Mangrove forest space expansion-atrophy monitoring method based on continuous time sequence spectrum index

Abstract

The invention relates to a mangrove forest space expansion-atrophy monitoring method based on a continuous time sequence spectral index. The mangrove forest space expansion-atrophy monitoring method comprises the following steps: step (1), preprocessing a remote sensing image; (2) generating an original time sequence; (3) removing abnormal observation data in the time sequence; step (4), constructing a vegetation index time sequence sensitive to the mangrove forest within the year; step (5), extracting a mangrove forest potential area every year based on a threshold method; step (6), identifying and correcting potential areas of the mangrove forest in data abnormal years; (7) capturing the years of space expansion and atrophy of the mangrove forest; (8) drawing a space-time distribution diagram of expansion and atrophy of the mangrove forest; and (9) quantitatively evaluating the annual mangrove forest area increase and loss. The method draws the spatial distribution diagram of mangrove forest expansion and atrophy based on long-time series multispectral remote sensing data, can accurately monitor the time point when the mangrove forest is subjected to atrophy and expansion events, and ascertains the spatial and temporal characteristics of the mangrove forest distribution range.

Description

Technical Field

[0001] The present invention belongs to the technical field of spatial monitoring of coastal wetland vegetation, and particularly relates to an identification algorithm for the spatial expansion and shrinkage of mangroves. By combining mangrove-sensitive spectral indices, continuous time series, and time series interpolation methods, precise and efficient monitoring of the spatio-temporal distribution of mangroves in coastal wetlands is achieved. Background Art

[0002] As a type of coastal wetland with the highest biodiversity and unique ecological functions globally, the mangrove ecosystem is not only a key barrier for maintaining the ecological balance of the coastal sea, but also plays an irreplaceable role in ecological value and functional synergy in areas such as regulating the global carbon cycle, enhancing the ecological resilience of the coastal zone, and mitigating climate change. However, under the dual influence of frequent extreme climate events and human activities (such as the expansion of unregulated aquaculture, coastal urbanization, and industrial and agricultural pollution), the global mangrove area has decreased by approximately 35% in the past 40 years, and its degradation rate far exceeds that of other coastal ecosystems. Against this background, building a large-scale and high-precision mangrove dynamic monitoring system has become an urgent task for wetland ecological protection and system restoration; accurately quantifying the trend of mangrove area evolution not only provides a scientific basis for the high-quality achievement of global sustainable development goals, but also is a strategic requirement for maintaining the global biodiversity treasure house and enhancing the blue carbon sequestration service capacity.

[0003] The quantitative assessment of the dynamic changes in mangrove area has traditionally relied on field survey methods. Although in-situ surveys can accurately obtain mangrove habitat parameters and achieve area measurement with centimeter-level spatial accuracy, due to the special intertidal habitat in the mangrove distribution area (such as the risk of tidal fluctuations and difficulties in accessing the muddy substrate), this method has inherent defects such as fragmented operation scope, poor spatio-temporal continuity, and high labor costs. With the innovation of remote sensing technology, the mangrove monitoring system has gradually evolved towards an integrated space-air-ground direction: Unmanned Aerial Vehicle (UAV) remote sensing, with its sub-meter resolution and flexible operation characteristics, has become an emerging solution for fine-scale mangrove monitoring at the medium and small scales. However, its single operation coverage area is usually less than 10 square kilometers, making it difficult to support the spatio-temporal evolution analysis of regional ecosystems. To break through the spatial scale limitation, the academic community has begun to integrate high spatio-temporal resolution satellite images with machine learning algorithms to construct intelligent interpretation models. However, it is worth noting that such methods still face dual technical bottlenecks - the strong dependence of machine learning models on labeled data leads to limited generalization ability in mangrove monitoring; existing research is mostly based on the static classification framework of single-temporal images, lacking an effective representation of the temporal response mechanism of the mangrove degradation-expansion dynamic process and sudden disturbance events (such as typhoon damage, pest outbreaks).

[0004] Time series contain the spatio-temporal attributes of mangroves. How to effectively and accurately mine time series information is a key technology for analyzing the spatio-temporal dynamics of mangroves. Current research mostly conducts area change statistics based on the mangrove distribution mapping results of discrete years (such as at 5-year intervals). Although this discrete analysis method can identify the macroscopic evolution trend, it is difficult to reveal the continuous spatio-temporal heterogeneity law of mangrove expansion and retreat in the special intertidal habitat. It is worth noting that remote sensing cloud platforms represented by Google Earth Engine and PIE-Engine, by integrating PB-level multi-source remote sensing data and built-in time series analysis modules, are triggering a paradigmatic innovation in mangrove dynamic monitoring - algorithms such as LandTrendr, CCDC, and VCT can automatically detect mangrove disturbance events (such as typhoon damage, returning ponds to forests) and quantify the annual area fluctuations by constructing continuous change trajectories of spectral indices such as NDVI and LSWI.

[0005] However, empirical studies have shown that there are still two bottlenecks in the existing methods: (1) The adaptability of the algorithms in the intertidal wetland scenario is insufficient, and the pixel mixing effect caused by tidal periodic interference leads to a high misjudgment rate in change detection. (2) These algorithms can only capture the characteristic information of the time series, but they cannot identify the mangrove growth areas. Therefore, improvement is urgently needed. Summary of the Invention

[0006] To solve the above problems, the primary object of the present invention is to provide a method for monitoring the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices. This method combines multi-source time series satellite remote sensing images and remote sensing spectral indices sensitive to the mangrove canopy, identifies the perennial mangrove growth areas, and on this basis, performs time series interpolation to weaken the influence of low-quality observation data on the identification accuracy, accurately identify the years when mangrove expansion and shrinkage events occur, and quantitatively evaluate the area increase and loss of mangroves.

[0007] Another object of the present invention is to provide a method for monitoring the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices, which can accurately identify mangrove expansion and shrinkage events in different geographical scenarios and analyze the spatio-temporal succession patterns of mangroves in different regions. Specifically, for the differences in mangrove populations in different countries and the differences in coastal land cover type scenarios, identify the mangrove growth areas, including areas that grew and then disappeared, and areas where mangroves are still growing, and explore the laws of mangrove expansion and shrinkage, and analyze the categories of mangrove expansion and shrinkage.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for monitoring the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices, the method comprising the following steps:

[0010] Step (1): Remote sensing image preprocessing;

[0011] Based on the Google Earth Engine platform, using Landsat 5 / 7 / 8 / 9 and Sentinel-2 remote sensing images, and selecting the study area and time for monitoring, preprocess these images, including radiometric normalization, removing cloud, snow, and shadow pixels.

[0012] Step (2): Generate the original time series;

[0013] Construct the original time series of two remote sensing image datasets respectively;

[0014] Step (3): Remove abnormal observation data in the time series;

[0015] Due to the uncertainty of tidal changes in the coastal zone, the tide may cover the mangrove canopy, resulting in mangrove pixels in the satellite image becoming water pixels or semi-water and semi-vegetation pixels, affecting the quality of the time series. Use the Normalized Difference Water Index to remove the abnormal points of water pixels and semi-water and semi-vegetation.

[0016] Step (4): Construct the time series of vegetation indices sensitive to mangroves within a year;

[0017] Select vegetation indices sensitive to mangroves, such as the Mangrove Vegetation Index (MVI), and calculate the annual mean of the vegetation index for each pixel based on the annual spectral index synthesis method to construct the interannual time series of the vegetation index.

[0018] Step (5): Extract the potential mangrove areas for each year based on the threshold method;

[0019] Use the threshold method to confirm the threshold (Threshold) for extracting the potential mangrove areas with the mangrove sensitivity index. Pixels with vegetation index values greater than this threshold are marked as potential mangrove areas. Taking MVI as an example, after repeatedly adjusting the experimental parameters, the MVI threshold based on Landsat images is 3 - 4; the MVI threshold based on Sentinel-2 images is 2.5 - 4.

[0020] Step (6): Identification and correction of potential mangrove areas in years with abnormal data;

[0021] The annual observation data of remote sensing images may have low quality. Some pixels have been identified as potential mangrove areas for many years, but in one or two years, due to the low quality of the annual observation data of remote sensing images (the threshold of the vegetation index sensitive to mangroves is too low), the potential mangrove areas are not identified. Therefore, time series interpolation is used to correct the areas that are not correctly identified.

[0022] Step (7): Capture the years of mangrove spatial expansion and shrinkage;

[0023] If the number of times a pixel is recognized for consecutive years exceeds a certain threshold Y, then the pixel is determined to be a mangrove growth area. The initial year of consecutive recognition is the time of mangrove expansion, and the year when consecutive recognition is interrupted is the time of mangrove shrinkage (disappearance). If a pixel is continuously detected as a potential mangrove during the period from the year corresponding to (end year of change detection - Y) to the end year of change detection, the year corresponding to the result of (end year of change detection - Y) will be determined as the initial year of mangrove expansion in this area.

[0024] Step (8): Draw the spatio-temporal distribution map of mangrove expansion and shrinkage;

[0025] Resample the mangrove expansion and shrinkage distribution maps extracted from Landsat and Sentinel-2 time series images to a 10m spatial resolution, and merge them to draw the final spatio-temporal distribution map of mangrove expansion and shrinkage.

[0026] Step (9): Quantitatively evaluate the annual increase and loss of mangrove area;

[0027] Based on the Google Earth Engine platform, the area of mangrove expansion and shrinkage each year.

[0028] Furthermore, in the said step (3), the NDWI calculation formula is as follows:

[0029] NDWI = (G - NIR) / (G + NIR)

[0030] Where G is the green band of the remote sensing image, and NIR is the near-infrared band of the remote sensing image.

[0031] Furthermore, in the said step (4), the MVI calculation formula is as follows:

[0032] MVI = (NIR - G) / (SWIR - G)

[0033] Where NIR is the near-infrared band of the remote sensing image, and G is the green band of the remote sensing image.

[0034] The paradigm for constructing the interannual time series of MVI is as follows:

[0035] TS0 = Mean[TS Intra-annual (MVI)]

[0036] Where TS0 is the interannual time series of MVI, and TS Intra-annual (MVI) is the intra-annual time series of MVI, and Mean means synthesizing the intra-annual time series by taking the mean.

[0037] Furthermore, in the said step (5), the calculation formula for extracting the potential mangrove area of each year in the time series is as follows:

[0038] TS1 = TS0 (MVI ≥ Threshold)

[0039] Where TS1 is the time - series collection of the potential mangrove areas and non - mangrove potential areas identified each year. The pixel points of the potential mangrove areas are marked as 1, and the pixel points of the non - mangrove potential areas are marked as 0. Therefore, TS1 is a time series composed of 0s and 1s.

[0040] Furthermore, in step (6), time - series interpolation is used to correct the areas that are not correct. The interpolation formula is as follows:

[0041] TS2 = Interpolation(TS1, k)

[0042] Where TS2 is the corrected time series, and k is the interpolation time length. Generally, only 1 year is interpolated, that is, the discontinuous points in the time series.

[0043] Furthermore, in step (7), the years of mangrove expansion - shrinkage are detected based on the method that the number of times a pixel is continuously identified for more than a certain threshold Y. The calculation formula is as follows:

[0044]

[0045] Where Year Expansion and Year Shrink are the detected years of mangrove expansion and shrinkage respectively, Year Expansion and Year Shrink are the initial year and the interrupted year when continuously identified as potential mangrove areas respectively. Y is the threshold for determining expansion and shrinkage based on the number of consecutive identification years. The threshold is 4 based on Landsat time - series images and 2 based on Sentinel - 2 time - series images.

[0046] Compared with the prior art, the advantages of the present invention are as follows:

[0047] The present invention monitors the spatio - temporal distribution of mangrove expansion - shrinkage on a large - scale spatial scale based on multi - source time - series remote sensing images, quantitatively evaluates the increase and decrease of mangrove area, and makes up for the defects of small - scale in - situ ground measurement, high cost, and the need for a large number of ground - measured sample points for machine - learning classification. At the same time, obtaining the spatio - temporal distribution and area change of mangroves on a large scale can provide a basis for formulating efficient scientific protection policies for mangroves and promoting the sustainable development of mangroves.

[0048] Meanwhile, based on spectral indices sensitive to mangrove canopies, multi-source remote sensing time series datasets, and time series interpolation methods, the present invention realizes the recognition of spatio-temporal succession patterns of mangroves in different geographical locations, solves the problem of false positive results caused by tides in the application of current time series algorithms in the intertidal zone, and can monitor the expansion and shrinkage of mangroves under different coastal habitat backgrounds. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flowchart of a method for monitoring the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices provided by an embodiment of the present invention.

[0050] Figure 2 It is the spatio-temporal distribution of mangrove expansion and shrinkage and the difference in monitoring results from other change monitoring algorithms provided by an embodiment of the present invention.

[0051] Figure 3 It is the spatio-temporal distribution of mangrove expansion and shrinkage monitored at different geographical locations provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention. The principles and features of the present invention will be described below in conjunction with the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0053] The implementation steps of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0054] Figure 1 As shown, it is a flowchart of a method for monitoring the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices implemented by the present invention. As shown in the figure, the method includes the following steps:

[0055] Step (1): Remote sensing image preprocessing;

[0056] Based on the Google Earth Engine platform, using Landsat 5 / 7 / 8 / 9 and Sentinel-2 remote sensing images, and selecting the study area and time to be monitored, preprocess the images;

[0057] Step (2): Synthesize the original time series;

[0058] Construct the original time series of two remote sensing image datasets respectively;

[0059] Step (3): Remove the abnormal observation data from the time series;

[0060] Due to the uncertainty of tidal changes in the coastal zone, the tide may cover the mangrove canopy, causing the mangrove pixels in the satellite image to become water pixels or semi-water and semi-vegetation pixels, which affects the quality of the time series. Use the Normalized Difference Water Index (NDWI) to remove the abnormal points of water pixels and semi-water and semi-vegetation. The calculation formula 1.1 of NDWI is as follows:

[0061] NDWI = (G - NIR) / (G + NIR) (1.1)

[0062] In the formula, G is the green band of the remote sensing image, and NIR is the near-infrared band of the remote sensing image.

[0063] Step (4): Construct the time series of the vegetation index sensitive to mangroves within a year;

[0064] Select the vegetation index sensitive to mangroves, such as the Mangrove Vegetation Index (MVI). Based on the annual spectral index synthesis method, calculate the mean value of the vegetation index of each pixel every year to construct the inter-annual time series of the vegetation index.

[0065] The calculation formula 1.2 of MVI is as follows:

[0066] MVI = (NIR - G) / (SWIR - G) (1.2)

[0067] In the formula, NIR is the near-infrared band of the remote sensing image, and G is the green band of the remote sensing image.

[0068] The paradigm 1.3 for constructing the inter-annual time series of MVI is as follows:

[0069] TS0 = Mean[TS Intra-annual (MVI)] (1.3)

[0070] In the formula, TS0 is the inter-annual time series of MVI, and TS Intra-annual (MVI) is the intra-annual time series of MVI, and Mean represents the mean synthesis of the intra-annual time series.

[0071] Step (5): Extract the potential mangrove areas for each year based on the threshold method;

[0072] Use the threshold method to confirm the threshold (Threshold) for extracting the potential mangrove areas using the mangrove sensitivity index. The pixels with vegetation index values greater than this threshold are marked as potential mangrove areas. Taking MVI as an example, after repeatedly adjusting the experimental parameters, the MVI threshold based on Landsat images is 3 - 4; the MVI threshold based on Sentinel-2 images is 2.5 - 4.

[0073] The calculation formula 1.4 for the potential mangrove areas in each year of the time series is as follows:

[0074] TS1 = TS0 (MVT ≥ Threshold) (1.4)

[0075] In the formula, TS1 is the time series set of the potential mangrove areas and non-potential mangrove areas identified each year. The pixel points of the potential mangrove areas are marked as 1, and the pixel points of the non-potential mangrove areas are marked as 0. Therefore, TS1 is a time series composed of 0 and 1.

[0076] Step (6): Identification and correction of potential mangrove areas in data abnormal years;

[0077] The annual observation data of remote sensing images may have low quality. Some pixel points have been identified as potential mangrove areas for many years, but in one or two years, due to the low quality of the annual observation data of remote sensing images (the threshold of the vegetation index sensitive to mangroves is too low), the potential mangrove areas are not identified. Therefore, time series interpolation is used to correct the incorrect areas. The interpolation formula 1.5 is as follows:

[0078] TS2 = Interpolation(TS1, k) (1.5)

[0079] In the formula, TS2 is the corrected time series, and k is the interpolation time length, generally only interpolating for 1 year, that is, the discontinuous points in the time series.

[0080] Step (7): Capture the years of mangrove spatial expansion and shrinkage;

[0081] If the number of times a pixel is continuously identified for more than a certain threshold Y, then determine that pixel as a mangrove growth area. The initial year of continuous identification is the time of mangrove expansion, and the year when the continuous identification is interrupted is the time of mangrove shrinkage (disappearance). If a pixel is continuously detected as a potential mangrove during the period from the year corresponding to (change detection end year - Y) to the change detection end year, then the year corresponding to the result of (change detection end year - Y) will be determined as the initial year of mangrove expansion in this area. The calculation formula is as follows:

[0082]

[0083] In the formula, YearExpansion and Year Shrink are the years when the detected mangrove expansion and shrinkage occurred, respectively, and Year Expansion and Year Shrink are the initial year and the interruption year when the area was continuously identified as a potential mangrove area, respectively. Y is the number of consecutive identified years to determine the thresholds for expansion and shrinkage. The threshold based on Landsat time-series images is 4, and the threshold based on Sentinel-2 time-series images is 2.

[0084] Step (8): Plot the spatio-temporal distribution maps of mangrove expansion and shrinkage;

[0085] Resample the mangrove expansion and shrinkage distribution maps extracted from Landsat and Sentinel-2 time-series images to a spatial resolution of 10 m, and merge them to plot the final spatio-temporal distribution maps of mangrove expansion and shrinkage.

[0086] Step (9): Quantitatively evaluate the annual increase and loss of mangrove area;

[0087] Based on the Google Earth Engine platform, the areas of mangrove expansion and shrinkage each year.

[0088] It can be seen that the spatio-temporal distribution of mangrove expansion - shrinkage is as Figure 2 shown. Among them, there are a large number of noise points in the spatio-temporal expansion distribution of mangroves monitored by the LandTrendr algorithm and the CCDC algorithm, while the method proposed in the present invention can accurately identify the expansion and shrinkage of mangroves. Figure 3 is the spatio-temporal distribution of mangrove expansion - shrinkage monitored at different geographical locations, indicating that the method proposed in the present invention can be applied to different coastal wetland scenarios and has strong robustness and generalization ability.

[0089] The present invention realizes the extraction of the perennial growth area of mangroves based on multi-source long-time-series remote sensing images and mangrove vegetation indices. At the same time, the time interpolation method is used to eliminate the influence of low-quality observation values on the recognition results, improve the recognition accuracy of mangrove expansion and shrinkage, and reduce the area of false positive detection in the recognition results. Compared with other time-series change detection algorithms, the method proposed in the present invention has stronger robustness; the spatio-temporal distribution maps of mangrove expansion - shrinkage are plotted based on different geographical location scenarios, demonstrating that the method proposed in the present invention has good generalization ability in detecting the spatio-temporal succession of mangroves.

[0090] In summary, the advantages of the present invention are as follows:

[0091] 1. It has stronger robustness. By combining the vegetation spectral index sensitive to mangrove phenology, continuous multi-source data time series, and time series interpolation method, the present invention reduces the false positive detection rate of identifying the spatio-temporal dynamics of mangroves, determines the mangrove growth area, captures the years when mangrove expansion and shrinkage events occur, draws the spatio-temporal distribution map of mangrove expansion-shrinkage, quantitatively evaluates the increase and decrease of mangrove area, and makes up for the defects of small ground in-situ measurement range, high cost, and the need for a large number of ground measured sample points in machine learning classification; at the same time, obtaining the spatio-temporal distribution and area change of mangroves on a large scale can provide a basis for formulating efficient scientific protection policies for mangroves and promoting the sustainable development of mangroves.

[0092] 2. It has good generalization. Based on the spectral index sensitive to the mangrove canopy, multi-source remote sensing time series data set, and time series interpolation method, the present invention realizes the identification of the spatio-temporal succession pattern of mangroves in different geographical locations, solves the problem of false positive results caused by tides when the current time series algorithm is applied in the intertidal zone, can realize the monitoring of mangrove expansion-shrinkage under different coastal habitat backgrounds, can be applied to coastal wetlands in different geographical locations, realizes the change detection of the spatio-temporal succession of mangroves in a large-scale range, and explores the differences in the succession characteristics of mangroves in different scenarios.

[0093] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0094] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A monitoring method for the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices, characterized in that It includes the following steps: Step (1): Preprocessing of remote sensing images; The preprocessing includes radiometric normalization processing, removing cloud, snow and shaded pixels; Step (2): Generating the original time series; Respectively construct the original time series of two remote sensing image datasets; Step (3): Removing abnormal observation data in the time series; Step (4): Constructing the time series of vegetation indices sensitive to mangroves within a year; Select the mangrove vegetation index MVI, calculate the annual mean value of the vegetation index for each pixel based on the annual spectral index synthesis method, and construct the inter-annual time series of the vegetation index; Step (5): Extracting the potential mangrove areas each year based on the threshold method; Use the threshold method to confirm the threshold Threshold for extracting the potential mangrove areas with the mangrove sensitivity index. Pixels with vegetation index values greater than this threshold are marked as potential mangrove areas. The MVI threshold based on Landsat images is 3 - 4; the MVI threshold based on Sentinel-2 images is 2.5 - 4; Step (6): Identification and correction of potential mangrove areas in years with abnormal data; The annual observation data of remote sensing images may have low quality. Use time series interpolation to correct the areas that are not correct; Step (7): Capturing the years of mangrove spatial expansion and shrinkage; If the number of times a pixel is continuously identified for more than a certain threshold Y, then determine that pixel as a mangrove growth area. The initial year of continuous identification is the time of mangrove expansion, and the year when continuous identification is interrupted is the time of mangrove shrinkage; If a pixel is continuously detected as a potential mangrove during the corresponding year to the end year of change detection, then the corresponding year will be determined as the initial year of mangrove expansion in this area; Step (8): Plotting the spatio-temporal distribution maps of mangrove expansion and shrinkage; Step (9): Quantitatively evaluating the annual increase and loss of mangrove area.

2. The monitoring method for the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices according to claim 1, characterized in that In the said step (1), based on the Google Earth Engine platform, use Landsat 5 / 7 / 8 / 9 and Sentinel-2 remote sensing images, and select the monitored study area and time to preprocess these images.

3. The monitoring method for mangrove spatial expansion and shrinkage based on continuous time series spectral indices according to claim 1, characterized in that In the said step (3), use the Normalized Difference Water Index to remove water pixel points and abnormal points of semi-water and semi-vegetation.

4. The monitoring method for mangrove spatial expansion - shrinkage based on continuous time - series spectral indices according to claim 3, characterized in that In the said step (3), The calculation formula of the Normalized Difference Water Index NDWI is as follows: NDWI = (G - NIR) / (G + NIR) Where G is the green band of the remote sensing image, and NIR is the near-infrared band of the remote sensing image.

5. The monitoring method for mangrove spatial expansion and shrinkage based on continuous time series spectral indices according to claim 1, wherein In the said step (4), the calculation formula of MVI is as follows: MVI = (NIR - G) / (SWIR - G) Where NIR is the near-infrared band of the remote sensing image, G is the green band of the remote sensing image, and SWIR is the shortwave infrared band; The paradigm for constructing the inter-annual time series of MVI is as follows: TS0 = Mean[TS Intra-annual (MVI)] where TS0 is the interannual time series of MVI, and TS Intra-annual (MVI) is the intrayear time series of MVI, and Mean represents the mean synthesis of the intrayear time series.

6. The monitoring method for mangrove spatial expansion - shrinkage based on continuous time - series spectral indices according to claim 1, wherein In the said step (5), the calculation formula for extracting the potential mangrove areas of each year in the time series is as follows: TS1 = TS0(MVI≥Threshold) Where TS1 is the time series collection of the potential mangrove areas and non-mangrove potential areas identified each year. The pixel points of the potential mangrove areas are marked as 1, and the pixel points of the non-mangrove potential areas are marked as 0. Therefore, TS1 is a time series composed of 0 and 1.

7. The monitoring method for mangrove spatial expansion - shrinkage based on continuous time - series spectral indices according to claim 1, characterized in that In the step (6), time series interpolation is used to correct the areas that are not correct. The interpolation formula is as follows: TS2 = Interpolation(TS1, k) Where TS2 is the corrected time series, and k is the interpolation time length, generally only interpolating for 1 year, that is, the discontinuous points in the time series.

8. The monitoring method for the spatial expansion and shrinkage of mangroves based on continuous time series spectral indices according to claim 1, characterized in that In the step (7), the corresponding year refers to the end year of change detection - Y. The years of mangrove expansion - shrinkage are detected based on the method that the number of times a pixel is continuously identified for multiple years exceeds a certain threshold Y. The calculation formula is as follows: where Year Expansion and Year Shrink are the years of detected mangrove expansion and shrinkage respectively, Year Expansion and Year Shrink are the initial year and the interrupted year of the mangrove potential area continuously identified respectively, Y is the threshold of the continuously identified years to determine expansion and shrinkage, the threshold based on Landsat time-series images is 4, and the threshold based on Sentinel-2 time-series images is 2.

9. The monitoring method for mangrove spatial expansion - shrinkage based on continuous time - series spectral indices according to claim 1, characterized in that In the step (8), the mangrove expansion and shrinkage distribution maps extracted from Landsat and Sentinel-2 time series images are resampled to a 10m spatial resolution and merged to draw the final spatio-temporal distribution map of mangrove expansion and shrinkage.

10. The monitoring method for mangrove spatial expansion - shrinkage based on continuous time - series spectral indices according to claim 1, characterized in that In the step (9), based on the Google Earth Engine platform, the area of mangrove expansion and shrinkage each year.

Images