Method for acquiring instantaneous coastline position based on optical remote sensing image

Abstract

The invention provides a method for obtaining an instantaneous coastline position based on an optical remote sensing image, and belongs to the technical field of image processing, and the method comprises the following steps: 1, obtaining an MGWI water body index image according to the optical remote sensing image; 2, processing the MGWI water body index image by using a clustering algorithm to obtain a sea-land segmentation binary image; step 3, preliminarily extracting boundary information according to the sea-land segmentation binary image, obtaining an image boundary possibly existing in the image, and obtaining a preliminary instantaneous coastline; and step 4, processing the initial instantaneous coastline, obtaining a boundary between the sea and the land in the sea-land segmentation binary image, and obtaining an instantaneous coastline position in the remote sensing image. According to the method for obtaining the instantaneous coastline position based on the optical remote sensing image, the difference between the water body and the land in the optical remote sensing image can be highlighted. The instant coastline position in the remote sensing image is rapidly obtained, and the efficiency of obtaining the coastline position based on the remote sensing image is greatly improved.

Description

Technical Field

[0001] The invention belongs to the field of surveying technology, and in particular relates to a method for acquiring instantaneous coastline position based on optical remote sensing images. Background Art

[0002] The change of coastline is a dynamic process, which is easily affected by natural and human factors such as tidal currents, coastal topography, hydrodynamic processes, marine engineering, coastal reclamation, etc. In coastal systems operating under natural conditions, human pressure is very low or even non-existent, and accretion and erosion of the coastline will not immediately cause risks. However, when human activities are introduced, the change of coastline may have far-reaching effects. In recent years, the frequent human transformation of the coastline and the impact of changes in natural factors such as global warming have accelerated the change of coastline.

[0003] Affected by the tide, the position of the coastline changes dynamically. The position of the coastline on the remote sensing image is the instantaneous coastline position when the remote sensing image is imaged. The traditional method of obtaining the coastline position based on remote sensing images is mainly manual visual interpretation. This method relies on manual interpretation and is inefficient when facing the problem of obtaining the coastline position over a large range. With the continuous innovation of computer technology and remote sensing technology, the automatic interpretation ability of remote sensing images has gradually increased, and various automatic interpretation algorithms have been continuously applied to the work of obtaining the coastline position, such as water body index method, object-oriented, edge detection and other methods. However, the edge detection method is easily affected by noise and is more suitable for constructing simple coastlines; the object-oriented method has high requirements for data, and the acquisition rules are relatively complex and have poor real-time performance.

[0004] The purpose of calculating the water index is to highlight the difference between the ocean and the land, so as to obtain the boundary between the ocean and the land as the coastline. The location of the coastline on the remote sensing image will also be obscured by clouds and the sea is frozen, resulting in blurred boundaries, which will affect the accuracy of the coastline location. Therefore, the technology needs to be further improved and developed. Summary of the invention

[0005] In order to solve the technical problems that the existing coastline position detection methods are easily affected by noise, and the object-oriented methods have high requirements on data, and the acquisition rules are relatively complex and have poor real-time performance, the present invention proposes a method for obtaining the instantaneous coastline position based on optical remote sensing images, which can solve the above problems.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions to achieve the above problems: A method for obtaining instantaneous coastline position based on optical remote sensing images, comprising the following steps: Step 1: Obtain MGWI water index images based on optical remote sensing images, including: (11) Obtain green band data and near infrared band data from optical remote sensing images respectively; (12) Calculate the green band data Near infrared band data The difference between , get the difference matrix; (13) Calculate the parameter matrix based on the difference matrix ; (14) According to the parameter matrix The MGWI water index matrix, i.e., the MGWI water index image, is calculated; Step 2: Use clustering algorithm to process the MGWI water index image to obtain a binary map of land and sea segmentation; Step 3: Preliminarily extract boundary information based on the land-sea segmentation binary map, obtain possible image boundaries in the image, and obtain a preliminary instantaneous coastline; Step 4: Process the preliminary instantaneous coastline to obtain the boundary between the ocean and the land in the ocean-land segmentation binary map, and obtain the instantaneous coastline position in the remote sensing image.

[0007] In some embodiments, step 2 includes: (21) Use the K-means clustering algorithm to perform binary clustering on the MGWI water index image, and normalize the clustering results to obtain a preliminary land-sea segmentation binary map; (22) Using the K-means clustering algorithm to process the preliminary land-sea segmentation binary map, and removing patches caused by terrestrial water bodies in the land part; (23) Perform morphological closing operation on the image obtained in step (22) to smooth the coastline edge and obtain a binary image of land and sea segmentation.

[0008] In some embodiments, in step (21), the cluster centers of the two types of data after clustering are sorted by size, the data whose cluster center values ​​are greater than a set threshold are classified as the ocean part and assigned a first pixel value, and the data whose cluster center values ​​are not greater than the set threshold are classified as the land part and assigned a second pixel value, thereby obtaining a preliminary land-sea segmentation binary map.

[0009] In some embodiments, step (22) includes: A sliding window is established in the preliminary land-sea segmentation binary map. When at least one pixel value in the sliding window is the first pixel value, it is recorded and the sliding window is slid around until the sliding window does not have the first pixel value, and the pixel values ​​of all pixels surrounded by the first pixel value are filled with the first pixel value.

[0010] In some embodiments, step (23) includes: Widen the image obtained in step (22), and fill a circle of pixels with the second pixel value around the image; The filled image is closed and the filled circle of pixels is removed to obtain a binary image of land and sea segmentation.

[0011] In some embodiments, step three includes: (31) Use the raster-to-feature operation to convert the land-sea segmentation binary map into a surface feature; (32) Obtain the surface element of the land part, and delete the surface element whose pixel value is the second pixel value; (33) Convert the surface elements of the land part into line elements to obtain the instantaneous coastline position in the remote sensing image where the image boundary may exist, which is the preliminary instantaneous coastline.

[0012] In some embodiments, step 4 includes: (41) Obtain the remote sensing image boundary based on the optical remote sensing image and generate the corresponding line element; (42) superimposing the remote sensing image boundary with the preliminary instantaneous coastline to obtain a full-border image including the remote sensing image boundary and the preliminary instantaneous coastline; (43) The pixels at the border of the remote sensing image in the full-border image are inverted to obtain the instantaneous coastline position in the remote sensing image.

[0013] In some embodiments, step (43) further includes searching the length of the line elements of the image after the inversion operation, and retaining the longest line element, which is the instantaneous coastline position in the remote sensing image.

[0014] In some embodiments, the parameter matrix in step (13) The calculation method is: ; The calculation method of the MGWI water index matrix in step (14) is: ; Wherein, n is a preset constant.

[0015] In some embodiments, when the closed operation is performed on the padded image in step (23), the operator used is that two adjacent values ​​located on the diagonal line are different from each other.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are: the method for obtaining the instantaneous coastline position based on optical remote sensing images of the present invention is based on the green band data in the optical remote sensing images. Near infrared band data The difference calculation of MGWI water index image can highlight the difference between water and land in optical remote sensing images. Combined with clustering algorithm, the instantaneous coastline position in remote sensing images can be quickly obtained, which greatly improves the efficiency of obtaining coastline position based on remote sensing images.

[0017] Other features and advantages of the present invention will become more apparent after reading the detailed description of the embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of a method for obtaining instantaneous coastline position based on optical remote sensing images provided by an embodiment of the present invention; Figure 2 is a process flow chart of removing plaques using a sliding window in one embodiment of the present invention; Figure 3 is a closing operator used in one embodiment of the present invention. DETAILED DESCRIPTION

[0019] The specific implementation modes of the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0020] Embodiment 1: This embodiment proposes a method for obtaining the instantaneous coastline position based on optical remote sensing images. Figure 1 As shown, the following steps are included: Step 1: Obtain MGWI water index images based on optical remote sensing images, including:

[0021] S11, respectively obtaining green band data and near infrared band data in the optical remote sensing image. The band numbers of the green band and near infrared band of the optical remote sensing data can be determined according to different sensor data, and then the green band data and near infrared band data in the optical remote sensing image can be obtained.

[0022] S12. Calculate green band data Near infrared band data The difference between the green band and the green band data is obtained. Near infrared band data The difference can highlight the difference between water and land in optical remote sensing images, which is more conducive to the segmentation of land and sea.

[0023] S13, calculate the parameter matrix based on the difference matrix .

[0024] S14, according to the parameter matrix The MGWI water index matrix, that is, the MGWI water index image, is calculated.

[0025] Step 2: Use clustering algorithm to process the MGWI water index image to obtain a binary map of land and sea segmentation.

[0026] Step three, preliminarily extract boundary information based on the land-sea segmentation binary map, obtain the image boundaries that may exist in the image, and obtain the preliminary instantaneous coastline, thus realizing the preliminary extraction of the instantaneous coastline, which is beneficial to the efficiency of subsequent extraction steps.

[0027] Step 4: Process the preliminary instantaneous coastline to obtain the boundary between the ocean and land in the land-sea segmentation binary map, and obtain the instantaneous coastline position in the remote sensing image. This solution quickly realizes the acquisition of the instantaneous coastline position in the remote sensing image through preliminary and fine extraction of the instantaneous coastline, greatly improving the efficiency of obtaining the coastline position based on remote sensing images.

[0028] This embodiment was tested with 78 scenes of Landsat satellite series data in Jiaozhou Bay area over 34 years, so the coastline positions were successfully obtained. The accuracy of instantaneous coastline position acquisition is better than the resolution of one data pixel. It is proved that the proposed method can effectively obtain coastline position information, greatly reducing manpower consumption, and providing a new technical means for obtaining coastline position of remote sensing images.

[0029] In some embodiments, step 2 includes: S21. Use the K-means clustering algorithm to perform binary clustering on the MGWI water index image, and normalize the clustering results to obtain a preliminary land-sea segmentation binary map.

[0030] S22. Use the K-means clustering algorithm to process the preliminary land-sea segmentation binary map to remove patches caused by terrestrial water bodies in the land part.

[0031] S23, performing morphological closing operation on the image obtained in step S22, smoothing the edge of the coastline, and obtaining a binary image of land and sea segmentation.

[0032] In some embodiments, in step S21, the cluster centers of the two types of data after clustering are sorted by size, and the data with a cluster center value greater than a set threshold is classified as the ocean part and assigned a first pixel value, and the data with a cluster center value not greater than the set threshold is classified as the land part and assigned a second pixel value, thereby obtaining a preliminary land-sea segmentation binary map. In the MGWI water index image, the pixel values ​​of the land part are small, and the pixel values ​​of the ocean part are large. This scheme sorts the cluster centers of the two types after clustering by size, and defines the class with a larger cluster center value as the ocean part; and defines the class with a smaller cluster center value as the land part, thereby obtaining a preliminary land-sea segmentation binary map.

[0033] In order to improve classification efficiency and classification accuracy, in some embodiments, K-means processes the preliminary land-sea segmentation binary map using an improved flood filling algorithm. Specifically, step S22 includes: A sliding window is established in the preliminary land-sea segmentation binary map. When at least one pixel value in the sliding window is the first pixel value, it is recorded and the sliding window is slid around until the sliding window does not have the first pixel value, and the pixel values ​​of all pixels surrounded by the first pixel value are filled with the first pixel value.

[0034] The principle of improving the flood filling algorithm is as follows Figure 2 As shown: a 2×2 sliding window is established from the upper left corner, such as the part in the red box in the figure. If at least one pixel in this window has a value of the first pixel, it is recorded and the window is moved left and downward until there is no pixel value in the window that is the first pixel value. All pixels over which the window has slid are regarded as an individual, such as the part in the blue box in the figure. These individuals are filled with the flood algorithm separately, and the pixel values ​​of all the pixels inside the window surrounded by the first pixel value are filled with the first pixel value.

[0035] Since the image has a boundary, there may be a coastline at the boundary position, or there may be no coastline, and at least one side adjacent to the boundary position no longer has a pixel value. In order to be able to identify the boundary portion of the image, in some embodiments, step S23 includes: The image obtained in step S22 is widened, and a circle of pixels with the second pixel value is filled around the image. By widening the image, pixels around the image boundary can be filled with pixels, so as to achieve the purpose of identifying the coastline affected by the image boundary.

[0036] The filled image is closed and the filled circle of pixels is removed to obtain a binary image of land and sea segmentation.

[0037] The morphological closing operation is performed on the preliminary land-sea segmentation binary image to smooth the coastline edge. In the calculation, the pixels at the image boundary are first filled with a circle of pixels with a width of 1 and a pixel value of 0 to protect the coastline information near the image boundary. Then the closing operation is performed. Since the coastline is usually not a simple straight line, but has a curved or branched shape, the operator used in this embodiment is as follows Figure 3 As shown in the figure, the operator focuses on the pixel relationship in the diagonal direction, which can better maintain or repair the diagonal shape or lines in the image without over-smoothing the coastline. On this basis, it eliminates small protrusions and better preserves the geometric features of the coastline.

[0038] In some embodiments, step three includes: S31. Use the raster-to-feature operation to convert the land-sea segmentation binary map into a surface feature. Use the raster-to-feature operation to convert the smoothed land-sea segmentation binary map into a surface feature. Add a "value" attribute to each surface feature to store the pixel value of the corresponding raster position of the surface feature.

[0039] S32, obtain the area elements of the land part, and delete the area elements whose pixel value is the second pixel value. Search with the "value" attribute, and delete the area elements whose "value" value is the second pixel value.

[0040] S33, converting the area elements of the land part into line elements, obtaining the instantaneous coastline position in the remote sensing image where the image boundary may exist, which is the preliminary instantaneous coastline. Converting the area element information of the land part into line elements, obtaining the instantaneous coastline position in the remote sensing image where the image boundary may exist.

[0041] In some embodiments, step 4 includes: S41. Acquire the remote sensing image boundary according to the optical remote sensing image, and generate corresponding line elements.

[0042] S42, superimposing the remote sensing image boundary with the preliminary instantaneous coastline to obtain a full-boundary image including the remote sensing image boundary and the preliminary instantaneous coastline.

[0043] S43, performing an inversion operation on pixels at the remote sensing image boundary in the full boundary image to remove the remote sensing image boundary and obtain the instantaneous coastline position in the remote sensing image containing only the preliminary instantaneous coastline.

[0044] In some embodiments, step S43 further includes searching the length of the line elements of the image after the inversion operation, and retaining the longest line element, which is the instantaneous coastline position in the remote sensing image.

[0045] Compared with the classic water body index NDWI, the method for obtaining the instantaneous coastline position based on optical remote sensing images proposed in this embodiment contains more edge information and is more suitable for obtaining the coastline position. Compared with the high-precision coastline position information interpreted visually, the error is less than one data pixel resolution and the accuracy is higher.

[0046] In step 1, the Gaussian kernel function (RBF) is introduced, and the expression is: .

[0047] In order to make the Gaussian kernel function suitable for optical remote sensing image processing, the Gaussian kernel function is improved by combining the advantages of the kernel function in dealing with nonlinear problems and the idea of ​​similarity of the Gaussian kernel function, and a new optical remote sensing water index MGWI is proposed to highlight the difference between water and land in optical remote sensing images.

[0048] In some embodiments, the parameter matrix in step S13 The calculation method is: .

[0049] The calculation method of the MGWI water index matrix in step S14 is: .

[0050] Wherein, n is a preset constant.

[0051] To adjust the parameters of MGWI's sensitivity to water content, The value of depends on the quality of the image. The minimum value of n is 10, and the value of n increases as the quality of the remote sensing image deteriorates.

[0052] In some embodiments, when the closed operation is performed on the filled image in step S23, the operator used is that two adjacent values ​​located on the diagonal line are different from each other.

[0053] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by ordinary technicians in this technical field within the essential scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A method for obtaining instantaneous coastline position based on optical remote sensing images, characterized in that: The following steps are involved: Step 1: Obtain MGWI water index images based on optical remote sensing images, including: (11) Obtain green band data and near infrared band data from optical remote sensing images respectively; (12) Calculate the green band data Near infrared band data The difference between , get the difference matrix; (13) Calculate the parameter matrix based on the difference matrix ; (14) According to the parameter matrix The MGWI water index matrix, i.e., the MGWI water index image, is calculated; Step 2: Use clustering algorithm to process the MGWI water index image to obtain a binary map of land and sea segmentation; Step 3: Preliminarily extract boundary information based on the land-sea segmentation binary map, obtain possible image boundaries in the image, and obtain a preliminary instantaneous coastline; Step 4: Process the preliminary instantaneous coastline to obtain the boundary between the ocean and the land in the ocean-land segmentation binary map, and obtain the instantaneous coastline position in the remote sensing image.

2. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 1, characterized in that: Step 2 includes: (21) Use the K-means clustering algorithm to perform binary clustering on the MGWI water index image, and normalize the clustering results to obtain a preliminary land-sea segmentation binary map; (22) Using the K-means clustering algorithm to process the preliminary land-sea segmentation binary map, and removing patches caused by terrestrial water bodies in the land part; (23) Perform morphological closing operation on the image obtained in step (22) to smooth the coastline edge and obtain a binary image of land and sea segmentation.

3. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 2, characterized in that: In step (21), the cluster centers of the two types of data after clustering are sorted by size, and the data whose cluster center values ​​are greater than the set threshold are classified as the ocean part and assigned the first pixel value, and the data whose cluster center values ​​are not greater than the set threshold are classified as the land part and assigned the second pixel value, thereby obtaining a preliminary land-sea segmentation binary map.

4. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 2, characterized in that: Step (22) includes: A sliding window is established in the preliminary land-sea segmentation binary map. When at least one pixel value in the sliding window is the first pixel value, it is recorded and the sliding window is slid around until the sliding window does not have the first pixel value, and the pixel values ​​of all pixels surrounded by the first pixel value are filled with the first pixel value.

5. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 2, characterized in that: Step (23) includes: Widen the image obtained in step (22), and fill a circle of pixels with the second pixel value around the image; The filled image is closed and the filled circle of pixels is removed to obtain a binary image of land and sea segmentation.

6. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 3, characterized in that: Step three includes: (31) Use the raster-to-feature operation to convert the land-sea segmentation binary map into a surface feature; (32) Obtain the surface element of the land part, and delete the surface element whose pixel value is the second pixel value; (33) Convert the surface elements of the land part into line elements to obtain the instantaneous coastline position in the remote sensing image where the image boundary may exist, which is the preliminary instantaneous coastline.

7. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 1, characterized in that: Step 4 includes: (41) Obtain the remote sensing image boundary based on the optical remote sensing image and generate the corresponding line element; (42) superimposing the remote sensing image boundary with the preliminary instantaneous coastline to obtain a full-border image including the remote sensing image boundary and the preliminary instantaneous coastline; (43) The pixels at the border of the remote sensing image in the full-border image are inverted to obtain the instantaneous coastline position in the remote sensing image.

8. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 7, characterized in that: Step (43) also includes searching the length of the line elements of the image after the inversion operation, and retaining the longest line element, which is the instantaneous coastline position in the remote sensing image.

9. The method for obtaining instantaneous coastline position based on optical remote sensing images according to any one of claims 1 to 8, characterized in that: The parameter matrix in step (13) The calculation method is: ； The calculation method of the MGWI water index matrix in step (14) is: ； Wherein, n is a preset constant.

10. The method for obtaining instantaneous coastline position based on optical remote sensing images according to claim 5, characterized in that: When the closed operation is performed on the padded image in step (23), the operator used is that the two adjacent values ​​located on the diagonal are different.

Images