is a historical normalized difference vegetation index in the current image;

[0040] a first execution unit configured to return to "determine all historical normalized difference vegetation indexes corresponding to each pixel based on the current image" if yes;

[0041] a second execution unit configured to determine the cloud pollution prediction value calculation formula by using a least square method based on the first parameter determination data set and the second parameter determination data set if no.

[0042] An apparatus, comprising:

[0043] one or more processors;

[0044] a memory device having one or more programs stored thereon;

[0045] the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the reconstruction method of the NDVI image as described above.

[0046] A storage medium has a computer program stored thereon, wherein the computer program is executed by a processor to implement the reconstruction method of the NDVI image as described above.

[0047] According to the specific embodiments of the present application, the following technical effects are disclosed.

[0048] The application discloses a reconstruction method, system and device of NDVI images and a storage medium. Compared with the traditional reconstruction method of NDVI images, the present application uses the least square method and the cloud pollution prediction value calculation formula determined based on the historical normalized difference vegetation index time series images and the period number of the corresponding period, thereby improving the reconstruction accuracy of the NDVI images. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor under the premise of the drawings.

[0050] Figure 1 The reconstruction method of NDVI images provided for the embodiment 1 of the present application is shown in the flowchart.

[0051] Figure 2 The original NDVI image of the experimental area in the first period with cloud pollution is shown in the figure.

[0052] Figure 3 The reconstructed NDVI image of the experimental area in the first period is shown in the figure.

[0053] Figure 4 The original NDVI image of the experimental area in the second period with cloud pollution is shown in the figure.

[0054] Figure 5 The reconstructed NDVI image of the experimental area in the second period is shown in the figure.

[0055] Figure 6 The reconstruction system structure diagram of NDVI images provided for the embodiment 2 of the present application is shown in the figure. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0057] The present application aims to provide a NDVI image reconstruction method, system, device and storage medium, aiming to improve the reconstruction accuracy of NDVI image.

[0058] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0059] Embodiment 1

[0060] Figure 1 The NDVI image reconstruction method flowchart provided by Embodiment 1 of the present application is shown in FIG. 1. As shown in FIG. 1, the NDVI image reconstruction method in the present embodiment comprises: Figure 1

[0061] Step 101: obtaining an original normalized difference vegetation index time series image; the normalized difference vegetation index time series image comprises normalized difference vegetation index images of vegetation to be monitored in different periods; the normalized difference vegetation index time series image comprises a plurality of pixels, and each pixel has a corresponding normalized difference vegetation index actual value in different periods.

[0062] Step 102: determining the normalized difference vegetation index image of the pixel with cloud pollution in the original normalized difference vegetation index time series image as a target image.

[0063] Step 103: obtaining the normalized difference vegetation index predicted value of the pixel with cloud pollution according to the period number of the corresponding period of the target image and the cloud pollution prediction value calculation formula; the cloud pollution prediction value calculation formula is determined based on the historical normalized difference vegetation index time series image and the period number of the corresponding period by using the least square method.

[0064] Step 104: replacing the normalized difference vegetation index actual value of the pixel with cloud pollution with the normalized difference vegetation index predicted value to obtain a reconstructed target image.

[0065] Step 105: determining the reconstructed normalized difference vegetation index time series image according to the reconstructed target image and the original normalized difference vegetation index time series image.

[0066] ​As an optional implementation, the process of determining the cloud pollution prediction value calculation formula comprises:

[0067] An initial expression of the cloud pollution prediction value calculation formula is constructed; parameters for preparing the cloud pollution prediction value calculation formula include a first coefficient, a second coefficient, a normalized difference vegetation index prediction value, and a period number.

[0068] A historical normalized difference vegetation index time series image of a pixel without cloud pollution is obtained; the historical normalized difference vegetation index time series image is a historical normalized difference vegetation index image of N periods of a whole growth cycle of vegetation for training; N>1.

[0069] Specifically, the historical normalized difference vegetation index time series image composed of the historical normalized difference vegetation index images of the N periods has N bands, and early images are placed in front bands, and are sorted in order from early to late; for example, when the growth cycle is from June to July, the first image is on June 1, and the first period refers to the image on June 1; if the image collection time interval is 5 days, the second period is on June 6.

[0070] All historical normalized difference vegetation indexes corresponding to each pixel are determined based on a current image; the initial image is the historical normalized difference vegetation index time series image.

[0071] A current maximum normalized difference vegetation index of each pixel and a corresponding current maximum period number are determined.

[0072] Image synthesis is performed on all current maximum normalized difference vegetation indexes to obtain a current maximum normalized difference vegetation index image, and image synthesis is performed on all current maximum period numbers to obtain a current maximum period number image.

[0073] Specifically, (1) for each pixel, all NDVI values (for the same pixel, NDVI values may be different in different periods, and need to be calculated) are traversed to find the maximum value and record the image period number corresponding to the maximum value; each pixel point is traversed to obtain the NDVI maximum value of each pixel point in all periods and the image period number corresponding thereto; image synthesis is performed by using ENVI software to obtain two band values, the first band is the NDVI maximum value, and the second band is the period number, that is, two images are obtained, one image contains the NDVI maximum value, and the other image contains the period number corresponding to the maximum value.

[0074] (2)Cyclic synthesis of NDVI maximum value: remove the previous maximum value (remove the maximum value found for each pixel in the original image of all periods), according to the method of (1), again to carry out the synthesis of NDVI maximum value, and constantly cycle (until the complete traversal of all periods), will get a plurality of NDVI maximum value image and period image.

[0075] (3) Band synthesis of multiple maximum value images (output an image with 2N bands, select the maximum value in the subsequent parameter inversion): N period image, the first N bands are NDVI maximum value sorting, the last N bands are maximum value corresponding period, process the synthesized image, and set the NDVI value less than 0 to 0.

[0076] Add the current maximum normalized difference vegetation index image to the first parameter determination data set, and add the current maximum period image to the second parameter determination data set.

[0077] Delete all current maximum normalized difference vegetation index in the current image.

[0078] Determine whether there is a historical normalized difference vegetation index in the current image.

[0079] If yes, return to "determine all historical normalized difference vegetation indexes corresponding to each pixel based on the current image".

[0080] If not, determine the first coefficient and the second coefficient based on the first parameter determination data set and the second parameter determination data set by using the least square method, so as to determine the cloud pollution prediction value calculation formula.

[0081] Specifically, all normalized difference vegetation indexes of all current maximum normalized difference vegetation index images in the first parameter determination data set are taken as the ordinate, and all current maximum period images in the second parameter determination data set are taken as the abscissa, the normalized difference vegetation index and the period are one-to-one corresponding, and the period-index curve is drawn, for example, one point of the curve is recorded as (T1, NDVI1). To ensure the increment, (T2, NDVI2) needs to satisfy T2

[0082] The determination formula of a and b is:

[0083]

[0084]

[0085] This led to the determination of the formula for calculating cloud pollution prediction values.

[0086] As an optional implementation method, the formula for calculating cloud pollution prediction values ​​is expressed as follows:

[0087]

[0088] Wherein, NDVI is the predicted value of the Normalized Difference Vegetation Index, a is the first coefficient, b is the second coefficient, and T is the number of periods.

[0089] Using a certain region in Heilongjiang Province, China as an experimental area, such as Figures 2-5 As shown, Figure 2 The original NDVI image of the experimental area during the first period, showing cloud contamination. Figure 3 The image shows the reconstructed NDVI image of the experimental area in the first phase. Figure 4 The original NDVI image of the experimental area with cloud contamination in the second period; Figure 5 This is the reconstructed NDVI image of the experimental area in the second phase. (Comparison) Figure 2 and Figure 3 ,contrast Figure 4 and Figure 5 As can be seen, the normalized difference vegetation index image reconstructed using the method in this embodiment is clearer.

[0090] Example 2

[0091] Figure 6 This is a schematic diagram of the NDVI image reconstruction system provided in Embodiment 2 of the present invention. Figure 6 As shown, the NDVI image reconstruction system in this embodiment includes:

[0092] The original image acquisition module 201 is used to acquire the original normalized differential vegetation index (NDVI) time series images. The NDVI time series images include NDVI images of the vegetation to be monitored at different times. The NDVI time series images include multiple pixels, and each pixel has a corresponding actual value of the NDVI at different times.

[0093] The target image determination module 202 is used to determine the normalized difference vegetation index image of the original normalized difference vegetation index time series image containing cloud pollution as the target image.

[0094] The prediction value calculation module 203 is used to obtain the normalized difference vegetation index prediction value of the pixels with cloud pollution based on the period number of the target image and the cloud pollution prediction value calculation formula. The cloud pollution prediction value calculation formula is determined by using the least squares method based on the historical normalized difference vegetation index time series image and the period number of the corresponding period.

[0095] The target image reconstruction module 204 is used to replace the actual value of the normalized difference vegetation index of pixels with cloud pollution with the predicted value of the normalized difference vegetation index to obtain the reconstructed target image.

[0096] The reconstructed image determination module 205 is used to determine the reconstructed normalized difference vegetation index time series image based on the reconstructed target image and the original normalized difference vegetation index time series image.

[0097] Optionally, the prediction calculation module 203 includes a cloud pollution prediction calculation formula determination submodule, which includes:

[0098] The initial expression construction unit is used to construct the initial expression of the cloud pollution prediction value calculation formula; the parameters prepared by the cloud pollution prediction value calculation formula include: the first coefficient, the second coefficient, the normalized difference vegetation index prediction value, and the period number.

[0099] The historical image acquisition unit is used to acquire historical normalized difference vegetation index time series images of pixels without cloud pollution; the historical normalized difference vegetation index time series images are historical normalized difference vegetation index images of N periods of the entire growth cycle of the training vegetation; N>1.

[0100] The historical index determination unit is used to determine all historical normalized differential vegetation indices corresponding to each pixel based on the current image; wherein, the initial image is a time series image of historical normalized differential vegetation indices.

[0101] The maximum index and period number determination unit is used to determine the current maximum normalized difference vegetation index and the current maximum period number of each pixel.

[0102] The image synthesis unit is used to synthesize images of all current maximum normalized difference vegetation indices to obtain images of the current maximum normalized difference vegetation index, and to synthesize images of all current maximum period numbers to obtain images of the current maximum period number.

[0103] The dataset determination unit is used to add the image with the current maximum normalized difference vegetation index to the first parameter determination dataset and the image with the current maximum period to the second parameter determination dataset.

[0104] The deletion unit is used to delete all current maximum normalized difference vegetation indices in the current image;

[0105] The judgment unit is used to determine whether the historical normalized difference vegetation index exists in the current image.

[0106] The first execution unit is used to return to the historical index determination unit if the condition is met.

[0107] The second execution unit is used to determine the cloud pollution prediction value calculation formula by using the least squares method to determine the dataset based on the first parameter and the second parameter, and to determine the first coefficient and the second coefficient, if not.

[0108] Example 3

[0109] An apparatus comprising:

[0110] One or more processors.

[0111] A storage device on which one or more programs are stored.

[0112] When one or more programs are executed by one or more processors, the one or more processors implement the NDVI image reconstruction method as described in Example 1.

[0113] Example 4

[0114] A storage medium storing a computer program, wherein the computer program, when executed by a processor, implements a method for reconstructing NDVI images as described in Example 1.

[0115] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0116] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for reconstructing NDVI images, characterized in that, The method includes: Obtain the original normalized difference vegetation index (NDVI) time series image; the NDVI time series image includes NDVI images of the vegetation to be monitored at different periods; the NDVI time series image includes multiple pixels, and each pixel has a corresponding actual value of NDVI at different periods. The normalized difference vegetation index (NDVI) image containing cloud-polluted pixels in the original normalized difference vegetation index time series image is identified as the target image. Based on the period number corresponding to the target image and the cloud pollution prediction value calculation formula, the normalized difference vegetation index (NDVI) prediction value of the pixels with cloud pollution is obtained; the cloud pollution prediction value calculation formula is determined using the least squares method based on historical normalized difference vegetation index time series images and the corresponding period number; the process of determining the cloud pollution prediction value calculation formula includes: An initial expression for the cloud pollution prediction calculation formula is constructed; the parameters prepared by the cloud pollution prediction calculation formula include: a first coefficient, a second coefficient, a normalized difference vegetation index prediction value, and a period number; Obtain historical normalized difference vegetation index time series images of pixels without cloud pollution; the historical normalized difference vegetation index time series images are historical normalized difference vegetation index images of N periods of the entire growth cycle of the vegetation used for training; N>1; Based on the current image, determine all historical normalized difference vegetation indices corresponding to each pixel; where the initial image is a time series image of historical normalized difference vegetation indices. Determine the current maximum normalized difference vegetation index and the current maximum period number for each pixel; Image synthesis is performed on all current maximum normalized difference vegetation indices to obtain the current maximum normalized difference vegetation index image, and image synthesis is performed on all current maximum period numbers to obtain the current maximum period number image; The image with the current maximum normalized difference vegetation index is added to the dataset determined by the first parameter, and the image with the current maximum period number is added to the dataset determined by the second parameter. Delete all current maximum normalized difference vegetation indices in the current image; Determine whether the historical normalized difference vegetation index exists in the current image; If so, return "Determine all historical normalized differential vegetation indices corresponding to each pixel based on the current image"; If not, then the least squares method is used to determine the dataset based on the first parameter and the second parameter, and to determine the first coefficient and the second coefficient, thereby determining the calculation formula for the cloud pollution prediction value; The actual values ​​of the Normalized Differential Vegetation Index (NDVI) of the pixels with cloud pollution are replaced with the predicted values ​​of the NDVI to obtain the reconstructed target image. Based on the reconstructed target image and the original normalized difference vegetation index time series image, the reconstructed normalized difference vegetation index time series image is determined.

2. The method for reconstructing NDVI images according to claim 1, characterized in that, The expression for the formula used to calculate the cloud pollution prediction value is as follows: ； in, This is the predicted value of the normalized difference vegetation index. As the first coefficient, As the second coefficient, Number of periods.

3. A reconstruction system for NDVI images, characterized in that, The system includes: The original image acquisition module is used to acquire the original normalized differential vegetation index (NDVI) time series images; the NDVI time series images include NDVI images of the vegetation to be monitored at different periods; the NDVI time series images include multiple pixels, and each pixel has a corresponding actual value of the NDVI at different periods. The target image determination module is used to determine the normalized difference vegetation index image of the pixels with cloud pollution in the original normalized difference vegetation index time series image as the target image. The prediction value calculation module is used to obtain the normalized difference vegetation index prediction value of the pixels with cloud pollution based on the period number of the target image and the cloud pollution prediction value calculation formula; the cloud pollution prediction value calculation formula is determined by using the least squares method based on the historical normalized difference vegetation index time series image and the period number of the corresponding period. The prediction calculation module includes a cloud pollution prediction calculation formula determination submodule, which includes: An initial expression construction unit is used to construct an initial expression for the cloud pollution prediction calculation formula; the parameters prepared by the cloud pollution prediction calculation formula include: a first coefficient, a second coefficient, a normalized difference vegetation index prediction value, and a period number; The historical image acquisition unit is used to acquire historical normalized differential vegetation index time series images of pixels without cloud pollution; the historical normalized differential vegetation index time series images are historical normalized differential vegetation index images of N periods of the entire growth cycle of the training vegetation; N>1. The historical index determination unit is used to determine all historical normalized differential vegetation indices corresponding to each pixel based on the current image; wherein, the initial image is a time series image of historical normalized differential vegetation indices; The maximum index and period number determination unit is used to determine the current maximum normalized difference vegetation index and the current maximum period number of each pixel. The image synthesis unit is used to synthesize images of all current maximum normalized difference vegetation indices to obtain images of the current maximum normalized difference vegetation index, and to synthesize images of all current maximum period numbers to obtain images of the current maximum period number. The dataset determination unit is used to add the image with the current maximum normalized difference vegetation index to the dataset determined by the first parameter, and to add the image with the current maximum period number to the dataset determined by the second parameter. The deletion unit is used to delete all current maximum normalized difference vegetation indices in the current image; The judgment unit is used to determine whether the historical normalized difference vegetation index exists in the current image; The first execution unit is used to return "determine all historical normalized differential vegetation indices corresponding to each pixel based on the current image" if the condition is met. The second execution unit is used to determine the cloud pollution prediction value calculation formula by using the least squares method to determine the dataset based on the first parameter and the second parameter, if not, and to determine the first coefficient and the second coefficient. The target image reconstruction module is used to replace the actual value of the Normalized Differential Vegetation Index (NDVI) of the pixels with cloud pollution with the predicted value of the NDVI, so as to obtain the reconstructed target image. The reconstructed image determination module is used to determine the reconstructed normalized difference vegetation index time series image based on the reconstructed target image and the original normalized difference vegetation index time series image.

4. A device, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the NDVI image reconstruction method as described in any one of claims 1 to 2.

5. A storage medium, characterized in that, It stores a computer program, wherein the computer program, when executed by a processor, implements the NDVI image reconstruction method as described in any one of claims 1 to 2.

Images