Hyperspectral satellite-based mudflat shoreline fine extraction method for mudflat in turbid water area

Abstract

The invention provides a turbid water area mud flat shoreline fine extraction method based on a hyperspectral satellite, and relates to the technical field of image processing, and the method comprises the following steps: S1, obtaining a plurality of Landsat series satellite original hyperspectral satellite remote sensing images according to a research target area and research time; s2, obtaining a preprocessed image; s3, performing multi-scale segmentation on the preprocessed image and calculating a normalized differential water body index of image data; and S4, performing threshold classification on the image after multi-scale segmentation in combination with a normalized differential water body index, adjusting and setting a proper normalized differential water body index, distinguishing a water body from a non-water body, and extracting a shoreline. According to the method, the long-time change rule of the coastline can be better researched and analyzed, the change trend of the coastline can be predicted, a basis is provided for development of coastal areas, and the method has a good application prospect in the fields of coastline research, coastline engineering construction and the like.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method for extracting turbid water tidal flat shorelines in a refined manner based on a hyperspectral satellite. Background Art

[0002] With the continuous development of coastal areas over the years, coastlines have undergone significant changes. As a key topographic feature, accurate coastline data is crucial for analyzing long-term coastline changes. With the continuous improvement and development of coastline mapping technology, common surveying instruments such as total stations and rangefinders can measure the coordinates of various locations. These coordinates can be used to construct complete coastline maps. However, these methods are labor-intensive, require long measurement cycles, and are susceptible to environmental and weather constraints, making large-scale coastline mapping difficult. In recent years, satellite remote sensing technology has rapidly developed, and coastline mapping using satellite remote sensing imagery has become a new approach to coastline mapping. Compared to traditional instrumental mapping, satellite remote sensing shoreline mapping is not only less expensive but also meets the need for long-term, large-scale shoreline observations over a specific area over long time spans. It is not limited by natural conditions such as regional location and environment, and can achieve high temporal and spatial resolution.

[0003] In remote sensing image processing, direct spectral comparison methods have been widely used for coastline extraction. These methods involve comparing raw data images through a series of algorithmic transformations to obtain comparison results. However, these methods are limited by low spatial resolution and sensitivity to environmental factors such as waves and tides. In turbid waters with numerous mudflats, such as those in the Yalu River estuary and the Yangtze River estuary, this accuracy can be even worse, necessitating further refinement of the extraction. The analysis of coastlines over long time series requires accurate coastline documentation, which is essential for studying nearshore coastline change. Therefore, to address the problem of poor nearshore coastline accuracy in turbid waters, a method for refining mudflat shoreline extraction in these waters is urgently needed. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to propose a method for fine-grained extraction of turbid water mudflat coastlines based on hyperspectral satellites to solve the technical problem of poor accuracy of existing coastline extraction methods.

[0005] The technical means adopted in the present invention are as follows:

[0006] A hyperspectral satellite-based method for extracting shorelines from muddy waters and mudflats comprises the following steps:

[0007] S1. Acquire several original hyperspectral satellite remote sensing images from Landsat series satellites for the research target area and research time;

[0008] S2, using remote sensing image processing tools to preprocess the original hyperspectral satellite remote sensing image to obtain a preprocessed image;

[0009] S3. Use remote sensing image processing tools to perform multi-scale segmentation on the pre-processed image and calculate the normalized difference water index of the image data;

[0010] S4. Combine the normalized difference water index to perform threshold classification on the multi-scale segmented image, adjust and set the appropriate normalized difference water index, distinguish water bodies from non-water bodies and extract the coastline.

[0011] Furthermore, in S1, the selection criteria for the original hyperspectral satellite remote sensing images from the Landsat series satellites include: cloud coverage near the study area is less than 10%, the image acquisition date is in the same tidal period each year, and the tidal level difference is less than 1 meter.

[0012] Furthermore, the specific steps of S2 are as follows:

[0013] S21. Import the original hyperspectral satellite remote sensing images of the Landsat series satellites into image processing related tools to convert the radiation data in the images into water surface remote sensing reflectance data;

[0014] S22, performing geometric correction on the image data processed in S21 to eliminate or correct geometric errors of the remote sensing image;

[0015] S23, performing radiometric calibration on the multiple image data processed in S22, and converting the digital quantization values ​​of the images into radiometric brightness values;

[0016] S24. Perform atmospheric correction on the multiple image data processed in S23 to eliminate radiation errors caused by atmospheric influence.

[0017] Furthermore, S3 specifically includes the following steps:

[0018] S31, importing the pre-processed image into a remote sensing image processing tool, selecting corresponding red band, green band, and blue band for combination according to the band data contained in different satellite image data, and observing the image;

[0019] S32, establishing a multi-scale segmentation process;

[0020] S33. Set the image layer weight, segmentation scale, and homogeneity standard parameters in the multi-scale segmentation editing interface;

[0021] S34, executing a multi-scale segmentation process to segment the remote sensing image into a large number of small unit image data;

[0022] S35, observe the remote sensing image segmentation effect and adjust the parameters in S33;

[0023] S36. Based on the large amount of small unit image data adjusted in S35, a feature-corrected normalized difference water index is created, and the normalized difference water index of the image data is calculated. The calculation method of the corrected normalized difference water index is as follows:

[0024]

[0025] Where: MNDWI represents the green band of the modified normalized difference water index, R rs (Green) represents the green band in Landsat satellite remote sensing images, R rs (SWIR1) represents the first shortwave infrared band of Landsat satellite remote sensing images.

[0026] Furthermore, S4 specifically includes the following steps:

[0027] S41. Open the fill option of the object layer and observe;

[0028] S42, adjusting the value range of the normalized difference water index until the water body and the non-water body are separated;

[0029] S43, after determining the value range of S42, a new classification process is created, and the MNDWI value range obtained in S42 is used to classify water bodies;

[0030] S44. Create a new process to output vector layer and select the shp format file to output the water body range;

[0031] S45. Import the shp file into a remote sensing image processing tool for editing and processing to obtain accurately extracted shoreline data.

[0032] The present invention also provides a storage medium, which includes a stored program, wherein when the program is run, any of the above-mentioned methods for fine-tuning the extraction of turbid water area mudflats and coastlines based on hyperspectral satellites is executed.

[0033] The present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes any one of the above-mentioned methods for fine-tuning the extraction of turbid water area mudflats and coastlines based on hyperspectral satellites through the computer program.

[0034] Compared with the prior art, the present invention has the following advantages:

[0035] The Landsat series satellite remote sensing images used in this method are easy to obtain, have relatively rich spectral band information, and have high resolution, which can meet the water depth dataset requirements in most application scenarios;

[0036] Compared with traditional instrument mapping, this method of satellite remote sensing shoreline mapping is not only less expensive, but also can meet the needs of long-term, large-scale shoreline observation of a certain area of ​​water over a long period of time. It is not restricted by natural conditions such as regional location and environment, and the mapping accuracy can achieve a high temporal and spatial resolution.

[0037] This method uses the modified normalized difference water index (MNDWI). Compared with the general normalized difference water index (NDWI), it combines additional bands or spectral indices to better distinguish water and non-water surfaces in turbid waters. This method provides higher accuracy and can detect small-scale coastline changes. The NDWI is improved by replacing the near-infrared band (NIR) (0.76-0.90μm) band in the original NDWI with the shortwave infrared (SWIR1) band to achieve better shoreline extraction effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0039] Figure 1 is a flow chart of the method of the present invention;

[0040] Figure 2 A location map of the Yalu River estuary area provided in an embodiment of the present invention;

[0041] Figure 3 For the embodiment of the present invention, eCognition software is used for multi-scale segmentation;

[0042] Figure 4 For the embodiment of the present invention, eCognition software is used for threshold classification;

[0043] Figure 5 The shoreline data in shp format extracted by the eCognition software in the embodiment of the present invention;

[0044] Figure 6 The refined Yalu River shoreline data of different years extracted and processed according to the embodiment of the present invention; DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0046] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0047] like Figure 1 As shown, the present invention provides a method for extracting shorelines of turbid waters and mudflats based on hyperspectral satellites, comprising the following steps:

[0048] S1. Acquire several original hyperspectral satellite remote sensing images from Landsat series satellites for the target research area and research period. The images are required to basically meet the selection criteria: (1) cloud coverage near the research area is less than 10%, (2) the image acquisition date is at the same tidal period each year, and (3) the tidal level difference is less than 1 meter.

[0049] S11. Register an account and log in to the Geospatial Data Cloud (https: / / www.gscloud.cn / ) or the U.S. Geological Survey website (https: / / earthexplorer.usgs.gov / );

[0050] S12. Select the required satellite image data according to the image selection area, time and standard. For example, the area is as follows Figure 2 As shown, the cloud cover of the selected images was below 10%, and the image selection date and tidal stage were similar for each year;

[0051] S13, download the corresponding satellite image data and import it into the remote sensing image processing tool Envi for viewing;

[0052] S14: Eliminate satellite image data with poor quality, reselect appropriate satellite image data for download and viewing, and repeat the steps until the collected image data set meets the requirements.

[0053] S2, using the remote sensing image processing platform Envi to preprocess the original hyperspectral satellite remote sensing image to obtain a preprocessed image;

[0054] S2. Use remote sensing image processing tools and remote sensing image processing platform Envi to pre-process the original hyperspectral satellite remote sensing images of the Landsat series;

[0055] S21. In order to convert the radiation data in the image into water surface remote sensing reflectance data, the acquired image data is imported into the remote sensing image processing tool environmental visualization image software Envi;

[0056] S22. Select the image registration workflow in Envi and use Envi to perform geometric correction on the acquired multiple image data to eliminate or correct the geometric errors of the remote sensing images. The remote sensing data of some satellites, such as Landsat-8, have already been geometrically corrected, so you can choose to skip this step based on the satellite image properties.

[0057] S23, select the radiometric calibration module in Envi, use Envi to perform radiometric calibration on the acquired multiple image data, and convert the digital quantization value of the image into physical quantities such as radiometric brightness value;

[0058] S24. Select the multispectral data FLAASH atmospheric correction module in Envi, and use Envi to perform atmospheric correction on the acquired multiple image data to eliminate the radiation error caused by atmospheric influence.

[0059] S3. Use remote sensing image processing tool eCognition to perform multi-scale segmentation on the pre-processed image and calculate the normalized differential water index of the image data;

[0060] S31, importing the satellite image dataset processed in S2 into the remote sensing image processing tool remote sensing image processing platform eCognition, selecting the corresponding red band, green band, and blue band according to the band data contained in different satellite image data, and observing the image;

[0061] S32, establishing a multi-resolution segmentation process;

[0062] S33. Set the image layer weight, segmentation scale, and homogeneity standard parameters in the multi-scale segmentation editing interface;

[0063] S34, executing a multi-scale segmentation process to segment the remote sensing image into a large number of small unit image data and opening the outline of the object levels;

[0064] S35. Observe the effect of remote sensing image segmentation and adjust the parameters in S33. Adjust the corresponding parameters of multi-scale segmentation until the image segmentation effect is accurate. Set Image Layer weight to 1, 1, 1, 1, 1, 1, Scale parameter to 42, shape to 0.1, and compactness to 0.5. Figure 3 This is an example of the results of multi-scale segmentation using eCognition software in an embodiment of the present invention;

[0065] S36. Based on the large amount of small unit image data adjusted in S35, a modified normalized difference water index (MNDWI) is created for the target object under the custom feature in the feature view interface, and the MNDWI value of the image data is calculated. The calculation method of the modified normalized difference water index is as follows:

[0066]

[0067] Where: MNDWI represents the green band of the modified normalized difference water index, R rs (Green) represents the green band in Landsat satellite remote sensing images, R rs (SWIR1) represents the first shortwave infrared band (SWIR1) of Landsat satellite remote sensing images.

[0068] S4. Combine the normalized difference water index to perform threshold classification on the multi-scale segmented image, adjust and set the appropriate normalized difference water index, distinguish water bodies from non-water bodies and extract the coastline. Figure 4 This is an illustration of threshold classification using eCognition software in an embodiment of the present invention.

[0069] S41: Turn on the fill option of the object levels for observation;

[0070] S42: Adjust the value range of MNDWI until the water body and non-water body separation effect is good;

[0071] S43: After the range is determined, a new classification (assign class) process is created in the decision tree window to classify water bodies using the specified MNDWI value range;

[0072] S44: Create a new export vector layer process in the decision tree window and select the shp format file for outputting the water body range. Figure 5 The water body data in shp format extracted by the eCognition software in the embodiment of the present invention;

[0073] S45: Import the shp file into the ArcGIS remote sensing image processing tool for editing and processing to obtain the accurately extracted shoreline data. Figure 6 The refined Yalu River shoreline data of different years after processing is extracted by the embodiment of the present invention, and the area is divided into five different parts: S1, S2, S3, S4 and S5. The S1 part covers the aquaculture area in the western part of the Yalu River estuary, the S2 section includes the Dandong Port construction area, the S3 section is the inland estuary area, the S4 section constitutes the east bank of the Yalu River estuary, and S5 is the Silk Island area.

[0074] In order to evaluate the applicability of the present invention, the shoreline data change analysis conducted in conjunction with this embodiment is as follows;

[0075] The Digital Shoreline Analysis System (DSAS) in ArcGIS was used to analyze the shorelines in 2000, 2010, and 2020, yielding the End Point Rate (EPR) and Net Shoreline Movement (NSM) of shoreline change over the 20-year period. Table 1 shows the change in the length of the Yalu River shoreline from 2000 to 2020. Table 2 shows the change in the erosion or accumulation trend of the Yalu River shoreline from 2000 to 2020.

[0076] Table 1 Changes in the length of the Yalu River coastline from 2000 to 2020

[0077]

[0078]

[0079] Table 2 Changes in erosion or accumulation trends along the Yalu River coastline from 2000 to 2020

[0080]

[0081] The EPR focuses on the rate of change over time, providing a temporal perspective, showing the annual rate of change over the above period; while the NSM provides a spatial perspective by quantifying the total movement of the coastline, measuring the long-term net coastline change between 2000 and 2020.

[0082] Changes in coastline length

[0083] As can be seen from the table, the total length of coastline along sections S1 to S4 increased from 2000 to 2020, with the total length of sections S1 to S4 increasing by 59 km. This increase represents a 22.6% increase compared to 2000, with an average annual increase of 2.95 km. The most significant change occurred between 2010 and 2020, with coastline length increasing by 45.5 km. Looking at cross-sectional changes, the coastline along section S2 increased by 396.9% of its original length from 2000 to 2020 due to the construction of Dandong Port. Therefore, the primary cause of the change in coastline length in the Yalu River estuary region from 2000 to 2020 was the increase in artificial coastline resulting from local port development.

[0084] Shoreline accumulation / erosion changes

[0085] From 2000 to 2020, the overall coastline of the Yalu River showed an accumulation trend. The temporal and spatial distribution of EPR is heterogeneous. Comparing the sections, S3 (inside the estuary) eroded slightly, with the highest erosion rate of 57.05m per year and an average net coastline erosion of 38.37m. Other parts showed accumulation, among which the construction activities of the local port (S2) and land reclamation in part of S4 contributed particularly significantly to the accumulation of the coastline. In addition, S5 (Satin Island) experienced slight erosion during the period 2000-2020, with a maximum erosion rate of 24.99m per year and an average net coastline erosion of 52.49m. From an anthropogenic perspective, aquaculture areas, ports and reclamation areas are the main projects of artificial coastline construction and the main influencing factors of coastline expansion. Due to the construction of a large number of artificial buildings, the Yalu River coastline has been greatly affected. From 2000 to 2020, the length of the coastline in the S1-S4 section increased by 22.6%. The accumulation of the coastline in this area is also mainly caused by human activities. Human activities are the main contributor to the accumulation of the coastline at the Yalu River estuary.

[0086] The long-term series of turbid waters and refined mudflat coastlines obtained using this method can facilitate coastline observation and analysis in a more convenient, rapid, time-saving and labor-saving manner.

[0087] To sum up, in the above embodiments of the present application, a method for fine-grained extraction of turbid water flat coastlines based on hyperspectral satellites is adopted, which mainly includes utilizing Envi, eCognition, and ArcGIS image processing software, the spectral characteristics and spatial characteristics of the Landsat series hyperspectral satellite remote sensing images, and combining a common water body index - the Normalized Difference Water Index (NDWI) and improving it to the modified Normalized Difference Water Index (MNDWI). The hyperspectral satellite image data is subjected to data preprocessing, image segmentation, threshold classification, and coastline extraction, which can effectively and finely extract the turbid water coastline, and has good application prospects in the field of turbid water coastline extraction and analysis under long time series.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for extracting mudflat shorelines in turbid waters based on hyperspectral satellites, characterized in that: The steps include: S1. Acquire several original hyperspectral satellite remote sensing images from Landsat series satellites for the research target area and research time; S2, using remote sensing image processing tools to preprocess the original hyperspectral satellite remote sensing image to obtain a preprocessed image; S3. Use remote sensing image processing tools to perform multi-scale segmentation on the pre-processed image and calculate the normalized difference water index of the image data; S4. Combine the normalized difference water index to perform threshold classification on the multi-scale segmented image, adjust and set the appropriate normalized difference water index, distinguish water bodies from non-water bodies and extract the coastline.

2. The method for extracting turbid waters and mudflats based on hyperspectral satellites according to claim 1 is characterized in that: In S1, the selection criteria for the original hyperspectral satellite remote sensing images from the Landsat series satellites include: cloud coverage near the study area is less than 10%, the image acquisition date is at the same tidal period each year, and the tidal level difference is less than 1 meter.

3. The method for extracting turbid waters and mudflats based on hyperspectral satellites according to claim 1 is characterized in that: The specific steps of S2 are as follows: S21. Import the original hyperspectral satellite remote sensing images of the Landsat series satellites into image processing related tools to convert the radiation data in the images into water surface remote sensing reflectance data; S22, performing geometric correction on the image data processed in S21 to eliminate or correct geometric errors of the remote sensing image; S23, performing radiometric calibration on the multiple image data processed in S22, and converting the digital quantization values ​​of the images into radiometric brightness values; S24. Perform atmospheric correction on the multiple image data processed in S23 to eliminate radiation errors caused by atmospheric influence.

4. The method for extracting turbid waters and mudflats based on hyperspectral satellites according to claim 1 is characterized in that: S3 specifically includes the following steps: S31, importing the pre-processed image into a remote sensing image processing tool, selecting corresponding red band, green band, and blue band for combination according to the band data contained in different satellite image data, and observing the image; S32, establishing a multi-scale segmentation process; S33. Set the image layer weight, segmentation scale, and homogeneity standard parameters in the multi-scale segmentation editing interface; S34, executing a multi-scale segmentation process to segment the remote sensing image into a large number of small unit image data; S35, observe the remote sensing image segmentation effect and adjust the parameters in S33; S36. Based on the large amount of small unit image data adjusted in S35, a feature-corrected normalized difference water index is created, and the normalized difference water index of the image data is calculated. The calculation method of the corrected normalized difference water index is as follows: Where: MNDWI represents the green band of the modified normalized difference water index, R rs (Green) represents the green band in Landsat satellite remote sensing images, R rs (SWIR1) represents the first shortwave infrared band of Landsat satellite remote sensing images.

5. The method for extracting turbid waters and mudflats based on hyperspectral satellites according to claim 1 is characterized in that: S4 specifically includes the following steps: S41. Open the fill option of the object layer and observe; S42, adjusting the value range of the normalized difference water index until the water body and the non-water body are separated; S43, after determining the value range of S42, a new classification process is created, and the MNDWI value range obtained in S42 is used to classify water bodies; S44. Create a new process to output vector layer and select the shp format file to output the water body range; S45. Import the shp file into a remote sensing image processing tool for editing and processing to obtain accurately extracted shoreline data.

6. A storage medium, characterized in that The storage medium includes a stored program, wherein when the program is run, the method for fine-grained extraction of turbid water tidal flat shorelines based on hyperspectral satellite according to any one of claims 1 to 5 is executed.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: The processor executes the method for fine-grained extraction of turbid water tidal flat shorelines based on hyperspectral satellites as described in any one of claims 1 to 5 through the operation of the computer program.

Images