A method and system for temporary land interpretation and reclamation verification

Abstract

This invention belongs to the field of natural resource management technology and discloses a method and system for interpreting temporary land use and verifying land reclamation. Utilizing the GEE cloud computing platform and Sentinel-2 satellite remote sensing imagery, it calculates and generates typical time-series characteristic curves of remote sensing characteristic indices for both temporary and non-temporary land use. It then calculates the time-series characteristic curve of the remote sensing characteristic index for the plot to be judged, and calculates the DTW similarity distance between the time-series characteristic curve of the plot and the typical time-series characteristic curves of temporary and non-temporary land use to determine whether the plot is temporary land use. For temporary land whose reclamation period has expired, it calculates the slope of the time-series characteristic curve based on the remote sensing characteristic indices at the time of expiration of the use period and the time of reclamation expiration to determine whether the land reclamation has been completed. This invention can significantly reduce the time and labor costs of manually visually interpreting and verifying land reclamation after the expiration of the temporary land use period, and improve the government's regulatory level and decision-making analysis capabilities.

Description

Technical Field

[0001] This invention belongs to the field of natural resource management technology, and relates to a method and system for interpreting and verifying temporary land use and reclamation, particularly to a technology and method for monitoring and supervising temporary land use based on satellite remote sensing data. Background Technology

[0002] Temporary land use is an important mode of land use in my country. It refers to state-owned land or land owned by farmers' collectives that needs to be used temporarily for construction projects and geological exploration. It provides great convenience for engineering construction and project implementation, playing a positive role in ensuring the smooth progress of construction projects and geological exploration. The usage period for temporary land is generally within two years, but can be up to four years in special circumstances. For temporary land use that occupies cultivated land, forest land, or agricultural land, failure to promptly reclaim the land after the expiration of the usage period can lead to land resource loss and damage to cultivated land. Traditional methods of supervising temporary land use mainly rely on land inspection and enforcement. After on-site inspections, enforcement opinions are issued to the project user unit or competent department, which then supervises rectification. This method is difficult to comprehensively, timely, accurately, and objectively supervise temporary land use projects.

[0003] Satellite remote sensing boasts advantages such as rapid response, wide coverage, short cycle time, high timeliness, and repeatable observation capabilities. It can continuously track, monitor over large areas, and accurately identify various land cover types, providing data support for the monitoring and supervision of temporary land use. However, due to the significant differences in the usage cycles and the complexity of temporary land use types, current methods for interpreting and verifying temporary land use and land reclamation using remote sensing technology generally rely on manual visual interpretation, and automation of these processes has not yet been achieved.

[0004] Based on the above analysis, the existing technologies have the following problems and shortcomings: Due to the large differences in reclamation cycles and the complexity of temporary land use types, current remote sensing interpretation and reclamation verification of temporary land use rely on manual visual interpretation and have not yet been automated. Traditional methods of temporary land use supervision are insufficient for comprehensive, timely, accurate, and objective monitoring of temporary land use projects. Manual visual remote sensing interpretation requires significant manpower and time, and is susceptible to subjective human factors that can lead to inconsistent standards and errors. Summary of the Invention

[0005] To overcome the difficulties in land reclamation verification during temporary land use supervision, this invention discloses a method and system for interpreting temporary land use and verifying reclamation, particularly a method and system based on time-series satellite remote sensing imagery. The technical solution is as follows:

[0006] This invention is implemented as follows: a method for interpreting temporary land use and verifying land reclamation, comprising:

[0007] S1. Using the GEE cloud computing platform and Sentinel-2 satellite remote sensing images, calculate and generate sample features of typical time-series characteristic curves of remote sensing characteristic indices for temporary and non-temporary land use.

[0008] S2. Calculate the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged. After linear interpolation or linear compression, calculate the DTW similarity distance between the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged and the typical time series characteristic curves of temporary land and non-temporary land. Based on the DTW distance criterion, determine whether the land parcel is temporary land.

[0009] S3. For temporary land use that has reached the end of its reclamation period, calculate the slope of the time-series characteristic curve based on the remote sensing characteristic index of the time of use expiration and the time of reclamation expiration to determine whether the temporary land use has completed land reclamation.

[0010] Before step S1, the Sentinel-2 satellite remote sensing images undergo preprocessing including image screening, cloud removal, and image cropping; temporary and non-temporary land use samples are acquired, and time series information of the entire life cycle of temporary land use is generated; the Sentinel-2 satellite remote sensing image screening is based on the GEE platform, which sets the date according to the vector range of temporary land use, the approval time, and the expiration time, and filters image data with cloud cover of less than 20% during the temporary land use period and the land reclamation period, and uses the QA60 band to perform cloud removal processing on the acquired Sentinel-2 satellite remote sensing images respectively.

[0011] In step S1, sample features of typical time-series characteristic curves of remote sensing feature indices for temporary and non-temporary land use are calculated and generated, including:

[0012] Based on the vector range shp and kml files of temporary land use, the temporal NDVI, NDBI and NDSI remote sensing characteristic indices of temporary land use during the land use cycle, as well as the annual temporal NDVI, NDBI and NDSI remote sensing characteristic indices of non-temporary land use, were obtained using the GEE cloud platform and Sentinel-2 satellite remote sensing imagery.

[0013] By performing maximum-minimum normalization on the time-series characteristic curves, the time-series MMNDVI, MMNDBI, and MMNDSI values ​​of typical temporary and non-temporary land use patches are obtained.

[0014] Sample characteristics of typical time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, for both temporary and non-temporary land use categories, were obtained by calculating the mean.

[0015] Furthermore, the temporal NDVI, NDBI, and NDSI remote sensing characteristic indices of temporary land use during the land use cycle are obtained, including:

[0016] The Sentinel-2 satellite sensor includes 12 spectral bands. By using band operations, the Normalized Difference Vegetation Index (NDVI), Normalized Building Index (NDBI), and Normalized Soil Index (NDSI) are calculated to obtain the NDVI, NDBI, and NDSI values ​​of satellite remote sensing data at different time phases.

[0017] The calculation expressions for NDVI, NDBI, and NDSI are as follows:

[0018]

[0019] In the formula, NIR is the near-infrared band surface reflectance of Sentinel-2 satellite remote sensing image, RED is the red band surface reflectance of Sentinel-2 satellite remote sensing image, SWIR is the short-infrared band surface reflectance of Sentinel-2 satellite remote sensing image, and MIR is the mid-infrared band surface reflectance of Sentinel-2 satellite remote sensing image.

[0020] Furthermore, the temporary land use area in the Sentinel-2 satellite remote sensing image includes multiple pixels, and the NDVI, NDBI, and NDSI values ​​calculated within the vector range are averaged.

[0021] For the time-series NDVI, NDBI, and NDSI remote sensing characteristic curves generated for each plot, the maximum and minimum normalized vegetation index (MMNDVI), maximum and minimum normalized building index (MMNDBI), and maximum and minimum normalized soil index (MMNDSI) for each plot in each time phase are obtained using maximum and minimum normalization.

[0022] The expression for the MMNDVI characteristic index is as follows:

[0023]

[0024] In the formula, MMNDVI represents the maximum and minimum normalized vegetation index, NDVI represents the normalized vegetation index at a certain time; max1 represents the maximum value in the NDVI time series data, and min1 represents the minimum value in the NDVI time series data; through maximum and minimum normalization processing, the values ​​of all calculated MMNDVI remote sensing feature indices are between 0 and 1.

[0025] The expression for the MMNDBI feature index is as follows:

[0026]

[0027] In the formula, MMNDBI represents the maximum and minimum normalized building index, NDBI represents the normalized building index at a certain time; max2 represents the maximum value in the NDBI time series data, and min2 represents the minimum value in the NDBI time series data; through maximum and minimum normalization processing, the values ​​of all calculated MMNDBI remote sensing feature indices are between 0 and 1.

[0028] The expression for the MMNDSI characteristic index is as follows:

[0029]

[0030] In the formula, MMNDSI represents the maximum and minimum normalized soil index, NDSI represents the normalized soil index at a certain time, max3 represents the maximum value in the NDSI time series data, and min3 represents the minimum value in the NDSI time series data. Through maximum and minimum normalization processing, the values ​​of all calculated MMNDSI remote sensing feature indices are between 0 and 1.

[0031] Based on the determined temporary land use vector, the temporary land use cycle, and the calculated NDVI, NDBI, and NDSI remote sensing characteristic indices for each patch, and after mean processing and maximum-minimum normalization processing, the time-series characteristic curves of MMNDVI, MMNDBI, and MMNDSI for the temporary land use category are plotted to obtain typical time-series sample characteristics of temporary land use; 30 sample points of non-temporary land use are randomly selected, and the time-series characteristic curves of MMNDVI, MMNDBI, and MMNDSI for non-temporary land use throughout the year are plotted.

[0032] In step S2, determining whether a plot of land is for temporary use based on the DTW distance criterion includes:

[0033] Calculate three remote sensing characteristic indices, including NDVI, NDBI, and NDSI, for each time phase of the land use patch to be judged, and generate time series characteristic curves of the three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, for the land use patch to be judged.

[0034] By comparing the DTW distance similarity between the time series characteristic curves of the three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, of the plot to be determined and the time series characteristic curves of typical temporary land use and non-temporary land use, it is determined whether the plot to be determined is temporary land use.

[0035] Furthermore, after obtaining typical temporal sample characteristics of non-temporary land use from the temporal characteristic curves of the three remote sensing feature indices, the similarity is calculated by combining the temporal characteristic curves of the three remote sensing feature indices and using the Dynamic Time Warping (DTW) method to determine whether the plot to be determined is temporary land use.

[0036] For two time series X n ={c1,c2…c n} and Y m ={q1,q2…q m Given two time series with lengths |X| and Y|, calculate the similarity between them and determine the regularized path W = W1, W2, ..., W... K Where Max(|X|,|Y|)≤K≤|X|+|Y|, W K The form is (i,j), where i represents the i-th coordinate in X and j represents the j-th coordinate in Y; the normalization path W starts from W1 = (1,1) and ends at W1 = (|X|,|Y|) to ensure that every coordinate in X and Y appears in W; where c n For the first time series X n The value of the nth phase in the equation, q m For the second time series Y m The value of the m-th phase in W K Let X be the Kth point in the path, Max(|X|,|Y|) be the larger of the two time series lengths, and |X|+|Y| be the sum of the two time series lengths.

[0037] In a regular path W, the i and j of W(i,j) are monotonically increasing, represented as:

[0038] W K =(i,j)

[0039] W K+1 =(i′,j′)

[0040] i≤i′≤i+1,j≤j′≤j+1

[0041] In the formula, i and j represent the position indices of the Kth point in the first and second time series, respectively; W K+1 Let i' be the (K+1)th point in the path, and let i' and j' be the position indices of the (K+1)th point in the first and second time series, respectively. i≤i'≤i+1 and j≤j'≤j+1 indicate that the movement at each step in the path is monotonically increasing.

[0042] The shortest optimal reduction path obtained is the DTW distance, which is:

[0043] D(i,j)=Dist(i,j)+min[D(i-1,j),D(i,j-1),D(i-1,j-1)]

[0044] In the formula, D(i,j) is the cumulative DTW distance between the time series at indices i and j, Dist(i,j) is the local Euclidean distance between the i-th point in the first time series and the j-th point in the second time series, min[] represents the minimum cumulative distance among the three possible paths, i.e., the minimum cumulative distance from the top, left, or diagonal direction, D(i-1,j) is the cumulative DTW distance between the time series at indices i-1 and j, D(i,j-1) is the cumulative DTW distance between the time series at indices i and j-1, and D(i-1,j-1) is the cumulative DTW distance between the time series at indices i-1 and j-1.

[0045] Calculate the DTW distance between the time series characteristic curves of the three remote sensing characteristic indices of the pending land parcel and the time series characteristic curves of the three remote sensing characteristic indices of temporary land and non-temporary land. If the DTW distance with the temporary land is smaller, the pending land parcel is determined to be temporary land; otherwise, it is determined to be non-temporary land.

[0046] In step S3, determining whether the temporary land use has been reclaimed includes:

[0047] Based on the time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, for temporary land use, the slopes of the three remote sensing characteristic indices at the time of expiration of the temporary land use period and the time of expiration of the reclamation period are calculated respectively; where the expiration of the reclamation period refers to the expiration of one year of use; by combining the differences in the slopes of the three remote sensing characteristic indices, relevant information on whether the temporary land use has completed land reclamation upon the expiration of the reclamation period is determined.

[0048] Furthermore, information related to determining whether temporary land use that has reached the end of its reclamation period has been reclaimed includes: for patches of temporary land use that have reached the end of their reclamation period, the slopes of three characteristic curves are calculated using three remote sensing characteristic indices at the end of the use period and the end of the reclamation period, respectively. The calculation expressions are as follows:

[0049]

[0050] In the formula, K MMNDVI ,K MMNDBI ,K MMNDSI The curvatures of the time-series characteristic curves for the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively; MMNDVI n ,MMNDBI n ,MMNDSI n The values ​​at the end of the reclamation period on the time-series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI are respectively. m ,MMNDBI m,MMNDSI m , where m and m are the values ​​of the time of use expiration on the time series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively, and m and m are the times of reclamation expiration and use expiration, respectively.

[0051] The slope values ​​of the time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, are used to determine whether land reclamation has been completed.

[0052] When K is satisfied MMNDVI >δ1,K MMNDBI <δ2,K MMNDSI <δ3 indicates that the land reclamation has been completed and the reclamation effect is good;

[0053] When K is satisfied MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI ≥δ3 indicates that land reclamation has not yet been completed for this plot;

[0054] When K cannot be satisfied simultaneously MMNDVI >δ1,K MMNDBI <δ2,K MMNDSI <δ3 or K MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI When the value is ≥δ3, the interpretation is switched to manual.

[0055] Among them, δ1, δ2, and δ3 are the thresholds of the slopes of three remote sensing feature indices used to determine whether land reclamation has been completed, which are statistically derived from 30 temporary land samples that have completed land reclamation. In practical applications, these indices are obtained by statistical calculation based on temporary land samples from different regions.

[0056] Another object of the present invention is to provide a temporary land use interpretation and reclamation verification system, which is used to regulate the temporary land use interpretation and reclamation verification method. The system includes:

[0057] The time-series remote sensing index sample feature generation module is used to calculate and generate sample features of typical time-series feature curves of remote sensing feature indices for temporary and non-temporary land use using the GEE cloud computing platform and Sentinel-2 satellite remote sensing images.

[0058] The temporary land use judgment module is used to calculate the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged. After linear interpolation or linear compression, it calculates the DTW similarity distance between the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged and the typical time series characteristic curve of temporary land use and non-temporary land use. Based on the DTW distance criterion, it judges whether the land parcel is temporary land use.

[0059] The land reclamation judgment module is used to determine whether land reclamation has been completed for temporary land that has reached the end of its reclamation period. This is done by calculating the slope of the time-series characteristic curve based on the remote sensing characteristic index of the time of use expiration and the time of reclamation expiration.

[0060] Existing verification schemes for temporary land reclamation primarily rely on on-site inspections, while verification methods based directly on satellite remote sensing imagery or UAV remote sensing mainly depend on manual visual interpretation, requiring significant manpower and time costs. To address these issues, this invention offers the following advantages and positive effects: By analyzing the differences in the time-series characteristic curves of remote sensing feature indices between temporary and non-temporary land, and utilizing the differences in the remote sensing feature index characteristic curves of reclaimed and unreclaimed temporary land in time-series satellite remote sensing imagery, this invention achieves rapid screening of temporary land and quick, automated judgment of whether land reclamation has been completed. This overcomes the technical bottleneck of relying on manual visual interpretation when directly verifying temporary land based on satellite remote sensing imagery. Furthermore, this method effectively avoids the problem of subjective human inference leading to inaccurate verification results.

[0061] The satellite remote sensing imagery used in this method is Sentinel-2 optical remote sensing imagery, but it can also be extended to other optical satellite remote sensing data with red, near-infrared, short-infrared, and mid-infrared bands. Furthermore, due to limitations in the spatial resolution of Sentinel-2 optical satellite remote sensing imagery and the size of temporary land use areas, this method is more suitable for identifying temporary land use areas greater than 0.7 hectares and less than 8 hectares, and for the automatic verification of land reclamation.

[0062] This invention was validated using 50 temporary land parcels in Zhejiang Province that had been verified as having completed land reclamation. The accuracy rate for identifying temporary land parcels reached 84%, with an error rate of only 4%. The accuracy rate for identifying temporary land parcels undergoing land reclamation reached 86%. This demonstrates that the method of this invention can effectively identify temporary land parcels and determine whether land reclamation has been completed within the prescribed period, significantly reducing labor costs. This invention is a method for indirectly using satellite remote sensing imagery to screen whether land reclamation of temporary land parcels has been completed, providing a new technical approach for the monitoring and supervision of temporary land parcels.

[0063] This invention provides a novel methodology for verifying whether temporary land use patches have completed land reclamation, specifically addressing the occupation of arable land, forest land, grassland, and agricultural land. This significantly reduces the time cost of manually visually verifying land reclamation after the expiration of the temporary land use period. The method for temporary land use identification and reclamation verification based on time-series satellite remote sensing imagery utilizes this method to continuously track and monitor problems arising during the use of temporary land, particularly assessing land reclamation after the expiration of the temporary land use period. This effectively supplements the shortcomings of traditional regulatory models in terms of coverage and timeliness, saving substantial manpower and resources, and significantly improving government regulatory capabilities and decision-making abilities.

[0064] This invention automates and streamlines the verification process for temporary land use by utilizing remote sensing imagery and related algorithms. This significantly improves the efficiency and accuracy of temporary land use verification and allows for the acquisition of most temporary land use and reclamation verification results without manual verification, thus significantly reducing labor costs. It is expected to bring significant economic and social benefits, particularly in the fields of land resource management and environmental protection, where it has broad commercial application prospects.

[0065] This invention is the first to automate the interpretation of temporary land use and the verification of land reclamation based on time-series remote sensing images, and integrates a temporary land use verification system, filling the gap in the automation of temporary land use supervision and verification based on time-series remote sensing images.

[0066] Traditional temporary land use verification relies on manual on-site inspections, which are time-consuming, labor-intensive, and easily affected by human factors. By utilizing time-series remote sensing imagery and related algorithms, this invention constructs a temporary land use interpretation and reclamation verification system based on the GEE cloud platform. This system automates the process of temporary land use interpretation and land reclamation verification, significantly reducing labor and time costs and overcoming the limitations of traditional temporary land use verification techniques. Attached Figure Description

[0067] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure;

[0068] Figure 1 This is a flowchart of the temporary land use interpretation and reclamation verification method provided in the embodiments of the present invention;

[0069] Figure 2 This is a time-series characteristic curve of the remote sensing characteristic indices of MMNDVI, MMNDBI, and MMNDSI for the entire life cycle of temporary land use provided in this embodiment of the invention.

[0070] Figure 3This is a year-round time-series characteristic curve of the remote sensing characteristic indices MMNDVI, MMNDBI, and MMNDSI for non-temporary land use (vegetated area) provided in the embodiments of the present invention;

[0071] Figure 4 This is a flowchart of a method for determining temporary land use based on a time-series characteristic curve of three combined remote sensing feature indices, provided by an embodiment of the present invention.

[0072] Figure 5 This is a flowchart of a method provided by an embodiment of the present invention for determining whether land reclamation has been completed for temporary land after the reclamation period has expired, based on the slope of the temporal characteristic curve of three joint remote sensing characteristic indices. Detailed Implementation

[0073] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0074] The innovation of this invention lies in the following: This invention utilizes temporal Sentinel-2 optical remote sensing imagery and related algorithms to propose, for the first time, an automated judgment of whether temporary land use has been completed by combining three remote sensing feature indices (MMNDVI, MMNDBI, and MMNDSI) and the slope value of the temporal feature curve. Based on the GEE cloud computing platform, it has for the first time achieved automation of temporary land use interpretation and land reclamation verification, overcoming the problem of relying on manual interpretation for remote sensing verification of temporary land use in the past.

[0075] This invention is the first to combine remote sensing feature indices calculated using three formulas, including NDVI, NDBI, and NDSI, into the feature construction and judgment of temporary land use. Using these as criteria for judging temporary land use solves the current deficiency that temporary land use can only be judged manually.

[0076] Furthermore, since temporary land use areas in Sentinel-2 satellite remote sensing imagery typically include multiple pixels, it is necessary to average the NDVI, NDBI, and NDSI values ​​calculated within the vector area. To mitigate the influence of different geographical locations on remote sensing indices, for the time-series NDVI, NDBI, and NDSI remote sensing characteristic curves generated for each plot, the maximum and minimum normalized vegetation index (MMNDVI), maximum and minimum normalized building index (MMNDBI), and maximum and minimum normalized soil index (MMNDSI) for each plot in each time phase are obtained using maximum and minimum normalization.

[0077] Example 1, such as Figure 1 As shown, the temporary land use interpretation and reclamation verification method provided in this embodiment of the invention includes the following steps:

[0078] S1. Using the GEE cloud computing platform and Sentinel-2 satellite remote sensing images, calculate and generate sample features of typical time-series characteristic curves of remote sensing characteristic indices for temporary and non-temporary land use.

[0079] Based on the vector extent shp or kml files of temporary land use, the temporal NDVI, NDBI, and NDSI remote sensing characteristic indices of temporary land use within the land use cycle, as well as the annual temporal NDVI, NDBI, and NDSI remote sensing characteristic indices of non-temporary land use, are obtained using the GEE (Google Earth Engine, GEE) cloud platform and Sentinel-2 satellite remote sensing imagery. By performing max-min normalization on the temporal characteristic curves, the temporal MMNDVI, MMNDBI, and MMNDSI values ​​of typical temporary and non-temporary land use patches are obtained. Then, by calculating the mean, the sample characteristics of the typical temporal characteristic curves of the three remote sensing characteristic indices, MMNDVI, MMNDBI, and MMNDSI, for both temporary and non-temporary land use categories are obtained.

[0080] S2. Calculate the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged. After linear interpolation or linear compression, calculate the DTW similarity distance between the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged and the typical time series characteristic curves of temporary land and non-temporary land. Based on the DTW distance criterion, determine whether the land parcel is temporary land.

[0081] To determine whether a plot of land is for temporary use, three remote sensing characteristic indices (NDVI, NDBI, and NDSI) are calculated for each time phase of the plot. Temporal characteristic curves of the three remote sensing characteristic indices (MMNDVI, MMNDBI, and MMNDSI) for the plot are generated. By comparing the DTW distance similarity between the temporal characteristic curves of the three remote sensing characteristic indices of the plot and the temporal characteristic curves of typical temporary land use and non-temporary land use, the determination of whether the plot is for temporary use is made.

[0082] S3. For temporary land use that has reached the end of its reclamation period, calculate the slope of the time-series characteristic curve based on the remote sensing characteristic index of the time of use expiration and the time of reclamation expiration to determine whether the temporary land use has completed land reclamation.

[0083] To determine whether land reclamation has been completed for temporary land use after the reclamation period has expired, the slopes of the three remote sensing characteristic indices (MMNDVI, MMNDBI, and MMNDSI) are calculated based on their temporal characteristic curves. These indices are compared between the time of expiration of the temporary land use period and the time of expiration of the reclamation period (generally one year after the expiration of the use period). The differences in the slopes of the three remote sensing characteristic indices are then used to determine whether land reclamation has been completed for the temporary land use after the reclamation period has expired, and other relevant information.

[0084] Before step S1, preprocessing work such as image screening, cloud removal, and image cropping of Sentinel-2 satellite remote sensing images is required to obtain samples of temporary and non-temporary land use with high availability, and to generate time series information on the entire life cycle of temporary land use. Sentinel-2 satellite remote sensing image screening is based on the GEE platform. Dates are set according to the temporary land use vector range, approval time, and expiration time. Image data with cloud cover of less than 20% during the temporary land use period and land reclamation period are selected for subsequent calculations. Then, cloud removal processing is performed on the acquired Sentinel-2 satellite remote sensing images using the QA60 band to reduce the impact of cloud cover on data statistics.

[0085] In step S1, the remote sensing characteristic indices such as NDVI, NDBI, and NDSI of Sentinel-2 satellite remote sensing images for each time phase are calculated. The Sentinel-2 satellite sensor includes 12 spectral bands. The Normalized Differential Vegetation Index (NDVI), Normalized Differential Built-Up Index (NDBI), and Normalized Differential Soil Index (NDSI) are calculated using band operations to obtain the NDVI, NDBI, and NDSI values ​​of satellite remote sensing data for different time phases.

[0086] The calculation expressions for NDVI, NDBI, and NDSI are as follows:

[0087]

[0088] In the formula, NIR is the near-infrared band surface reflectance of Sentinel-2 satellite remote sensing image, RED is the red band surface reflectance of Sentinel-2 satellite remote sensing image, SWIR is the short-infrared band surface reflectance of Sentinel-2 satellite remote sensing image, and MIR is the mid-infrared band surface reflectance of Sentinel-2 satellite remote sensing image.

[0089] In addition, since the temporary land use area in Sentinel-2 satellite remote sensing imagery generally includes multiple pixels, it is necessary to average the NDVI, NDBI, and NDSI values ​​calculated within the vector range.

[0090] For the time-series NDVI, NDBI, and NDSI remote sensing characteristic curves generated for each plot, the maximum and minimum normalized vegetation index (MMNDVI), maximum and minimum normalized building index (MMNDBI), and maximum and minimum normalized soil index (MMNDSI) for each plot in each time phase are obtained using maximum and minimum normalization.

[0091] MMNDVI (Max-Min Normalization Differential Vegetation Index) is a time-series modified version of the NDVI index. It's a remote sensing index that uses the volatility of the index to measure changes in land cover. It objectively reflects changes in vegetation cover, weakening the influence of different geographical locations on NDVI, and relying solely on the index's variation patterns to determine changes in land cover over a specific time period. Generally, a higher MMNDVI value indicates richer land cover vegetation. The corresponding MMNDVI characteristic index expression is as follows:

[0092]

[0093] In the formula, MMNDVI represents the maximum and minimum normalized vegetation index, NDVI represents the normalized vegetation index at a certain time; max1 represents the maximum value in the NDVI time series data, and min1 represents the minimum value in the NDVI time series data; through maximum and minimum normalization processing, the values ​​of all calculated MMNDVI remote sensing feature indices are between 0 and 1.

[0094] MMNDBI (Max-Min Normalization Differential Built-Up Index) is a time-series modified version of the NDBI index. It's a remote sensing index used to measure the density or abundance of buildings on the Earth's surface, reflecting changes in building density or abundance. It weakens the influence of geographical location on the NDBI, relying solely on the index's variation patterns to determine changes in surface buildings over a specific time period. Generally, a higher MMNDBI value indicates a richer building population. The corresponding MMNDBI characteristic index expression is as follows:

[0095]

[0096] In the formula, MMNDBI represents the maximum and minimum normalized building index, NDBI represents the normalized building index at a certain time; max2 represents the maximum value in the NDBI time series data, and min2 represents the minimum value in the NDBI time series data; through maximum and minimum normalization processing, the values ​​of all calculated MMNDBI remote sensing feature indices are between 0 and 1.

[0097] MMNDSI (Max-Min Normalization Differential Soil Index) is a remote sensing index used to measure the abundance of bare soil on the land surface. It is a time-series modification of the NDSI index, reducing the influence of different geographical locations on the NDSI and relying solely on the index's variation patterns to determine changes in bare soil over a specific time period. Generally, a higher MMNDSI value indicates a higher content of bare soil on the land surface. The corresponding MMNDSI characteristic index expression is as follows:

[0098]

[0099] In the formula, MMNDSI represents the maximum and minimum normalized soil index, NDSI represents the normalized soil index at a certain time, max3 represents the maximum value in the NDSI time series data, and min3 represents the minimum value in the NDSI time series data. Through maximum and minimum normalization processing, the values ​​of all calculated MMNDSI remote sensing feature indices are between 0 and 1.

[0100] Based on the established temporary land use vector, the usage period of temporary land use, and the calculated NDVI, NDBI, and NDSI remote sensing characteristic indices for each patch, the MMNDVI, MMNDBI, and MMNDSI time-series characteristic curves for each temporary land use category are plotted using mean processing and maximum-minimum normalization processing, thus obtaining typical time-series sample characteristics of temporary land use. At the same time, 30 sample points of non-temporary land use are randomly selected, and the annual MMNDVI, MMNDBI, and MMNDSI time-series characteristic curves of non-temporary land use are plotted using the above method.

[0101] In step S2, following the above method, time-series feature curves of three remote sensing feature indices for the undetermined land parcel are generated, thus obtaining typical time-series sample features of non-temporary land use. Then, the time-series feature curves of the three remote sensing feature indices are combined, and the Dynamic Time Warping (DTW) method is used to calculate similarity to determine whether the undetermined land parcel is temporary land use. DTW distance-based similarity determination can be used for data of unequal length and emphasizes feature matching; compared to Euclidean distance, it aligns data more naturally.

[0102] DTW uses a time warping function W(n) that satisfies certain conditions to describe the temporal correspondence between a test template and a reference template, and solves for the warping function that minimizes the cumulative distance when the test template and the reference template match. For two time series X... n ={c1,c2…c n} and Y m ={q1,q2…q m Given two time series with lengths X| and |Y|, calculate the similarity between them and determine the regularized path W = W1, W2, ..., W... K Where Max(|X|,|X|)≤K≤|X|+|Y|, W K The form is (i,j), where i represents the i-th coordinate in X and j represents the j-th coordinate in Y; the normalization path W starts from W1 = (1,1) and ends at W1 = (|X|,|Y|) to ensure that every coordinate in X and Y appears in W; where c n For the first time series X n The value of the nth phase in the equation, q m For the second time series Y m The value of the m-th phase in W K Let X be the Kth point in the path, Max(|X|,|Y|) be the larger of the two time series lengths, and |Y|+|Y| be the sum of the two time series lengths.

[0103] In a regular path W, the i and j of W(i,j) are monotonically increasing, represented as:

[0104] W K =(i,j)

[0105] W K+1 =(i′,j′)

[0106] i≤i′≤i+1,j≤j′≤j+1

[0107] In the formula, i and j represent the position indices of the Kth point in the first and second time series, respectively; W K+1 Let i' be the (K+1)th point in the path, and let i' and j' be the position indices of the (K+1)th point in the first and second time series, respectively. i≤i'≤i+1 and j≤j'≤j+1 indicate that the movement at each step in the path is monotonically increasing.

[0108] The shortest optimal reduction path obtained is the DTW distance, which is:

[0109] D(i,j)=Dist(i,j)+min[D(i-1,j),D(i,j-1),D(i-1,j-1)]

[0110] In the formula, D(i,j) is the cumulative DTW distance between the time series at indices i and j, Dist(i,j) is the local Euclidean distance between the i-th point in the first time series and the j-th point in the second time series, min[] represents the minimum cumulative distance among the three possible paths, i.e., the minimum cumulative distance from the top, left or diagonal direction, D(i-1,j) is the cumulative DTW distance between the time series at indices i-1 and j, D(i,j-1) is the cumulative DTW distance between the time series at indices i and j-1, and D(i-1,j-1) is the cumulative DTW distance between the time series at indices i-1 and j-1.

[0111] Calculate the DTW distance between the time series characteristic curves of the three remote sensing characteristic indices of the pending land parcel and the time series characteristic curves of the three remote sensing characteristic indices of temporary land and non-temporary land. If the DTW distance with the temporary land is smaller, the pending land parcel is determined to be temporary land; otherwise, it is determined to be non-temporary land.

[0112] In step S3, for patches of temporary land determined to have reached the end of their reclamation period, the slopes of three characteristic curves are calculated using three remote sensing characteristic indices at the end of the usage period and the end of the reclamation period, respectively. The calculation expressions are as follows:

[0113]

[0114] In the formula, K MMNDVI ,K MMNDBI ,K MMNDSI The curvatures of the time-series characteristic curves for the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively; MMNDVI n ,MMNDBI n ,MMNDSI n The values ​​at the end of the reclamation period on the time-series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI are respectively. m ,MMNDBI m ,MMNDSI m , where n and m are the values ​​at the end of the usage period on the time series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively, and n and m are the end of the reclamation period and the end of the usage period, respectively;

[0115] The slope values ​​of the time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, are used to determine whether land reclamation has been completed.

[0116] When K is satisfied MMNDVI >δ1,K MMNDBI <δ2,KMMNDSI <δ3 indicates that the land reclamation has been completed and the reclamation effect is good;

[0117] When K is satisfied MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI ≥δ3 indicates that land reclamation has not yet been completed for this plot;

[0118] When K cannot be satisfied simultaneously MMNDVI >δ1,K MMNDBI <δ2,K MMNDSI <δ3 or K MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI When the value is ≥δ3, the interpretation is switched to manual.

[0119] Among them, δ1, δ2, and δ3 are the thresholds of the slopes of three remote sensing feature indices used to determine whether land reclamation has been completed, which are statistically derived from 30 temporary land samples that have completed land reclamation. In practical applications, these indices are obtained by statistical calculation based on temporary land samples from different regions.

[0120] Example 2, the method for identifying temporary land use and verifying land reclamation using time-series satellite remote sensing images provided in this embodiment of the invention specifically includes the following steps 1 to 4:

[0121] Step 1: This invention first calculates the MMNDVI, MMNDBI, and MMNDSI remote sensing feature indices for temporary land use categories. These indices are used to extract the fluctuation characteristics of the time-series feature curves and reduce the interference of regional location differences, thereby obtaining typical time-series characteristics of temporary land use. Specifically, this includes the following steps:

[0122] (1.1) Based on the Sentinel-2 optical remote sensing image patches of temporary land use in Zhejiang Province, the data of the patches are integrated according to the approval time and expiration time.

[0123] (1.2) Manually inspect temporary map patches to ensure that the patch information is accurate.

[0124] (1.3) Select 10 temporary land use plots that have been confirmed to have completed land reclamation through on-site investigation for feature statistics.

[0125] (1.4) Load temporary map patches one by one on the Google Earth Engine (GEE) platform. Based on the approval and expiration time of the temporary land use, use the ee.ImageCollection function to filter images with cloud cover of less than 20%.

[0126] (1.5) Use the QA60 band of the GEE platform to perform cloud removal processing on Sentinel-2 satellite remote sensing images for each time phase.

[0127] (1.6) Calculate the remote sensing characteristic indices of MMNDVI, MMNDBI and MMNDSI for temporary land use patches, and plot the time series characteristic curves of NDVI, NDBI and NDSI for temporary land use from the start of approval to the reclamation acceptance stage.

[0128] For Sentinel-2 optical remote sensing satellite data, the MMNDVI index is calculated using the 8th and 4th bands of the Sentinel-2 optical remote sensing satellite data and the NDVI time-series characteristic curves. The calculation formula is as follows:

[0129]

[0130] In the formula, B8 and B4 are the near-infrared band and red band of Sentinel-2 optical remote sensing satellite data, respectively, max1 represents the maximum value in the NDVI time series data, and min1 represents the minimum value in the NDVI time series data.

[0131] MMNDBI is calculated using bands 11 and 8 of Sentinel-2 optical remote sensing satellite data and the NDBI time-series characteristic curve. The calculation formula is as follows:

[0132]

[0133] In the formula, B11 and B18 are the short infrared and near infrared bands of Sentinel-2 optical remote sensing satellite data, respectively; max2 represents the maximum value in NDBI time series data; and min2 represents the minimum value in NDBI time series data.

[0134] MMNDSI is calculated using bands 12 and 8 of Sentinel-2 optical remote sensing satellite data and NDSI time-series characteristic curves. The calculation formula is as follows:

[0135]

[0136] In the formula, B12 and B8 are the mid-infrared and near-infrared bands of Sentinel-2 optical remote sensing satellite data, respectively; max3 represents the maximum value in the NDSI time series data; and min3 represents the minimum value in the NDSI time series data.

[0137] (1.7) Perform linear interpolation and linear compression on the time-series characteristic curves of the 10 identified temporary land parcels, and make the length of their horizontal axis 50.

[0138] (1.8) Take the mean characteristic value of the 10 temporary land use patches at each horizontal axis position, and obtain the time series characteristic curves of the three remote sensing characteristic indices MMNDVI, MMNDBI, and MMNDSI for the temporary land use category, such as... Figure 2 As shown.

[0139] Next, this invention uses the GEE cloud platform to calculate the time-series characteristic curves of three remote sensing characteristic indices, namely MMNDVI, MMNDBI, and MMNDSI, for non-temporary land use (vegetation) throughout the year.

[0140] (1.9) In the GEE cloud platform, 30 non-temporary land (vegetation) samples were randomly selected in the area near the identified temporary land.

[0141] (1.10) Taking the Sentinel-2 remote sensing data from 2019 as an example, the time-series characteristic curves of MMNDVI, MMNDBI, and MMNDSI for 30 non-temporary land use samples were calculated respectively.

[0142] (1.11) The time series characteristic curves of 30 non-temporary land use samples are linearly compressed to make the horizontal axis length 12 (i.e., one value per month).

[0143] (1.12) Take the characteristic mean of 30 non-temporary land use samples at each horizontal axis position to obtain the annual time series characteristic curves of MMNDVI, MMNDBI, and MMNDSI for non-temporary land use categories, as shown below. Figure 3 As shown.

[0144] Step 2, as follows Figure 4 As shown, for each patch to be judged, the time-series characteristic curves of the three remote sensing feature indices of the patch are generated using the above method. Then, the DTW distance algorithm is used to calculate the similarity between the patch to be judged and temporary land use and non-temporary land use, so as to determine whether the patch is temporary land use based on the similarity. The specific steps are as follows:

[0145] (2.1) First, determine the time range to be cut off for the MMNDVI, MMNDBI and MMNDSI time series characteristic curves of the plot to be judged based on the approval start time and reclamation acceptance time.

[0146] (2.2) Input the land parcel to be judged on the GEE cloud platform and obtain the time series characteristic curves of MMNDVI, MMNDBI and MMNDSI of the land parcel to be judged.

[0147] (2.3) Calculate the similarity, i.e., the DTW distance, between the MMNDVI time series characteristic curve of the plot to be judged, the MMNDVI time series characteristic curve of temporary land use, and the MMNDVI time series characteristic curve of non-temporary land use, respectively, to obtain D. 临时用地MMNDVI and D 非临时用地MMNDVICalculate the similarity, i.e., the DTW distance, between the MMNDBI time series characteristic curve of the segment to be judged, the MMNDBI time series characteristic curve of temporary land use, and the MMNDBI time series characteristic curve of non-temporary land use, to obtain D. 临时用地MMNDBI and D 非临时用地MMNDBI Calculate the similarity, i.e., the DTW distance, between the MMNDSI time series characteristic curve of the segment to be judged, the MMNDSI time series characteristic curve of temporary land use, and the MMNDSI time series characteristic curve of non-temporary land use, to obtain D. 临时用地MMNDSI and D 非临时用地MMNDSI .

[0148] (2.4) Based on the calculated 6 DTW distance results, the plots to be judged are automatically identified, and the identification rules are as follows: Figure 4 As shown: For condition D 临时用地MMNDVI <D 非临时用地MMNDVI D 临时用地MMNDBI <D 非临时用地MMNDBI D 临时用地MMNDSI <D 非临时用地MMNDSI If two of these conditions are met, it is determined to be temporary land use; when D 临时用地MMNDVI >D 非临时用地MMNDVI D 临时用地MMNDBI >D 非临时用地MMNDBI D 临时用地MMNDSI >D 非临时用地MMNDSI时 If the land is deemed to be non-temporary, it will be classified as such; otherwise, it will be subject to manual interpretation.

[0149] (2.5) Using temporary land use patches in Zhejiang Province as the data source, 50 of them were selected as verification data. The accuracy and error rate of the discrimination method were calculated. The accuracy of the automatic discrimination of temporary land use in this invention reached 84%, and the error rate was only 4%.

[0150] Step 3, as follows Figure 5 As shown, the determination of whether land reclamation has been completed is based on the characteristic morphology of the time-series characteristic curves of MMNDVI, MMNDBI, and MMNDSI of temporary land use patches. The specific steps are as follows:

[0151] (3.1) For temporary land use that has reached the end of the reclamation period, extract the MMNDVI, MMNDBI and MMNDSI feature values ​​of the temporary land use patch at the end of the use period, as well as the MMNDVI, MMNDBI and MMNDSI feature values ​​at the end of the reclamation period.

[0152] (3.2) Calculate the slopes of MMNDVI, MMNDBI, and MMNDSI, respectively, K MMNDVI ,K MMNDBI ,K MMNDSI The calculation expression is as follows:

[0153]

[0154] In the formula, K MMNDVI ,K MMNDBI ,K MMNDSI The curvatures of the time-series characteristic curves for the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively; MMNDVI n ,MMNDBI n ,MMNDSI n The values ​​at the end of the reclamation period on the time-series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI are respectively. m ,MMNDBI m ,MMNDSI m , where n and m are the values ​​at the end of the usage period on the time series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively, and n and m are the end of the reclamation period and the end of the usage period, respectively;

[0155] (3.3) Determine whether land reclamation is complete based on three slope values, as follows: Figure 5 As shown: When K MMNDVI >δ1,K MMNDBI <δ2,K MMNDSI <δ3 indicates that the land reclamation of this plot is well completed; when K MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI ≥δ3 indicates that the land reclamation has not been completed; if only one of the conditions is met, manual interpretation is required.

[0156] (3.4) Using temporary land use patches in Zhejiang Province as the data source, 50 temporary land use patches that have completed land reclamation were selected as verification data to calculate the accuracy of automatic land reclamation judgment. The accuracy of automatic judgment of whether land reclamation has been completed by this invention reached 86%.

[0157] Step 4, result evaluation.

[0158] Currently, existing methods for verifying land reclamation for temporary land use rely on on-site inspections or manual visual interpretation of high-resolution satellite imagery. There is no automated technology for identifying temporary land use and determining its reclamation. This invention proposes a completely new technical process for identifying temporary land use and reclaiming land.

[0159] (4.1) Qualitative Evaluation: Currently, all methods rely on manual visual interpretation, and there is no automated technology for identifying temporary land use and reclamation. This invention, based on time-series Sentinel-2 optical satellite remote sensing imagery, constructs time-series characteristic curves by calculating three remote sensing feature indices: MMNDVI, MMNDBI, and MMNDSI. Combined with the DTW algorithm, it achieves automatic verification of temporary land use and land reclamation, overcoming the drawbacks of relying solely on manual verification, which consumes a large amount of manpower and time. This provides a new technical means for the automatic monitoring and verification of land reclamation for temporary land use.

[0160] (4.2) Quantitative Evaluation: The accuracy rate of the automatic identification of temporary land use in this invention is 84%, which meets the requirements for automatic identification of temporary land use. Although the accuracy of temporary land use identification still needs to be improved, the error rate is only 4%. For temporary land use that cannot be automatically identified by this invention, manual verification can continue, and the probability of direct errors is low. The accuracy rate of the automatic verification of land reclamation of temporary land use after the reclamation period has expired is 86%, which is highly accurate and has no verification errors. For temporary land use that cannot be automatically identified as having completed land reclamation, manual verification can continue.

[0161] Example 3: The temporary land use interpretation and reclamation verification system provided in this embodiment of the invention includes:

[0162] The time-series remote sensing index sample feature generation module is used to calculate and generate sample features of typical time-series feature curves of remote sensing feature indices for temporary and non-temporary land use using the GEE cloud computing platform and Sentinel-2 satellite remote sensing images.

[0163] The temporary land use judgment module is used to calculate the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged. After linear interpolation or linear compression, it calculates the DTW similarity distance between the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged and the typical time series characteristic curve of temporary land use and non-temporary land use. Based on the DTW distance criterion, it judges whether the land parcel is temporary land use.

[0164] The land reclamation judgment module is used to determine whether land reclamation has been completed for temporary land that has reached the end of its reclamation period. This is done by calculating the slope of the time-series characteristic curve based on the remote sensing characteristic index of the time of use expiration and the time of reclamation expiration.

[0165] To further demonstrate the positive effects of the above embodiments, the present invention conducts the following experiments based on the above technical solutions.

[0166] Taking a temporary land use patch in Zhejiang Province as an example, the start and end time of this patch is from December 15, 2019 to December 29, 2021. The time series NDVI, NDBI, and NDSI of this patch are obtained, and the time series feature length is 52. The time series features are subjected to max-min normalization processing to obtain MMNDVI, MMNDBI, and MMNDSI respectively, and then linear compression is performed. Then, the DTW distance is calculated between the MMNDVI, MMNDBI, and MMNDSI time series feature curves of temporary land use and non-temporary land use respectively. Calculations show that the DTW distances between MMNDVI, MMNDBI, and MMNDSI of this plot and the MMNDVI, MMNDBI, and MMNDSI of the temporary land use are 10.261, 9.828, and 13.757, respectively, and the DTW distances between MMNDVI, MMNDBI, and MMNDSI of this plot and the non-temporary land use are 17.742, 25.102, and 18.183, respectively, simultaneously satisfying D 临时用地MMNDVI <D 非临时用地MMNDVI D 临时用地MMNDBI <D 非临时用地MMNDBI D 临时用地MMNDSI <D 非临时用地MMNDSI The land parcel was correctly identified as temporary land use. The slope of the characteristic curve was calculated using the first and last MMNDVI, MMNDBI, and MMNDSI values ​​of the parcel. Since the months were consistent, δ1 = δ2 = δ3 = 0. K was then calculated. MMNDVI =-0.008≤δ1、K MMNDBI =0.001≥δ2、K MMNDSI =0.003≥δ3, indicating that the land reclamation of this plot has not been completed. Based on the actual situation, this interpretation is correct.

[0167] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention and within the spirit and principles of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for interpreting temporary land use and verifying land reclamation, characterized in that, The method includes: S1. Using the GEE cloud computing platform and Sentinel-2 satellite remote sensing imagery, calculate and generate sample features of typical time-series characteristic curves of remote sensing characteristic indices for temporary and non-temporary land use. S2. Calculate the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged. After linear interpolation or linear compression, calculate the DTW similarity distance between the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged and the typical time series characteristic curves of temporary land and non-temporary land. Based on the DTW distance criterion, determine whether the land parcel is temporary land. S3. For temporary land that has reached the end of its reclamation period, calculate the slope of the time-series characteristic curve based on the remote sensing characteristic index of the time of use expiration and the time of reclamation expiration to determine whether the temporary land has completed land reclamation. In step S2, determining whether a land parcel is for temporary use based on the DTW distance criterion includes: Calculate three remote sensing characteristic indices, including NDVI, NDBI, and NDSI, for each time phase of the land use patch to be judged, and generate time series characteristic curves of the three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, for the land use patch to be judged. By comparing the DTW distance similarity between the time series characteristic curves of the three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, of the plot to be determined and the time series characteristic curves of typical temporary land use and non-temporary land use, it is determined whether the plot to be determined is temporary land use.

2. The method for interpreting temporary land use and verifying reclamation according to claim 1, characterized in that, Before step S1, the Sentinel-2 satellite remote sensing images undergo preprocessing including image screening, cloud removal, and image cropping; temporary and non-temporary land use samples are acquired, and time series information of the entire life cycle of temporary land use is generated; the Sentinel-2 satellite remote sensing image screening is based on the GEE platform, which sets the date according to the vector range of temporary land use, the approval time, and the expiration time, and filters image data with cloud cover of less than 20% during the temporary land use period and the land reclamation period, and uses the QA60 band to perform cloud removal processing on the acquired Sentinel-2 satellite remote sensing images respectively.

3. The method for interpreting temporary land use and verifying reclamation according to claim 1, characterized in that, In step S1, sample features of typical time-series characteristic curves of remote sensing feature indices for temporary and non-temporary land use are calculated and generated, including: Based on the vector range shp and kml files of temporary land use, the temporal NDVI, NDBI and NDSI remote sensing characteristic indices of temporary land use during the land use cycle, as well as the annual temporal NDVI, NDBI and NDSI remote sensing characteristic indices of non-temporary land use, were obtained using the GEE cloud platform and Sentinel-2 satellite remote sensing imagery. By performing maximum-minimum normalization on the time-series characteristic curves, the time-series MMNDVI, MMNDBI, and MMNDSI values ​​of typical temporary and non-temporary land use patches are obtained. Sample characteristics of typical time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, for both temporary and non-temporary land use categories, were obtained by calculating the mean.

4. The method for interpreting temporary land use and verifying reclamation according to claim 3, characterized in that, Obtain the time-series NDVI, NDBI, and NDSI remote sensing characteristic indices of temporary land use within the land use cycle, including: The Sentinel-2 satellite sensor includes 12 spectral bands. By using band operations, the Normalized Difference Vegetation Index (NDVI), Normalized Building Index (NDBI), and Normalized Soil Index (NDSI) are calculated to obtain the NDVI, NDBI, and NDSI values ​​of satellite remote sensing data at different time phases. The calculation expressions for NDVI, NDBI, and NDSI are as follows: In the formula, NIR is the near-infrared band surface reflectance of Sentinel-2 satellite remote sensing image, RED is the red band surface reflectance of Sentinel-2 satellite remote sensing image, SWIR is the short-infrared band surface reflectance of Sentinel-2 satellite remote sensing image, and MIR is the mid-infrared band surface reflectance of Sentinel-2 satellite remote sensing image.

5. The method for interpreting temporary land use and verifying reclamation according to claim 4, characterized in that, The temporary land use area in the Sentinel-2 satellite remote sensing image includes multiple pixels. The NDVI, NDBI, and NDSI values ​​calculated within the vector area are averaged. For the time-series NDVI, NDBI, and NDSI remote sensing characteristic curves generated for each plot, the maximum and minimum normalized vegetation index (MMNDVI), maximum and minimum normalized building index (MMNDBI), and maximum and minimum normalized soil index (MMNDSI) for each plot in each time phase are obtained using maximum and minimum normalization. The expression for the MMNDVI characteristic index is as follows: In the formula, MMNDVI represents the maximum and minimum normalized vegetation index, NDVI represents the normalized vegetation index at a certain time; max1 represents the maximum value in the NDVI time series data, and min1 represents the minimum value in the NDVI time series data; through maximum and minimum normalization processing, the values ​​of all calculated MMNDVI remote sensing feature indices are between 0 and 1. The expression for the MMNDBI feature index is as follows: In the formula, MMNDBI represents the maximum and minimum normalized building index, NDBI represents the normalized building index at a certain time; max2 represents the maximum value in the NDBI time series data, and min2 represents the minimum value in the NDBI time series data; through maximum and minimum normalization processing, the values ​​of all calculated MMNDBI remote sensing feature indices are between 0 and 1. The expression for the MMNDSI characteristic index is as follows: In the formula, MMNDSI represents the maximum and minimum normalized soil index, NDSI represents the normalized soil index at a certain time, max3 represents the maximum value in the NDSI time series data, and min3 represents the minimum value in the NDSI time series data. Through maximum and minimum normalization processing, the values ​​of all calculated MMNDSI remote sensing feature indices are between 0 and 1. Based on the determined temporary land use vector, the temporary land use cycle, and the calculated NDVI, NDBI, and NDSI remote sensing characteristic indices for each patch, and after mean processing and maximum-minimum normalization processing, the time-series characteristic curves of MMNDVI, MMNDBI, and MMNDSI for the temporary land use category are plotted to obtain typical time-series sample characteristics of temporary land use; 30 sample points of non-temporary land use are randomly selected, and the time-series characteristic curves of MMNDVI, MMNDBI, and MMNDSI for non-temporary land use throughout the year are plotted.

6. The method for interpreting temporary land use and verifying reclamation according to claim 1, characterized in that, In step S2, after obtaining typical time-series sample characteristics of non-temporary land use from the time-series characteristic curves of the three remote sensing feature indices, the similarity is calculated by combining the time-series characteristic curves of the three remote sensing feature indices and using the Dynamic Time Warping (DTW) method to determine whether the plot to be determined is temporary land use. For two time series X n ={c1,c2…c n } and Y m ={q1,q2…q m Given two time series with lengths |X| and |Y|, calculate the similarity between them and determine the regularized path W = W1, W2, ..., W... K Where Max(|X|,|Y|)≤K≤|X|+|Y|, W K The form is (i,j), where i represents the i-th coordinate in X and j represents the j-th coordinate in Y; the normalization path W starts from W1 = (1,1) and ends at W1 = (|X|,|Y|) to ensure that every coordinate in X and Y appears in W; where c n For the first time series X n The value of the nth phase in the equation, q m For the second time series Y m The value of the m-th phase in W K Let X be the Kth point in the path, Max(|X|,|Y|) be the larger of the two time series lengths, and |X|+|Y| be the sum of the two time series lengths. In a regular path W, the i and j of W(i,j) are monotonically increasing, represented as: W K =(i,j) W K+1 =(i',j') i≤i'≤i+1,j≤j'≤j+1 In the formula, i and j represent the position indices of the Kth point in the first and second time series, respectively; W K+1 Let i' be the (K+1)th point in the path, and let i' and j' be the position indices of the (K+1)th point in the first and second time series, respectively. i≤i'≤i+1 and j≤j'≤j+1 indicate that the movement at each step in the path is monotonically increasing. The shortest optimal reduction path obtained is the DTW distance, which is: D(i,j)=Dist(i,j)+min[D(i-1,j),D(i,j-1),D(i-1,j-1)] In the formula, D(i,j) is the cumulative DTW distance between the time series at indices i and j, Dist(i,j) is the local Euclidean distance between the i-th point in the first time series and the j-th point in the second time series, min[] represents the minimum cumulative distance among the three possible paths, i.e., the minimum cumulative distance from the top, left or diagonal direction, D(i-1,j) is the cumulative DTW distance between the time series at indices i-1 and j, D(i,j-1) is the cumulative DTW distance between the time series at indices i and j-1, and D(i-1,j-1) is the cumulative DTW distance between the time series at indices i-1 and j-1. Calculate the DTW distance between the time series characteristic curves of the three remote sensing characteristic indices of the pending land parcel and the time series characteristic curves of the three remote sensing characteristic indices of temporary land and non-temporary land. If the DTW distance with the temporary land is smaller, the pending land parcel is determined to be temporary land; otherwise, it is determined to be non-temporary land.

7. The method for interpreting temporary land use and verifying reclamation according to claim 1, characterized in that, In step S3, determining whether the temporary land use has been reclaimed includes: Based on the time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, for temporary land use, the slopes of the three remote sensing characteristic indices at the time of expiration of the temporary land use period and the time of expiration of the reclamation period are calculated respectively; where the expiration of the reclamation period refers to the expiration of one year of use; by combining the differences in the slopes of the three remote sensing characteristic indices, relevant information on whether the temporary land use has completed land reclamation upon the expiration of the reclamation period is determined.

8. The method for interpreting temporary land use and verifying reclamation according to claim 7, characterized in that, Information for determining whether temporary land use that has reached the end of its reclamation period has been reclaimed includes: For patches of temporary land use that have reached the end of their reclamation period, the slopes of three characteristic curves are calculated using three remote sensing characteristic indices at the end of the use period and the end of the reclamation period, respectively. The calculation expressions are as follows: In the formula, K MMNDVI ,K MMNDBI ,K MMNDSI The curvatures of the time-series characteristic curves for the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively; MMNDVI n ,MMNDBI n ,MMNDBI n The values ​​at the end of the reclamation period on the time-series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI are respectively. m ,MMNDBI m ,MMNDSI m , where n and m are the values ​​at the end of the usage period on the time series characteristic curves of the remote sensing feature indices MMNDVI, MMNDBI, and MMNDSI, respectively, and n and m are the end of the reclamation period and the end of the usage period, respectively; The slope values ​​of the time-series characteristic curves of three remote sensing characteristic indices, including MMNDVI, MMNDBI, and MMNDSI, are used to determine whether land reclamation has been completed. When K is satisfied MMNDVI >δ1,K MMNDBI <δ2,K MMNDSI <δ3 indicates that the land reclamation has been completed and the reclamation effect is good; When K is satisfied MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI ≥δ3 indicates that land reclamation has not yet been completed for this plot; When K cannot be satisfied simultaneously MMNDVI >δ1,K MMNDBI <δ2,K MMNDSI <δ3 or K MMNDVI ≤δ1,K MMNDBI ≥δ2,K MMNDSI When the value is ≥δ3, the interpretation is switched to manual. Among them, δ1, δ2, and δ3 are the thresholds of the slopes of three remote sensing feature indices used to determine whether land reclamation has been completed, obtained by statistical calculation of temporary land use samples from different regions.

9. A system for interpreting temporary land use and verifying land reclamation, characterized in that, This system is used to regulate the temporary land use interpretation and reclamation verification method according to any one of claims 1-8. The system includes: The time-series remote sensing index sample feature generation module is used to calculate and generate sample features of typical time-series feature curves of remote sensing feature indices for temporary and non-temporary land use using the GEE cloud computing platform and Sentinel-2 satellite remote sensing imagery. The temporary land use judgment module is used to calculate the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged. After linear interpolation or linear compression, it calculates the DTW similarity distance between the time series characteristic curve of the remote sensing characteristic index of the land parcel to be judged and the typical time series characteristic curve of temporary land use and non-temporary land use. Based on the DTW distance criterion, it judges whether the land parcel is temporary land use. The land reclamation judgment module is used to determine whether land reclamation has been completed for temporary land that has reached the end of its reclamation period. This is done by calculating the slope of the time-series characteristic curve based on the remote sensing characteristic index of the time of use expiration and the time of reclamation expiration.

Images