Novel spartina alterniflora distribution remote sensing extraction method

Abstract

The invention relates to a novel spartina alterniflora distribution remote sensing extraction method, which comprises the following steps: acquiring satellite remote sensing data, and constructing NDVI (normalized difference vegetation index) time sequence data according to the satellite remote sensing data; constructing a spartina alterniflora remote sensing index according to the vegetation index stack; and obtaining digital elevation model data, and performing post-processing on the spartina alterniflora remote sensing index according to the digital elevation model data to obtain a spartina alterniflora distribution map. The method has the beneficial effects that the complexity of extracting the spartina alterniflora in the complex environment of the coastal wetland is effectively reduced through phenological feature selection, spartina alterniflora remote sensing index construction, histogram threshold segmentation, post-processing and the like.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing images, and more specifically, to a novel remote sensing extraction method for Spartina alterniflora distribution. Background Art

[0002] In recent years, the integrated prevention and control of Spartina alterniflora has become one of the hot issues at home and abroad. How to accurately obtain the distribution location of Spartina alterniflora has become an urgent problem to be solved. Given the dispersion and fragmentation of Spartina alterniflora in the tidal flats, it is difficult to monitor it through field survey methods. Remote sensing technology uses multi-temporal data obtained through platforms such as satellites and drones, making it possible to monitor Spartina alterniflora over a large area and for a long time, providing an effective method for real-time monitoring and analysis of multi-scale changes in Spartina alterniflora. Some scholars extract Spartina alterniflora by combining high spatial resolution and hyperspectral images with vegetation indices. Although high-resolution images can provide more accurate research results in space, the data cost is too high and has poor adaptability in large-scale research. Satellite remote sensing data such as Landsat and Sentinel-2 are widely used in remote sensing scientific research because of their moderate spatial resolution, good temporal continuity, and easy data acquisition.

[0003] However, due to the similarity of spectra between different vegetation, "different objects with the same spectrum" makes it difficult to accurately identify Spartina alterniflora using a single remote sensing image. Studies have shown that by extracting Spartina alterniflora in combination with phenological characteristics, classification errors caused by the similarity of vegetation spectra can be effectively avoided, thereby improving classification accuracy. Based on this finding, many scholars currently use remote sensing data to obtain the phenological characteristics of plants and then identify the distribution area of ​​Spartina alterniflora.

[0004] At present, the research on Spartina alterniflora usually uses different vegetation indices to fit its phenological characteristic curve, and combines supervised and unsupervised classification methods to train selected sample points. Although these methods effectively utilize the phenological characteristics of Spartina alterniflora, the accuracy of their extraction results greatly depends on the selection of sample points, which is highly complex and has great uncertainty. Summary of the invention

[0005] The purpose of the invention is to address the deficiencies of the prior art and to propose a novel remote sensing extraction method for the distribution of Spartina alterniflora.

[0006] In the first aspect, a novel remote sensing extraction method for the distribution of Spartina alterniflora is provided, comprising:

[0007] Step 1: Acquire satellite remote sensing data, and construct NDVI time series data based on the satellite remote sensing data;

[0008] Step 2: constructing a Spartina alterniflora remote sensing index according to the vegetation index stack;

[0009] Step 3: acquiring digital elevation model data, and post-processing the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

[0010] Preferably, step 1 comprises:

[0011] Step 1.1, obtain satellite remote sensing data, and use the median synthesis algorithm to pre-process the satellite remote sensing data to reduce the impact of clouds;

[0012] Step 1.2, calculating the normalized difference vegetation index NDVI for the preprocessed satellite remote sensing data, and obtaining NDVI time series data;

[0013] Step 1.3, filtering the NDVI time series data.

[0014] Preferably, step 2 comprises:

[0015] Step 2.1, apply two Fourier functions to fit the filtered NDVI time series data to obtain the phenological characteristic coefficients, and convert them into a periodic curve representing the vegetation phenological characteristics;

[0016] Step 2.2, normalizing the phenological characteristic coefficients;

[0017] Step 2.3, analyzing the performance characteristics of Spartina alterniflora in terms of phenological characteristic coefficients to construct a remote sensing index of Spartina alterniflora;

[0018] Step 2.4: Use the threshold method to segment the remote sensing index of Spartina alterniflora to obtain the initial distribution range of Spartina alterniflora.

[0019] Preferably, in step 3, the post-processing includes an opening operation and a closing operation; the opening operation is used to eliminate noise caused by different objects with the same spectrum, and the closing operation is used to fill holes caused by the same object with different spectra.

[0020] As a preference, it also includes:

[0021] Step 4: Evaluate the accuracy of the Spartina alterniflora distribution map.

[0022] Preferably, in step 2.3, the remote sensing index formula of Spartina alterniflora is:

[0023]

[0024] Among them, H is the remote sensing index of Spartina alterniflora, and a0 and b1 represent the normalized phenological characteristic coefficients.

[0025] Preferably, in step 4, the accuracy evaluation indicators include overall accuracy, producer accuracy, user accuracy and Kappa coefficient.

[0026] In a second aspect, a novel remote sensing extraction system for the distribution of Spartina alterniflora is provided, which executes any method described in the first aspect, including:

[0027] An acquisition module, used for acquiring satellite remote sensing data and constructing a vegetation index stack according to the satellite remote sensing data;

[0028] A construction module, used for constructing a remote sensing index of Spartina alterniflora according to the vegetation index stack;

[0029] The post-processing module is used to obtain digital elevation model data, and post-process the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

[0030] According to a third aspect, a computer storage medium is provided, wherein a computer program is stored in the computer storage medium; when the computer program is executed on a computer, the computer executes any method described in the first aspect.

[0031] In a fourth aspect, an electronic device is provided, characterized in that it includes:

[0032] Memory, used to store computer programs;

[0033] A processor is used to execute the computer program to implement any method as described in the first aspect.

[0034] The beneficial effects of the present invention are as follows: the present invention aims at the problems of the complexity of coastal wetland environment and the diversity of vegetation types, constructs a remote sensing index of Spartina alterniflora based on phenological characteristics, and makes full use of the knowledge of phenological characteristics to realize the remote sensing accurate extraction of Spartina alterniflora. The method effectively reduces the complexity of extracting Spartina alterniflora in the complex environment of coastal wetlands by selecting phenological characteristics, constructing the remote sensing index of Spartina alterniflora, histogram threshold segmentation, post-processing, etc. The process of the present invention is easy to implement, which is an effective supplement to the existing remote sensing information extraction method of Spartina alterniflora, is conducive to improving the positioning accuracy of the spatial range of Spartina alterniflora, and provides an effective tool for environmental management and ecological research, which is used to monitor and manage the dynamic changes of Spartina alterniflora in coastal wetlands. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A schematic diagram of a study area according to an embodiment of the present invention;

[0036] Figure 2 is a flow chart of an embodiment of the present invention;

[0037] FIG3 is a phenological characteristic coefficient diagram of an embodiment of the present invention;

[0038] Figure 4 It is a phenological characteristic curve diagram of an embodiment of the present invention;

[0039] FIG5 is a schematic diagram showing the distribution of different ground objects in the feature spaces a0 and b1 according to an embodiment of the present invention;

[0040] Figure 6 is a frequency histogram of the remote sensing index of Spartina alterniflora according to an embodiment of the present invention;

[0041] Figure 7 This is a diagram of the remote sensing information extraction results of Spartina alterniflora according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The present invention is further described below in conjunction with embodiments. The description of the following embodiments is only used to help understand the present invention. It should be noted that for ordinary persons in the art, without departing from the principle of the present invention, the present invention can also be modified in some ways, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

[0043] Embodiment 1:

[0044] Hangzhou Bay is located on the eastern coast of China. It is a trumpet-shaped estuary bay formed by the Qiantang River flowing into the East China Sea. It is bordered by the Zhoushan Islands to the east, Shanghai to the north, and the Ningshao Plain to the south. It has a subtropical monsoon climate, abundant rainfall, high vegetation diversity, and rich coastal ecosystems. Hangzhou Bay is located in the sedimentary basin of the Yangtze River Delta. The sediments are mainly silt and sand, forming rich coastal wetland and tidal flat resources, providing an ideal habitat for many animals and plants, and has a high ecological protection value. The study area is located in the middle section of the coastal wetlands on the south bank of Hangzhou Bay. It belongs to the northern part of Ningbo City, Zhejiang Province. It is in the key area where rivers and oceans meet. It is distributed with typical salt marsh vegetation such as Spartina alterniflora, Phragmites australis and Scirpus maritima, which has important ecological significance. The location of the study area is as follows: Figure 1 shown.

[0045] The data used in the present invention include Sentinel-2 satellite remote sensing data, digital elevation model (DigitalElevation Model, DEM) and field measurement data, as shown in Table 1. Sentinel-2 is composed of two satellites, Sentinel-2A and Sentinel-2B, which have wide-band, high-resolution multispectral imaging functions. Multispectral imagers (MultiSpectral Instrument, MSI) carried by two satellites. With 13 imaging bands, Sentinel-2 can provide valuable data support for land use / cover change, climate change monitoring and disaster response. The Sentinel-2 product used in the present invention is surface reflectance data that has been atmospherically corrected and radiation-corrected. The present invention also uses a 30m resolution digital elevation model (DigitalElevationModel, DEM) provided by NASA JPL, which provides undulation information on the surface or terrain surface. Using DEM, it is possible to effectively identify and exclude those misclassified areas that are not suitable for the growth of Spartina alterniflora, thereby improving the accuracy and practicality of ecological research and geographic analysis.

[0046] Table 1 Data used in the study

[0047]

[0048] Aiming at the problems of coastal wetland environmental complexity and vegetation type diversity, the present invention constructs a new Spartina alterniflora remote sensing index based on phenological characteristics and develops a fast and accurate Spartina alterniflora distribution remote sensing extraction method.

[0049] As an embodiment, the present application provides a novel remote sensing extraction method for the distribution of Spartina alterniflora, which uses Sentinel-2 satellite remote sensing data as a data source and is implemented through the GEE cloud platform, including:

[0050] Step 1: Acquire satellite remote sensing data, and construct a vegetation index stack based on the satellite remote sensing data.

[0051] Step 1 includes:

[0052] Step 1.1: Obtain satellite remote sensing data and preprocess them using the median synthesis algorithm to reduce the impact of clouds.

[0053] Specifically, Sentinel-2 products covering the target year of the study area were selected in GEE, and images with less than 80% cloud cover were screened to establish a dense time series image collection. In order to reduce cloud interference and retain key phenological characteristics, median filtering was used on remote sensing images within 10 days, which effectively improved the quality and practicality of remote sensing data and ensured the accuracy and scientificity of the analysis results.

[0054] Step 1.2: Calculate the Normalized Difference Vegetation Index (NDVI) for the preprocessed satellite remote sensing data and obtain NDVI time series data.

[0055] The acquired satellite remote sensing data is a collection of dense time series images, so the present invention calculates NDVI for the preprocessed dense time series image collection in step 1.2. NDVI is a commonly used remote sensing index, which uses the absorption and reflection characteristics of plant chlorophyll to specific spectral bands for calculation and is often used to evaluate the health status and growth vitality of vegetation. The NDVI calculation formula is as follows:

[0056]

[0057] Where NIR and R represent the reflectance of near infrared band and red band respectively.

[0058] Step 1.3, filtering the NDVI time series data.

[0059] After obtaining the NDVI time series data, in order to improve the data quality and accuracy, the Savitzky-Golay (SG) algorithm is used to filter the data.

[0060] As a smoothing technology for digital signal processing, SG filtering is mainly used in spectral data processing in the field of digital images and remote sensing. The filter was proposed by Abraham Savitzky and Marcel JEGolay in 1964. The core idea is to smooth the data by fitting a polynomial on a set of neighboring data points, thereby retaining the high-frequency characteristics of the signal, such as peaks and valleys. The operating principle of the SG filter is to take a certain number of neighboring points (window size) with the point as the center at each data point, and fit a polynomial within this window. The order of the fitted polynomial and the size of the window are the key parameters of the filter. These two parameters determine the effect and smoothness of the filter. The SG filter formula is as follows:

[0061]

[0062] Where i represents the index position of the data point in the data sequence; m represents the number of data points on one side of the filter window; k represents the index offset extending to the left and right of the window boundary from the center data point i, C krepresents the filter coefficient of the kth data point, N represents the total window size, y i+k Represents the value of the original data at position i+k, is the NDVI value after filtering.

[0063] In the present invention, the optimal configuration parameters of the SG filter are explored. The results show that when m=3, N=7, and the polynomial coefficient is equal to 4, the SG filter has the best filtering effect on NDVI. Specifically, for each data point, the present invention sets a seven-point window (total window size N=7) containing the point and the three points on its left and right, and applies a 4th-order polynomial to least squares fitting in this window. In this way, the center point data in each window is updated to the estimated value of the polynomial at that point, thereby generating a series of smooth and coherent filtered NDVI data sequences, which effectively depicts the changing trend of vegetation coverage over time.

[0064] Step 2: constructing a remote sensing index of Spartina alterniflora according to the vegetation index stack.

[0065] Step 2 includes:

[0066] Step 2.1: Use the Fourier function to fit the filtered NDVI time series data to obtain the phenological characteristic coefficient, and convert it into a periodic curve that represents the vegetation phenological characteristics.

[0067] After filtering with the SG algorithm, the noise-free and relatively smooth NDVI data is obtained. The present invention applies the Fourier function to fit the smoothed NDVI data to obtain a periodic curve that characterizes the phenological characteristics of vegetation. Fourier function fitting is a method of decomposing a complex waveform into a number of simple sine and cosine waves. In this way, the NDVI time series can be decomposed into a combination of the basic frequency and its higher harmonics. The Fourier function formula is as follows:

[0068]

[0069] Where, t represents time; f(t) represents the fitted function; a0 is a constant term; a n and b n are Fourier coefficients, representing the amplitude coefficients of cosine and sine respectively; n is an integer multiple of frequency, called the order of harmonics; ω is the frequency coefficient.

[0070] Specifically, the present invention selects a special form of Fourier function, a two-term Fourier function, to fit the NDVI time series data. The formula of the two-term Fourier function is as follows:

[0071] f(t)=a0+a1cos(ωt)+b1sin(ωt)+a2cos(2ωt)+b2sin(2ωt)

[0072] Where t represents the day of year (DOY); f(t) represents the fitted function; ω is the frequency coefficient; a0 is the average value or baseline level of the function f(t), representing the overall change of vegetation phenology; a1 is the amplitude of the cosine component of the frequency ω, representing the general change of vegetation; b1 is the amplitude of the sine component of the frequency ω, also representing the general change of vegetation; a2 is the amplitude of the cosine component of the frequency 2ω, representing the small fluctuation of vegetation; b2 is the amplitude of the sine component of the frequency 2ω, also representing the small fluctuation of vegetation.

[0073] It can be seen that the present invention applies two-term Fourier function fitting to each data point to obtain 5 coefficients representing phenological characteristics, and Figure 3a-3e As shown, they are converted into 5 independent phenological characteristic coefficient diagrams, corresponding to the phenological characteristic coefficients a0, a1, a2, b1, and b2.

[0074] Figure 3 shows that Spartina alterniflora on the south side of Hangzhou Bay shows a relatively consistent pattern in some specific phenological characteristic coefficients. It shows high value consistency in characteristics a0, a2, and b2, while characteristics a1 and b1 generally show low value characteristics.

[0075] Step 2.2, normalize the phenological characteristic coefficients to ensure consistency of calculation.

[0076] The analysis showed that Spartina alterniflora showed relatively consistent patterns in some specific phenological characteristic coefficients. In characteristics a0, a2, and b2, Spartina alterniflora showed high consistency, while characteristics a1 and b1 generally showed low characteristics. The five independent phenological characteristic coefficient graphs were normalized so that the data range of each graph was adjusted to between 0 and 1 to ensure that they had the same dimension. The normalization formula is as follows:

[0077]

[0078] In the formula, x * represents the normalized pixel value, x represents the original pixel value, and x min Indicates the minimum pixel value in the image, x max Indicates the maximum pixel value in the image.

[0079] Step 2.3: Analyze the performance characteristics of Spartina alterniflora in terms of phenological characteristic coefficients to construct the Spartina alterniflora remote sensing index.

[0080] Typical landforms in the study area, such as Spartina alterniflora, reed, farmland, water bodies, breeding ponds and buildings, were selected to draw landform characteristic curves and determine the characteristic coefficients suitable for constructing the Spartina alterniflora index.

[0081] Specifically, we selected typical landform types in the study area, such as Spartina alterniflora, reed, farmland, water body, breeding pond, and building, and drew landform characteristic curves ( Figure 4 ).pass Figure 4 It can be found that the spectrum curve of Spartina alterniflora presents a regular broken line shape. In terms of the a0 feature, Spartina alterniflora is generally higher, while although the reed is slightly lower than Spartina alterniflora, it also maintains a high level. In contrast, buildings, breeding ponds, water bodies, and farmlands show significantly low values ​​in the a0 feature. In terms of the b1 feature, the characteristic values ​​of Spartina alterniflora are generally low, the reed is slightly higher than Spartina alterniflora, while buildings, breeding ponds, water bodies, and farmlands are generally higher. In terms of the three characteristic values ​​of a1, a2, and b2, there is a certain volatility between different sample points.

[0082] According to the stability of the features of Spartina alterniflora at a0 and b1, a0 and b1 were further selected to draw the feature space map, and the distribution characteristics of various types of land objects on the feature space map were analyzed, as shown in Figure 5. Figure 5a This is the remote sensing index map of Spartina alterniflora. Figure 5b Figure 5 shows the spatial distribution of different landform features. It can be seen from Figure 5 that there are obvious distribution differences between Spartina alterniflora and other types of landforms in the feature space of a0 and b1. Spartina alterniflora is mainly concentrated in the high-value area of ​​a0 and the low-value area of ​​b1, reed is mainly concentrated in the middle-value area of ​​a0 and b1, and farmland mainly covers the middle-high value area of ​​a0 and the middle-value area of ​​b1. In contrast, seawater, aquaculture ponds and buildings are mainly located in the low-value area of ​​a0 and the middle-high value area of ​​b1.

[0083] The phenological characteristics of Spartina alterniflora present high values ​​in a0 and low values ​​in b1. This unique combination of high and low values ​​makes it clearly distinguishable from other land features in the feature space. Therefore, by using the differential feature combination of a0 and b1 and combining the construction techniques of multiple vegetation indices in optical remote sensing technology, the Spartina alterniflora index can be constructed, which can effectively classify land features, especially distinguish Spartina alterniflora from other types of land features on the south side of Hangzhou Bay.

[0084] The remote sensing index formula of Spartina alterniflora is as follows:

[0085]

[0086] Where H is the remote sensing index of Spartina alterniflora, and a0 and b1 represent the normalized phenological characteristic coefficients.

[0087] Step 2.4: Use the threshold method to segment the remote sensing index of Spartina alterniflora to obtain the initial distribution range of Spartina alterniflora.

[0088] The constructed Spartina alterniflora remote sensing index was used to draw the spatial distribution map of the Spartina alterniflora index, and the threshold segmentation technique was used to identify Spartina alterniflora. In order to determine the optimal segmentation threshold, the frequency histogram of the Spartina alterniflora index in the study area was analyzed. The frequency histogram can intuitively reflect the frequency information of the data and display the number of data points in each interval in the form of a bar chart, thereby visualizing the distribution characteristics of the data.

[0089] By normalizing and mathematically calculating the phenological characteristics, the H value of the sample points was calculated based on the histogram analysis and combined with remote sensing images, and the threshold range was determined according to the statistical results to extract the initial spatial distribution range of Spartina alterniflora.

[0090] For example, the frequency histogram of the Spartina alterniflora index is as follows: Figure 6 As shown. Combined Figure 6 , based on the histogram analysis, the H value of the sample point was calculated in combination with Sentinel-2 images, and the threshold range was determined based on the statistical results. After multiple experiments and precision evaluation, the threshold was set to 0.35, and the remote sensing index of Spartina alterniflora was segmented to obtain the initial distribution range of Spartina alterniflora.

[0091] Step 3: acquiring digital elevation model data, and post-processing the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

[0092] DEM is a kind of information that characterizes the undulating characteristics of the earth's surface. It has been widely used in the fields of terrain analysis, mapping and hydrological research. Due to the influence of image noise and the complexity of coastal wetlands, the rough extraction results of Spartina alterniflora obtained based on the threshold method have misclassification. The present invention uses the elevation data provided by DEM to post-process the remote sensing information of Spartina alterniflora, and effectively excludes the areas with higher altitudes that are not suitable for the growth of Spartina alterniflora.

[0093] Mathematical morphology is a theory and technology in the field of image processing and analysis, which is mainly used to study and process the shape structure of graphics. It originates from set theory, and extracts the geometric and topological features of images through operations such as expansion and contraction of sets, thereby achieving the purposes of image simplification, segmentation and feature extraction. The basic operations of mathematical morphology mainly include expansion, corrosion, opening operation and closing operation. The present invention uses opening operation and closing operation to post-process the rough extraction information of Spartina alterniflora, wherein the opening operation is used to eliminate the noise caused by the same spectrum of different objects, and the closing operation is used to fill the holes caused by the different spectrum of the same object.

[0094] Embodiment 2:

[0095] On the basis of Example 1, Example 2 of the present application provides a more specific novel remote sensing extraction method for the distribution of Spartina alterniflora, including:

[0096] Step 1: Acquire satellite remote sensing data, and construct a vegetation index stack based on the satellite remote sensing data.

[0097] Step 2: constructing a remote sensing index of Spartina alterniflora according to the vegetation index stack.

[0098] Step 3: acquiring digital elevation model data, and post-processing the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

[0099] Step 4: Evaluate the accuracy of the Spartina alterniflora distribution map.

[0100] Accuracy evaluation is a key step in the remote sensing information extraction process, which is used to verify and quantify the accuracy of the extraction results. The present invention selects sample points through field surveys and high-resolution satellite remote sensing images, and calculates its overall accuracy (OA), producer accuracy (PA), user accuracy (UA) and Kappa coefficient. OA is one of the most intuitive evaluation indicators, which represents the ratio of the number of correctly classified pixels to the total number of pixels. PA represents the proportion of pixels that are correctly classified into a certain category among all pixels whose ground object type is actually a certain category. UA represents the proportion of pixels that are marked as a certain category in the classification result and actually belong to the correct category. The Kappa coefficient is a more complex statistic, which is used to determine whether the classification result is consistent with the actual result. The calculation formula is shown below

[0101]

[0102] In the formula, is the number of correctly classified pixels, and N is the total number of pixels.

[0103]

[0104] In the formula, x ii is the number of pixels of a certain type that are correctly classified, The total number of true data pixels for this category.

[0105]

[0106] In the formula, x ii is the number of pixels of a certain type that are correctly classified, is the number of all pixels of this class in the classification result.

[0107]

[0108] Where N is the total number of observations, Xii is the number of samples correctly classified for each category, and x i+is the sum of the elements in the i-th row of the confusion matrix, x +i is the sum of the elements in the i-th column of the confusion matrix.

[0109] For example, the extraction results of Spartina alterniflora are shown in Figure 7.

[0110] Combining World Imagery and field survey data analysis, it was found that the distribution of Spartina alterniflora extracted ( Figure 7 ) is consistent with the actual distribution. Specifically, Spartina alterniflora is distributed in patches in the Hangzhou Bay Bridge and the coastal swamp area to its east, covering a wide range, while on the west side of the bridge, Spartina alterniflora is distributed in patches, with a smaller coverage and sporadic distribution. In addition, there is a small area of ​​Spartina alterniflora distributed in patches in the outer area of ​​the coastal swamp east of 121°20′0″E.

[0111] Based on the sample points, the producer accuracy, user accuracy, overall accuracy and Kappa coefficient are statistically calculated, as shown in Table 2.

[0112] Table 2 Confusion matrix

[0113]

[0114] Table 2 shows that the extraction results of Spartina alterniflora are relatively accurate, with producer accuracy and user accuracy of 91.50% and 89.71% respectively, an overall accuracy of 90.50%, and a Kappa coefficient of 0.81, indicating that not only the proportion of Spartina alterniflora information in the extraction results to the real Spartina alterniflora is large, and the omission error is small, but also the proportion of accurately identified Spartina alterniflora to the extraction results is large, and the misclassification error is small, so users can obtain more useful information from the extraction results.

[0115] It should be noted that the parts in this embodiment that are the same or similar to those in Embodiment 1 can be referenced to each other and will not be described in detail in this application.

[0116] Embodiment 3:

[0117] On the basis of Example 1, Example 3 of the present application provides a novel remote sensing extraction system for the distribution of Spartina alterniflora, comprising:

[0118] An acquisition module, used for acquiring satellite remote sensing data and constructing a vegetation index stack according to the satellite remote sensing data;

[0119] A construction module, used for constructing a remote sensing index of Spartina alterniflora according to the vegetation index stack;

[0120] The post-processing module is used to obtain digital elevation model data, and post-process the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

[0121] Specifically, the system provided in this embodiment is a system corresponding to the method provided in Embodiment 1. Therefore, the parts in this embodiment that are the same or similar to those in Embodiment 1 can be referenced to each other and will not be repeated in this application.

Claims

1. A novel remote sensing extraction method for the distribution of Spartina alterniflora, characterized in that: include: Step 1: Acquire satellite remote sensing data, and construct NDVI time series data based on the satellite remote sensing data; Step 2: constructing a Spartina alterniflora remote sensing index according to the vegetation index stack; Step 3: acquiring digital elevation model data, and post-processing the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

2. The novel remote sensing extraction method for the distribution of Spartina alterniflora according to claim 1, characterized in that: Step 1 includes: Step 1.1, obtain satellite remote sensing data, and use the median synthesis algorithm to pre-process the satellite remote sensing data to reduce the impact of clouds; Step 1.2, calculating the normalized difference vegetation index NDVI for the preprocessed satellite remote sensing data, and obtaining NDVI time series data; Step 1.3, filtering the NDVI time series data.

3. The novel remote sensing extraction method for the distribution of Spartina alterniflora according to claim 2, characterized in that: Step 2 includes: Step 2.1, apply two Fourier functions to fit the filtered NDVI time series data to obtain the phenological characteristic coefficients, and convert them into a periodic curve representing the vegetation phenological characteristics; Step 2.2, normalizing the phenological characteristic coefficients; Step 2.3, analyzing the performance characteristics of Spartina alterniflora in terms of phenological characteristic coefficients to construct a remote sensing index of Spartina alterniflora; Step 2.4: Use the threshold method to segment the remote sensing index of Spartina alterniflora to obtain the initial distribution range of Spartina alterniflora.

4. The novel remote sensing extraction method for the distribution of Spartina alterniflora according to claim 3 is characterized in that: In step 3, the post-processing includes an opening operation and a closing operation; the opening operation is used to eliminate the noise caused by different objects with the same spectrum, and the closing operation is used to fill the holes caused by the same object with different spectrum.

5. The novel remote sensing extraction method for the distribution of Spartina alterniflora according to claim 4 is characterized in that: Also includes: Step 4: Evaluate the accuracy of the Spartina alterniflora distribution map.

6. The novel remote sensing extraction method for the distribution of Spartina alterniflora according to claim 5 is characterized in that: In step 2.3, the remote sensing index formula of Spartina alterniflora is: Among them, H is the remote sensing index of Spartina alterniflora, and a0 and b1 represent the normalized phenological characteristic coefficients.

7. The novel remote sensing extraction method for the distribution of Spartina alterniflora according to claim 6 is characterized in that: In step 4, the accuracy evaluation indicators include overall accuracy, producer accuracy, user accuracy and Kappa coefficient.

8. A new remote sensing extraction system for the distribution of Spartina alterniflora, characterized in that it is used to: Used to perform the method according to any one of claims 1 to 7, comprising: An acquisition module, used for acquiring satellite remote sensing data and constructing a vegetation index stack according to the satellite remote sensing data; A construction module, used for constructing a remote sensing index of Spartina alterniflora according to the vegetation index stack; The post-processing module is used to obtain digital elevation model data, and post-process the Spartina alterniflora remote sensing index according to the digital elevation model data to obtain a Spartina alterniflora distribution map.

9. A computer storage medium, characterized in that The computer storage medium stores a computer program; when the computer program is executed on a computer, the computer executes any one of the methods described in claims 1 to 7.

10. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the method according to any one of claims 1 to 7.

Images