Vegetation growth path extraction method and system based on remote sensing image

Abstract

The invention relates to the technical field of image filtering extraction, in particular to a vegetation growth path extraction method and system based on a remote sensing image, and the method comprises the steps: obtaining the multispectral remote sensing data of a vegetation region at each time point, obtaining a spectral information interference influence difference value according to the correlation between each pixel and the spectral data of pixels in a local region, and obtaining a spectral information interference influence difference value; the method comprises the following steps: clustering all pixels at the same time point, calculating a normalized vegetation index of each pixel according to filtering parameters when filtering multispectral data of the pixels in each cluster by a technology, and constructing a screened and adjusted characteristic value in a process of extracting a vegetation growth path by adopting an optimal path search algorithm, and adjusting the initial screening value of each pixel when the initial seed point is selected, further screening the initial seed point, and setting the moving direction of each pixel to extract a vegetation growth path. According to the invention, the accuracy of vegetation growth path extraction is improved.

Description

Technical Field

[0001] The present application relates to the field of image filtering and extraction technology, and in particular to a method and system for extracting vegetation growth paths based on remote sensing images. Background Art

[0002] Advances in remote sensing technology have provided new approaches for extracting vegetation growth paths. Satellite and aerial remote sensing platforms enable the acquisition of large-scale, long-term vegetation index data, such as vegetation reflectance spectra. This data contains rich information on vegetation growth, coverage, and phenological changes.

[0003] Currently, extracting vegetation growth paths using remote sensing data has become a mainstream research direction. Researchers often attempt to capture the dynamic changes in vegetation growth by analyzing the temporal variations of vegetation indices. The Normalized Difference Vegetation Index (NDVI) is widely used. It uses the difference and ratio of reflectance between the near-infrared and red bands to intuitively reflect vegetation growth and coverage. In terms of data processing technology, high-resolution remote sensing image processing technology continues to advance. However, when extracting vegetation growth paths using remote sensing data, environmental noise can interfere with the presence of interfering factors, leading to deviations in vegetation reflectance spectral information and, in turn, interfering with the accurate determination of vegetation growth paths. For example, when using NDVI to calculate vegetation growth status, noise can cause abnormal fluctuations in the vegetation index, resulting in large errors in vegetation index calculation and analysis, which in turn affects the accuracy of vegetation growth extension path analysis, resulting in generally low accuracy in vegetation growth path extraction based on remote sensing data. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a vegetation growth path extraction method and system based on remote sensing images. The technical solutions adopted are as follows:

[0005] The present invention provides a method for extracting vegetation growth paths based on remote sensing images, comprising the following steps:

[0006] Obtain multispectral remote sensing data of vegetation areas at each time point;

[0007] The spectral curve of each pixel is fitted to the spectral data of each pixel at each frequency. According to the correlation and difference between the spectral curves of each pixel and the pixels in the local area, the difference value of the spectral information interference effect of each pixel at each time point is obtained;

[0008] All pixels at the same time point are clustered, and based on the fluctuation of the spectral data corresponding to each pixel at the same frequency in each cluster, combined with the average level of the difference value of the spectral information interference of all pixels in each cluster, the filtering parameters when filtering the spectral data of the pixels in the cluster are obtained. The normalized vegetation index of each pixel is calculated based on the filtered multispectral data to obtain the vegetation index time series of each pixel;

[0009] The optimal path search algorithm is used to extract the vegetation growth path, wherein the screening-adjusted characteristic value is obtained according to the average level of the difference value of the spectral information interference of each pixel at all time points, and the initial screening value of each pixel is obtained by using the distribution of the vegetation index time series of each pixel. The adjusted screening value of each pixel is obtained in combination with the characteristic value to screen the initial seed point, and the moving direction of each pixel at the current time point is obtained according to the difference in the normalized vegetation index of each pixel at each time point, and the vegetation growth path is obtained using the optimal path search algorithm.

[0010] Preferably, the method for obtaining the difference value of the spectral information interference effect of each pixel at each time point is:

[0011] For any time point, the spectral information interference effect difference value a of the xth pixel at any time point is x The calculation formula is: Where b x,y represents the normalized result of the Pearson correlation coefficient between the spectral curve of the x-th pixel and the spectral curve of the y-th pixel in its comparison window; c x,y represents the DTW distance between the spectral curve of the x-th pixel and the spectral curve of the y-th pixel in its comparison window, and m represents the number of pixels. A window of a preset size centered on each pixel serves as the comparison window for each pixel.

[0012] Preferably, the method for obtaining the filtering parameters is:

[0013] For each cluster at any time point, the coefficient of variation of the spectral data of all pixels at each frequency in the cluster is calculated, and the average value of the normalized results of the coefficient of variation obtained at all frequencies in the cluster is used as the first parameter. The average value of the normalized results of the difference values ​​affected by the spectral information interference of all pixels is used as the second parameter. The sum of the first parameter and the second parameter is used as the filtering parameter when filtering the spectral data of the pixels in the cluster.

[0014] Preferably, the vegetation index time series of each pixel is composed of the normalized vegetation index of each pixel at all time points arranged in chronological order.

[0015] Preferably, the method for obtaining the characteristic value of the screening adjustment is:

[0016] The average value of the normalized results of the difference values ​​of the spectral information interference effect of each pixel at all time points was calculated and used as the characteristic value of the screening adjustment of each pixel.

[0017] Preferably, the initial screening value of each pixel is the mean value of all elements in the vegetation index time series of each pixel.

[0018] Preferably, the method for calculating the adjusted screening value of each pixel is:

[0019] Among them D i Represents the adjusted screening value of the i-th pixel; d i and t i They represent the initial screening value and the characteristic value of the screening adjustment of the i-th pixel respectively; t v represents the eigenvalue of the screening adjustment of the v-th pixel; m represents the number of pixels.

[0020] Preferably, the screening of the initial seed points further comprises: sorting all pixels from large to small according to the adjusted screening value, and selecting a preset percentage of pixels as the initial seed points.

[0021] Preferably, the method for obtaining the moving direction of each pixel at the current time point is:

[0022] Among the adjacent pixels of each pixel, the direction of the adjacent pixel where the normalized vegetation index at the current time point increases the most compared with the previous time point is selected as the moving direction of each pixel at the current time point.

[0023] An embodiment of the present application also provides a vegetation growth path extraction system based on remote sensing images, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned vegetation growth path extraction methods based on remote sensing images.

[0024] As can be seen from the above, the vegetation growth path extraction method and system based on remote sensing images provided by this application have at least the following beneficial effects:

[0025] When using remote sensing data to extract vegetation growth paths, the present application fully considers the deviation of vegetation reflectance spectral information caused by environmental noise interference, which affects the accurate judgment of vegetation growth paths; therefore, based on the acquired remote sensing image data, the present application combines the aggregation characteristics of the growth state of the same area during vegetation growth to compare and analyze the deviation of spectral information data in different local areas, thereby obtaining analysis results of the deviation of spectral information data in different areas, and optimizes and adjusts the filtering processing of spectral data based on the analysis results. Compared with traditional filtering processing methods, the present application fully combines the aggregation characteristics of vegetation growth status, accurately analyzes the interference influence on spectral data in different areas, effectively improves the accuracy of spectral data processing, and then accurately calculates the normalized vegetation index, and further combines the interference characteristics under vegetation growth aggregation, optimizes and adjusts the selection of initial seed points in the A-star algorithm based on the calculated normalized vegetation index, avoids the influence of noise interference on vegetation growth path extraction, and improves the accuracy of vegetation growth path extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0027] Figure 1 This is a flowchart of the steps of the vegetation growth path extraction method based on remote sensing images provided in this application. DETAILED DESCRIPTION

[0028] In order to further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features and effects of the vegetation growth path extraction method and system based on remote sensing images proposed in this application. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics of one or more embodiments may be combined in any suitable form.

[0029] Unless otherwise specified and limited, terms such as "comprises", "includes" or any other variants thereof are intended to cover non-exclusive inclusion, so that a circuit structure, article or device comprising a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the article or device comprising the element. In addition, the term "and\or" used herein includes any and all combinations of one or more related listed items. All technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs.

[0030] The specific scheme of the vegetation growth path extraction method and system based on remote sensing images provided by this application is described in detail below with reference to the accompanying drawings.

[0031] See also Figure 1 , which shows a flowchart of a vegetation growth path extraction method based on remote sensing images provided by an embodiment of the present application, including the following steps:

[0032] Step 1: Obtain multispectral remote sensing data of vegetation areas at each time point.

[0033] First, the temporal resolution of MODIS data is high, which is conducive to obtaining the vegetation index time series change information of vegetation growth. Therefore, the remote sensing data obtained in the present embodiment is MODIS data. MODIS data are multispectral remote sensing data collected by the medium-resolution imaging spectrometer on Terra satellite and Aqua satellite, wherein the multispectral remote sensing data is a multispectral remote sensing map. Further in the present embodiment, preferably, radiation calibration is carried out to the data collected, and the digital quantization value (DN value) recorded by the sensor is converted into a radiance value with physical significance, ensuring that the data obtained at different times and by the sensor are comparable. FLAASH (Fast Line of Sight Atmospheric Analysis of Spectral Hypercubes) is adopted to correct the scattering and absorption of the atmosphere to the remote sensing signal, restore the real reflectivity of the vegetation, and by selecting ground control points, utilize polynomial transformation method to carry out geometric correction to the multispectral remote sensing image, eliminate the geometric deformation in the image, make the geographical location in the image consistent with the actual geographical location. It should be noted that specific correction process is existing known technology, and relevant redundant description is not done in the present embodiment.

[0034] Step 2: Fit the spectral curve of each pixel to the spectral data of each pixel at each frequency, and obtain the difference value of the spectral information interference effect of each pixel at each time point based on the correlation and difference between the spectral curves of each pixel and the pixels in the local area.

[0035] Accurately acquiring relevant information is crucial when studying vegetation growth and coverage. While widely used, MODIS data is susceptible to interference from various environmental factors. Pixels in different regions may exhibit anomalies in their reflectance spectrum, affecting the accuracy of vegetation indices calculated based on MODIS data. In actual vegetation growth, vegetation growth exhibits clustering characteristics, meaning that vegetation within clustered areas exhibits similar scattering and absorption information. Therefore, when applying interference denoising filtering to MODIS data, efficient data filtering is achieved by leveraging the spectral data characteristics of vegetation under these clustering characteristics.

[0036] Specifically, the MODIS data acquired in this example is multispectral data collected at different time intervals for the same vegetation region. For each point in time, the spectral data for each pixel across all frequency ranges is mapped to a two-dimensional rectangular coordinate system in chronological order to compare the reflectance and scattering characteristics of vegetation in different local areas. The mapping results are then curve-fitted using the least squares method to obtain a spectral curve for each pixel. This spectral curve intuitively reflects the scattering and reflectance characteristics of each pixel's location.

[0037] Similarly, taking any time point as an example, in order to fully reflect the interference of scattering and reflection at the location of each pixel, a comparative analysis of the local area is performed with each pixel as the center. Preferably, in this embodiment, a 7×7 comparison window is constructed with each pixel as the center. By calculating the Pearson correlation coefficient of the spectral curve between each pixel and each other pixel in its comparison window, and normalizing the calculated Pearson correlation coefficient, the Softmax function is used for normalization in this embodiment. In this embodiment, the normalized result of the Pearson correlation coefficient is used to measure the degree of correlation between the growth states of the pixels. The larger the normalized result of the Pearson correlation coefficient, the more likely it is that the pixel belongs to the same clustering area with similar growth state changes.

[0038] However, it's important to note that even when vegetation with similar growth conditions exhibits correlation in the absence of environmental noise, significant differences in spectral data can still occur due to interference. Therefore, we further calculated the DTW (Dynamic Time Warping) distance between the spectral curve of each pixel and the spectral curve of every other pixel within its comparison window. We also analyzed the correlation of spectral variation characteristics of different pixels under vegetation clustering characteristics based on MODIS data, and performed a weighted analysis of interference differences in local areas, thereby more accurately assessing and addressing interference factors in the data.

[0039] Furthermore, for any point in time, based on the correlation and difference between each pixel and the pixels in the local area regarding the spectral curve, the difference value of the spectral information interference effect of each pixel at the said point in time is calculated. In this embodiment, the calculation relationship is: Where a x represents the difference value of the spectral information interference of the x-th pixel at any time point; b x,y represents the normalized result of the Pearson correlation coefficient between the spectral curve of the x-th pixel and the spectral curve of the y-th pixel in its comparison window; c x,y It represents the DTW distance between the spectral curve of the x-th pixel and the spectral curve of the y-th pixel in its comparison window, where m represents the number of pixels. The larger the spectral information interference difference value, the greater the difference in spectral information between the corresponding pixel and other pixels due to interference.

[0040] Step three: cluster all pixels at the same time point, and according to the fluctuation of the spectral data corresponding to each pixel at the same frequency in each cluster, combine the average level of the difference value of the spectral information interference of all pixels in each cluster to obtain the filtering parameters when filtering the spectral data of the pixels in the cluster. Calculate the normalized vegetation index of each pixel based on the filtered multispectral data to obtain the vegetation index time series of each pixel.

[0041] Furthermore, if the local area where the pixel is located has a similar growth state to other pixels, the obtained spectral data information is correlated; however, under the influence of environmental interference, the interference influence in the same clustered area may cause spectral information data deviations with different degrees of influence. At this time, if the characteristic differences under the influence of noise interference in different areas are not optimized, it may cause large deviations in the spectral information data, thereby affecting the calculation of the normalized vegetation index.

[0042] Therefore, according to the difference in interference effects of different pixels in the vegetation clustering state in the multispectral data, a classification optimization process of filtering and noise reduction is performed; specifically, for each time point, all pixels at the same time point are used as input, and the k-means clustering algorithm is used to obtain the clustering results of all pixels to obtain each cluster cluster, wherein the distance metric between different pixels in this embodiment is the Euclidean distance between different pixels, and the number of clusters in this embodiment is set to 10, and the filter parameters are adjusted based on the deviation of the spectral information data difference of the pixels in each cluster in the clustering result. The specific adjustment process is as follows:

[0043] First, the spectral information data of the pixels of different clusters after clustering are taken as input, and a Gaussian filter is used for noise reduction processing. In this embodiment, the filtering parameters are determined when filtering the spectral data of all pixels in the same cluster. Specifically: for each cluster, the coefficient of variation of the spectral data of all pixels at each frequency in the cluster is calculated, and the coefficient of variation represents the fluctuation change characteristics of the spectral scattering and reflection characteristics of the vegetation in the local area; further, in order to accurately adjust the filtering effect of the local area under noise interference, for each cluster, the average value of the normalized results of the coefficient of variation obtained at all frequencies in the cluster is used as the first parameter; the average value of the normalized results of the difference values ​​affected by the spectral information interference of all pixels in the cluster is used as the second parameter; then the filtering parameters when filtering all pixels in the cluster can be determined by the first parameter and the second parameter. In this embodiment, preferably, the sum of the first parameter and the second parameter is used as the filtering parameter when filtering the spectral data of the pixels in the cluster. If the interference effect of the vegetation information obtained based on the aggregation characteristics of vegetation spectral information and the information change characteristics at different frequencies is greater, the filter parameter is set to be larger to reduce the impact of environmental noise interference on the spectral information data.

[0044] Furthermore, based on the results of the Gaussian filter processing of multispectral data, the Normalized Difference Vegetation Index (NDVI) for each pixel at each time point can be accurately calculated. This NDVI effectively reflects the growth and coverage of vegetation, with a larger NDVI indicating more vigorous vegetation growth at that time point. It should be noted that the specific calculation of the NDVI is well known to those skilled in the art and will not be further described in this example.

[0045] Based on the calculated normalized vegetation index of each pixel at each time point, a vegetation index time series of the normalized vegetation index is obtained. In this embodiment, the vegetation index time series of the pixel is composed of the normalized vegetation index of the pixel at all time points in chronological order, which reflects the change data of the vegetation growth status of each area in the detected area; then, the vegetation characteristic information data of different areas in the detected area is reflected based on the extracted vegetation index time series.

[0046] Step 4: Use the optimal path search algorithm to extract the vegetation growth path, wherein the screening-adjusted characteristic value is obtained based on the average level of the difference value of the spectral information interference of each pixel at all time points, and the initial screening value of each pixel is obtained by using the distribution of the vegetation index time series of each pixel. The adjusted screening value of each pixel is obtained in combination with the characteristic value to screen the initial seed point, and the moving direction of each pixel at the current time point is obtained based on the difference in the normalized vegetation index of each pixel at each time point, and the vegetation growth path is obtained using the optimal path search algorithm.

[0047] In the study of vegetation growth path extraction, based on the vegetation index time series of each pixel obtained by the above process, the A-star algorithm is used in this embodiment to accurately extract the vegetation growth path. It should be noted that although the A-star algorithm is mainly used to search for the optimal path from the starting point to the target point in the graph structure, the key is to use the heuristic function to evaluate the node priority and give priority to exploring the path that is more likely to lead to the target. In the vegetation growth path extraction scenario, each pixel can be regarded as a node in the graph. The process of vegetation extending from the area where one pixel is located to the area where another pixel is located is similar to moving from one node to the adjacent node in the graph. Select a suitable initial seed point as the starting point, and regard the ideal end point of the pixel corresponding to the area with the most vigorous vegetation growth as the target point. By defining the cost function, the effective extraction of the vegetation growth path can be achieved. The specific implementation process is as follows:

[0048] First, when setting the initial seed point, this embodiment optimizes and adjusts the selection of the initial seed point based on the results of the clustering feature analysis of the spectral data, targeting the difference values ​​of the spectral information interference effect of each pixel at different time points. Specifically, the average value of the normalized results of the difference values ​​of the spectral information interference effect of each pixel at all time points is calculated and used as the eigenvalue t for the screening adjustment of each pixel. Among them, the larger the eigenvalue t, the greater the impact of the spectral information interference deviation on the pixel at different time periods. From the analysis of long-term dynamic growth conditions, this deviation may cause significant interference to the extraction and analysis of vegetation growth paths.

[0049] Therefore, in this embodiment, the mean of all elements in the vegetation index time series for each pixel is calculated as the initial screening value for each pixel. The initial screening value is then adjusted based on the screening characteristic value t. That is, the larger the characteristic value of the interference effect, the smaller the corresponding screening value. The specific adjustment formula is: Among them, D i Represents the adjusted screening value of the i-th pixel; d i and t i They represent the initial screening value and the characteristic value of the screening adjustment of the i-th pixel respectively; t vrepresents the eigenvalue of the screening adjustment of the v-th pixel; m represents the number of pixels.

[0050] Furthermore, after the adjustment of the initial screening value is completed according to the above process of this embodiment, all pixels are sorted from large to small according to the adjusted screening value, and a preset percentage of pixels are selected as initial seed points. In this embodiment, the preset percentage is 10%.

[0051] Furthermore, the cost function of the A-star algorithm is defined as C = V(t+1) - V(t), which is the difference in the normalized vegetation index corresponding to the pixel at adjacent time points; where V(t+1) and V(t) represent the normalized vegetation index corresponding to the pixel at time points t+1 and t. The direction of pixel movement is determined based on this cost function: for each pixel, among the adjacent pixels of each pixel, the direction of the pixel with the largest increase in the normalized vegetation index at the current time point compared to the previous time point is selected as the movement direction of each pixel at the current time point. Starting from the initial seed point, the movement of the pixel point is decided according to the above rules at each time point. Through continuous iteration, the position changes of the pixel at different time points are continuously tracked, and a complete vegetation growth path is constructed.

[0052] This embodiment realizes the extraction of vegetation growth paths based on the A-star algorithm. However, unlike the traditional A-star algorithm, when extracting plant growth paths, it can use path planning and extract plant growth paths based on the normalized vegetation index of pixels at different times, avoiding the problem of inaccurate normalized vegetation index calculation results caused by noise interference, and improving the accuracy of vegetation growth extension path extraction.

[0053] Based on the same inventive concept as the above method, an embodiment of the present application also provides a vegetation growth path extraction system based on remote sensing images, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned vegetation growth path extraction methods based on remote sensing images.

[0054] It should be understood that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0055] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0056] The above content is only an implementation method of the present application and is not intended to limit the scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the scope of protection of the present application.

Claims

1. A vegetation growth path extraction method based on remote sensing images, characterized in that: The following steps are involved: Obtain multispectral remote sensing data of vegetation areas at each time point; The spectral curve of each pixel is fitted to the spectral data of each pixel at each frequency. According to the correlation and difference between the spectral curves of each pixel and the pixels in the local area, the difference value of the spectral information interference effect of each pixel at each time point is obtained; All pixels at the same time point are clustered, and based on the fluctuation of the spectral data corresponding to each pixel at the same frequency in each cluster, combined with the average level of the difference value of the spectral information interference of all pixels in each cluster, the filtering parameters when filtering the spectral data of the pixels in the cluster are obtained. The normalized vegetation index of each pixel is calculated based on the filtered multispectral data to obtain the vegetation index time series of each pixel; The optimal path search algorithm is used to extract the vegetation growth path, wherein the screening-adjusted characteristic value is obtained according to the average level of the difference value of the spectral information interference of each pixel at all time points, and the initial screening value of each pixel is obtained by using the distribution of the vegetation index time series of each pixel. The adjusted screening value of each pixel is obtained in combination with the characteristic value to screen the initial seed point, and the moving direction of each pixel at the current time point is obtained according to the difference in the normalized vegetation index of each pixel at each time point, and the vegetation growth path is obtained using the optimal path search algorithm.

2. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The method for obtaining the difference value of the spectral information interference effect of each pixel at each time point is: For any time point, the spectral information interference effect difference value a of the xth pixel at any time point is x The calculation formula is: Where b x,y represents the normalized result of the Pearson correlation coefficient between the spectral curve of the x-th pixel and the spectral curve of the y-th pixel in its comparison window; c x,y represents the DTW distance between the spectral curve of the x-th pixel and the spectral curve of the y-th pixel in its comparison window, and m represents the number of pixels. A window of a preset size centered on each pixel serves as the comparison window for each pixel.

3. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The method for obtaining the filtering parameters is: For each cluster at any time point, the coefficient of variation of the spectral data of all pixels at each frequency in the cluster is calculated, and the average value of the normalized results of the coefficient of variation obtained at all frequencies in the cluster is used as the first parameter. The average value of the normalized results of the difference values ​​affected by the spectral information interference of all pixels is used as the second parameter. The sum of the first parameter and the second parameter is used as the filtering parameter when filtering the spectral data of the pixels in the cluster.

4. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The vegetation index time series of each pixel is composed of the normalized vegetation index of each pixel at all time points arranged in chronological order.

5. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The method for obtaining the characteristic value of the screening adjustment is: The average value of the normalized results of the difference values ​​of the spectral information interference effect of each pixel at all time points was calculated and used as the characteristic value of the screening adjustment of each pixel.

6. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The initial screening value of each pixel is the mean value of all elements in the vegetation index time series of each pixel.

7. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The calculation method of the adjusted screening value of each pixel is: Among them D i Represents the adjusted screening value of the i-th pixel; d i and t i They represent the initial screening value and the screening-adjusted characteristic value of the i-th pixel respectively; t v represents the eigenvalue of the screening adjustment of the v-th pixel; m represents the number of pixels.

8. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The screening of the initial seed points further includes: sorting all pixels from large to small according to the adjusted screening value, and selecting a preset percentage of pixels as the initial seed points.

9. The vegetation growth path extraction method based on remote sensing images according to claim 1, characterized in that: The method for obtaining the moving direction of each pixel at the current time point is: Among the adjacent pixels of each pixel, the direction of the adjacent pixel where the normalized vegetation index at the current time point increases the most compared with the previous time point is selected as the moving direction of each pixel at the current time point.

10. A vegetation growth path extraction system based on remote sensing images, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the vegetation growth path extraction method based on remote sensing images as described in any one of claims 1 to 9 are implemented.

Images