Ship lock extraction method and device based on high-resolution remote sensing image

Abstract

The invention relates to a ship lock extraction method and device based on a high-resolution remote sensing image, and the method comprises the steps: extracting a water body image from the high-resolution remote sensing image of a target region; dividing the water body probability data of the target area by using a preset threshold value to obtain a water body area, and determining a target channel or river range in the target area based on the water body area and the water body image; based on the shape priori knowledge of the lock chamber of the ship lock, screening out a lock chamber area meeting a preset suspected condition from the target channel or river range; and performing ship lock extraction in the lock chamber area by combining the spectral characteristics of the ship lock and the prior knowledge of the water body to obtain a ship lock extraction result meeting a preset precision condition. Therefore, the technical problems that in the related technology, when an object with an irregular shape is recognized, the requirement for the image resolution is high, the object is difficult to separate from other various building backgrounds in a complex scene, and the accuracy of a recognition result is low due to the fact that remote sensing image classification is uncertain are solved.

Description

Technical Field

[0001] This application relates to the field of intelligent interpretation technology of remote sensing images, and in particular to a method and apparatus for extracting locks based on high-resolution remote sensing images. Background Technology

[0002] Ship locks are an important component of water conservancy projects, commonly used to overcome water level differences in inland waterways to ensure passage. They serve multiple functions, including regulating water levels and flow, promoting shipping and economic development, protecting water resources, and preventing and mitigating disasters. They are of great significance to water transport, water resource management, and economic development. Utilizing remote sensing to automatically identify ship locks can provide a scientific basis for ship lock water transport service management, hydrological environment assessment, and watershed water resource allocation.

[0003] Ship locks typically consist of upstream and downstream approach channels, upstream and downstream lock heads, lock chambers, and lock gates, and are generally constructed of materials such as concrete and metal, making them typical waterborne structures. Currently, there is considerable research on using remote sensing imagery to identify and extract waterborne structures.

[0004] Among related technologies, automatic identification of typical man-made targets on water can be achieved based on fuzzy theory and prior knowledge, using grayscale and geometric features of high-resolution remote sensing images; binarized edge detection can be performed on road skeleton lines in remote sensing images, and bridge edge lines can be quickly determined based on multi-dimensional information such as bridge features; bridges on water can be extracted based on high-resolution optical remote sensing images using directional enhanced linear structural units and mathematical morphology; bridge recognition models can be constructed by analyzing the topological relationships of bridge areas in remote sensing images, enabling efficient bridge extraction in complex scenes; fuzzy thresholding can be used for land-water segmentation, and bridge boundary and location parameters can be determined based on prior knowledge models; and an improved YOLOv5 convolutional neural network method can be used to accurately identify large ground objects such as bridges, dams, and ports in complex environments.

[0005] However, classification based on remote sensing imagery in related technologies often involves uncertainty. The main sources of this uncertainty include factors such as clouds, shadows, occlusion, and blurred boundaries of ground features. Furthermore, most related technologies are applied to water conservancy facilities with relatively regular shapes and simple scenes, such as bridges and roads. Research on remote sensing identification and extraction algorithms for ship locks is limited. Additionally, ship locks are often small "V" shapes, "W" shapes, or irregular rectangles, and the identification process requires high image resolution, making them difficult to separate from various other buildings in complex scenes, thus requiring improvement. Summary of the Invention

[0006] This application provides a method and apparatus for extracting locks based on high-resolution remote sensing images, in order to solve the technical problems in related technologies, such as the requirement for high image resolution when recognizing irregularly shaped objects, the difficulty in separating them from other building backgrounds in complex scenes, and the uncertainty in remote sensing image classification, which leads to low accuracy of recognition results.

[0007] The first aspect of this application provides a method for extracting ship locks based on high-resolution remote sensing imagery, comprising the following steps: extracting water body images from high-resolution remote sensing images of a target area; dividing the water body probability data of the target area using a preset threshold to obtain water body regions, and determining the target channel or river range within the target area based on the water body regions and the water body images; filtering out lock chamber regions that meet preset suspicion conditions from the target channel or river range based on prior knowledge of the lock chamber shape; and performing ship lock extraction within the lock chamber regions by combining the spectral characteristics of the ship lock and prior knowledge of the water body to obtain a ship lock extraction result that meets preset accuracy conditions.

[0008] Optionally, in one embodiment of this application, the step of extracting water body images from high-resolution remote sensing images of the target area includes: classifying the high-resolution remote sensing images to obtain binary classification images for each category; calculating the normalized water index of the high-resolution remote sensing images and performing binary segmentation on the normalized water index to obtain initial water body extraction images; calculating the difference between the binary classification images of each category and the initial water body extraction images to obtain calculation results, and obtaining the water body image based on the calculation results.

[0009] Optionally, in one embodiment of this application, the step of dividing the water body probability data of the target area using a preset threshold to obtain a water body area, and determining the target channel or river range in the target area based on the water body area and the water body image, includes: dividing the water body area into multiple connected domains and calculating the area of ​​each of the multiple connected domains; selecting connected domains with an area greater than a first preset area threshold as the initial channel or river range; and using the initial channel or river range as a mask to select the target channel or river range from the water body image.

[0010] Optionally, in one embodiment of this application, the step of selecting lock chamber areas that meet preset suspected conditions from the target channel or river range based on prior knowledge of the shape of the lock chamber includes: determining an empirical threshold based on the prior knowledge of the shape; using the empirical threshold to perform shape filtering on the target channel or river range, and selecting river segments with an area smaller than a second preset area threshold as the lock chamber areas that meet the preset suspected conditions.

[0011] Optionally, in one embodiment of this application, the step of combining the spectral characteristics of the lock and prior knowledge of the water body to extract the lock in the lock chamber area and obtain a lock extraction result that meets the preset accuracy conditions includes: obtaining the boundary points at both ends of each lock chamber in the lock chamber area that are approximately perpendicular to the river direction; extending along the river direction with each boundary point as the center to obtain the intersection point with the adjacent lock chamber; obtaining the region of interest where the lock is suspected based on all the intersection points on the boundary lines of the adjacent lock chambers that are approximately perpendicular to the river direction; and using the initial channel or river range and the target channel or river range as masks to filter out at least one interfering factor from the water body and non-river range from the region of interest to obtain the lock extraction result.

[0012] A second aspect of this application provides a lock extraction device based on high-resolution remote sensing imagery, comprising: a first extraction module for extracting water body images from high-resolution remote sensing imagery of a target area; a determination module for dividing the water body probability data of the target area using a preset threshold to obtain a water body region, and determining the target channel or river range within the target area based on the water body region and the water body image; a filtering module for filtering lock chamber regions that meet preset suspicion conditions from the target channel or river range based on prior knowledge of the lock chamber shape; and a second extraction module for performing lock extraction within the lock chamber regions by combining the spectral characteristics of the lock and prior knowledge of the water body, to obtain a lock extraction result that meets preset accuracy conditions.

[0013] Optionally, in one embodiment of this application, the first extraction module includes: a classification unit, used to classify the high-resolution remote sensing image to obtain a binary classification image for each category; a first calculation unit, used to calculate the normalized water index of the high-resolution remote sensing image and perform binary segmentation on the normalized water index to obtain an initial water body extraction image; and a second calculation unit, used to calculate the difference between the binary classification images of each category and the initial water body extraction image to obtain a calculation result, and obtain the water body image based on the calculation result.

[0014] Optionally, in one embodiment of this application, the determining module includes: a third calculation unit, configured to divide the water body region into multiple connected domains and calculate the area of ​​each of the multiple connected domains; a first filtering unit, configured to filter out connected domains with an area greater than a first preset area threshold as the initial channel or river range; and a second filtering unit, configured to use the initial channel or river range as a mask to filter out the target channel or river range from the water body image.

[0015] Optionally, in one embodiment of this application, the screening module includes: a determining unit, configured to determine an empirical threshold based on the prior knowledge of shape; and a third screening unit, configured to use the empirical threshold to perform shape filtering on the target channel or river range, and screen out river sections with an area smaller than a second preset area threshold as the gate chamber area that meets the preset suspected condition.

[0016] Optionally, in one embodiment of this application, the second extraction module includes: an acquisition unit, configured to acquire the boundary points at both ends of each lock chamber in the lock chamber region that are approximately perpendicular to the river direction, and extend along the river direction with each boundary point as the center to obtain the intersection point with the adjacent lock chamber; a fourth calculation unit, configured to obtain the region of interest where the suspected lock exists based on all the intersection points on the boundary lines of the adjacent lock chambers that are approximately perpendicular to the river direction; and a filtering unit, configured to use the initial channel or river range and the target channel or river range as masks to filter out at least one interfering factor from the region of interest that is not within the water body or the non-river range, respectively, to obtain the lock extraction result.

[0017] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the lock extraction method based on high-resolution remote sensing imagery as described in the above embodiments.

[0018] A fourth aspect of this application provides a computer-readable storage medium storing computer instructions for causing the computer to perform the lock extraction method based on high-resolution remote sensing imagery as described in the above embodiments.

[0019] A fifth aspect of this application provides a computer program product, including a computer program that, when executed, implements the above-described method for extracting locks based on high-resolution remote sensing images.

[0020] This application embodiment can extract water body images from high-resolution remote sensing images of the target area, divide the water body probability data of the target area using a preset threshold to obtain the water body region, and determine the target channel or river range in the target area based on the water body region and water body image. Combining prior knowledge of the shape of the lock chamber, lock chamber regions that meet preset suspicion conditions are selected. By combining the spectral characteristics of the lock and prior knowledge of the water body, lock extraction is performed in the lock chamber regions to obtain lock extraction results that meet preset accuracy conditions. Using satellite remote sensing for lock extraction can greatly reduce the workload and cost of manual operations, facilitate the daily dynamic maintenance of lock hubs, and provide more comprehensive data support for the construction and management of water transport infrastructure. Based on this, by combining fuzzy theory and prior knowledge of locks, lock targets can be extracted from high-resolution remote sensing images, thereby effectively extracting small lock targets from the images. This solves the technical problems in related technologies, such as the high image resolution requirements for recognizing irregularly shaped objects, the difficulty in separating them from various building backgrounds in complex scenes, and the uncertainty in remote sensing image classification, resulting in low accuracy of recognition results.

[0021] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0022] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0023] Figure 1 This is a flowchart of a lock extraction method based on high-resolution remote sensing imagery provided in an embodiment of this application;

[0024] Figure 2 This is a schematic diagram illustrating the principle of a lock extraction method based on high-resolution remote sensing imagery according to an embodiment of this application.

[0025] Figure 3 This is a schematic diagram illustrating the effect of water extraction according to an embodiment of this application;

[0026] Figure 4 This is a schematic diagram illustrating the effect of extracting the target channel or river range according to one embodiment of this application;

[0027] Figure 5 This is a schematic diagram illustrating the preliminary extraction effect of the gate chamber region according to an embodiment of this application;

[0028] Figure 6 This is a schematic diagram illustrating the principle of region of interest extraction according to an embodiment of this application;

[0029] Figure 7This is a schematic diagram illustrating the final effect of lock extraction according to an embodiment of this application;

[0030] Figure 8 This is a schematic diagram illustrating the effect of randomly generating verification points based on the lock extraction results according to one embodiment of this application;

[0031] Figure 9 This is a schematic diagram of a lock extraction device based on high-resolution remote sensing imagery provided in an embodiment of this application;

[0032] Figure 10 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0033] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0034] The following description, with reference to the accompanying drawings, describes a method and apparatus for extracting locks based on high-resolution remote sensing imagery according to embodiments of this application. Addressing the technical problems mentioned in the background art, such as the high image resolution requirements for recognizing irregularly shaped objects, the difficulty in separating them from various building backgrounds in complex scenes, and the uncertainty in remote sensing image classification leading to low accuracy of recognition results, this application provides a method for extracting locks based on high-resolution remote sensing imagery. In this method, water body images can be extracted from high-resolution remote sensing images of the target area. Water body probability data of the target area is divided using a preset threshold to obtain water body regions. Based on the water body regions and water body images, the target channel or river within the target area is determined. By considering the flow range and prior knowledge of the lock chamber shape, lock chamber areas that meet preset suspicion criteria are selected. Combining the spectral characteristics of the lock and prior knowledge of the water body, lock extraction is performed within these areas, yielding lock extraction results that meet preset accuracy requirements. Utilizing satellite remote sensing for lock extraction significantly reduces the workload and cost of manual operations, aids in the daily dynamic maintenance of lock hubs, and provides more comprehensive data support for the construction and management of water transport infrastructure. Furthermore, by combining fuzzy theory and prior knowledge of locks, lock targets can be extracted from high-resolution remote sensing images, effectively extracting small lock targets from the images. This solves the technical problems in related technologies, such as the high image resolution requirements for irregularly shaped object recognition, the difficulty in separating objects from various building backgrounds in complex scenes, and the uncertainty in remote sensing image classification, leading to low accuracy in recognition results.

[0035] Specifically, Figure 1This is a flowchart illustrating a method for extracting ship locks based on high-resolution remote sensing images, provided in an embodiment of this application.

[0036] like Figure 1 As shown, the lock extraction method based on high-resolution remote sensing imagery includes the following steps:

[0037] In step S101, water body images are extracted from high-resolution remote sensing images of the target area.

[0038] In actual implementation, the embodiments of this application can use remote sensing images to identify and extract locks, and on this basis, use an improved fuzzy classification algorithm to extract water images from high-resolution remote sensing images.

[0039] Among them, the EnFCM (Enhanced Fuzzy C-Means) algorithm is an efficient unsupervised classification method that can automatically segment images given the desired category. However, since the EnFCM algorithm can only classify images and cannot explicitly assign a category to the result, this application's embodiment calculates the NDWI of a high-resolution image and performs binary segmentation to obtain a coarse water body extraction result. Then, it calculates the similarity between each EnFCM classification category and the coarse water body extraction result to determine the water body category in the fuzzy classification result.

[0040] Optionally, in one embodiment of this application, extracting water body images from high-resolution remote sensing images of the target area includes: classifying the high-resolution remote sensing images to obtain binary classification images for each category; calculating the normalized water index of the high-resolution remote sensing images and performing binary segmentation on the normalized water index to obtain initial water body extraction images; calculating the difference between the binary classification images of each category and the initial water body extraction images to obtain calculation results, and obtaining water body images based on the calculation results.

[0041] As one possible implementation, the present application embodiment may include the following steps when extracting water body images:

[0042] Step S1: In this embodiment of the application, high-resolution remote sensing images can be classified using the Enhanced Fuzzy C-Means (EnFCM) algorithm to obtain a binary classification image for each category.

[0043] This application embodiment can preprocess the original image by linearly weighting and summing the pixels and their neighboring pixels to form an average image from the original image to its local neighboring pixels:

[0044]

[0045] Where, ξ k Let x represent the grayscale value of the k-th pixel in image ξ. j x represents kThe neighboring pixels, N k Indicates x k The set of pixels within the center window. x represents k The average gray value of the surrounding neighborhood, where α represents a parameter used to control the influence of the neighborhood.

[0046] In this embodiment, the objective function can be defined as J. s For any l, when satisfying Under the constraints, find J s Minimum value:

[0047]

[0048] Among them, v l Let u represent the i-th cluster center. il γ represents the fuzzy membership degree of gray value l relative to cluster i, c represents the number of clusters, q represents the number of gray levels in the given image, and γ l This represents the number of pixels with a grayscale value equal to l, where l = 1, ..., q.

[0049] Step S2: In this embodiment of the application, the Normalized Difference Water Index (NDWI) of the high-resolution image can be calculated, and binary segmentation can be performed on it to obtain a coarse water body extraction image, i.e., the initial water body extraction image. The formula for calculating the Normalized Difference Water Index (NDWI) is as follows:

[0050]

[0051] In this context, GREEN represents the green light band of the image, and NIR represents the near-infrared band of the image.

[0052] Step S3: In this embodiment of the application, the difference between the binary images of each type of EnFCM and the coarse water body extraction image is calculated. The image with the lowest difference is determined to be the accurate water body extraction image, i.e., the water body image. The mean absolute error (MAE) is used as the difference measure, and its calculation formula is as follows:

[0053]

[0054] Where m represents the number of pixels, x i This represents the actual observed value of the i-th pixel. This represents the predicted value of the i-th pixel.

[0055] In step S102, the water body probability data of the target area is divided using a preset threshold to obtain the water body area, and the target channel or river range in the target area is determined based on the water body area and the water body image.

[0056] Understandably, to extract a lock, the channel or river where the lock is located must first be located. The water extraction results in step S101 contain many non-river water bodies of varying sizes and shapes, making them difficult to filter using simple shape or area filtering methods. Dynamic World (DW) data has a low resolution, therefore, water-crossing structures such as locks and bridges in the image have little impact on river connectivity. Therefore, this embodiment of the application can use the DW probabilistic water body dataset to perform shape filtering on the preliminary EnFCM water body results, retaining only rivers with good connectivity and removing other small-area water bodies or those unrelated to rivers, thereby improving data quality and analytical accuracy, and facilitating subsequent lock extraction.

[0057] It should be noted that the preset threshold can be set by those skilled in the art based on their own relevant knowledge, and no specific restrictions are imposed here.

[0058] Optionally, in one embodiment of this application, water body probability data of the target area is divided using a preset threshold to obtain a water body area, and the target channel or river range in the target area is determined based on the water body area and the water body image, including: dividing the water body area into multiple connected domains and calculating the area of ​​each connected domain; selecting connected domains with an area greater than a first preset area threshold as the initial channel or river range; and using the initial channel or river range as a mask to select the target channel or river range from the water body image.

[0059] For example, the extraction of the target channel or river range may include the following steps:

[0060] Step S1: Based on the knowledge of those skilled in the art, set relevant thresholds to divide the DW data into water body and non-water body regions.

[0061] Step S2: Apply the 8-direction neighborhood search algorithm to divide the DW water body region into different connected domains, calculate the area of ​​each connected domain and filter by area, retaining the area with the larger area as the rough channel or river range, i.e. the initial channel or river range.

[0062] Step S3: Using the rough channel or river range as a mask, filter out the precise river range, i.e. the target channel or river range, from the precise water body image obtained in step S101.

[0063] In step S103, based on prior knowledge of the shape of the lock chamber, lock chamber areas that meet the preset suspected conditions are selected from the target channel or river range.

[0064] Furthermore, in this embodiment of the application, the suspected lock chamber area can be obtained by screening based on the general rules of lock chambers, such as the prior knowledge that the lock chamber area is approximately rectangular and the area of ​​the lock chamber is much smaller than the area of ​​the river section between other water-crossing structures.

[0065] The preset conditions can be set by those skilled in the art based on prior knowledge of the shape of the gate chamber, and no specific restrictions are imposed here.

[0066] Optionally, in one embodiment of this application, based on prior knowledge of the shape of the lock chamber, a lock chamber area that meets the preset suspected conditions is selected from the target channel or river range, including: determining an empirical threshold based on prior knowledge of the shape; using the empirical threshold to perform shape filtering on the target channel or river range, and selecting river sections with an area smaller than a second preset area threshold as lock chamber areas that meet the preset suspected conditions.

[0067] For example, in this application embodiment, information such as the lock chamber of the ship lock is usually approximately rectangular and the area of ​​the lock chamber is usually much smaller than the area of ​​the river section between other water-crossing structures can be used as shape prior knowledge, so as to obtain the suspected lock chamber area by filtering by area and shape.

[0068] Based on prior knowledge of shape, embodiments of this application can set an empirical threshold to perform shape filtering on the target channel or river range, and screen out river sections with smaller areas as suspected gate chamber areas, that is, gate chamber areas that meet the preset suspected conditions.

[0069] In step S104, the lock is extracted in the lock chamber area by combining the spectral characteristics of the lock and prior knowledge of the water body, and the lock extraction result that meets the preset accuracy conditions is obtained.

[0070] In actual implementation, the embodiments of this application can accurately extract the lock based on prior knowledge of the water body, such as the obvious difference between the spectral characteristics of the lock and the water body, and the fact that locks usually divide rivers into several sections.

[0071] It should be noted that the preset accuracy conditions can be set by those skilled in the art according to the actual situation, and no specific restrictions are imposed here.

[0072] Optionally, in one embodiment of this application, the lock is extracted in the lock chamber region by combining the spectral characteristics of the lock and prior knowledge of the water body to obtain the lock extraction result that meets the preset accuracy conditions. This includes: obtaining the boundary points at both ends of each lock chamber in the lock chamber region that are approximately perpendicular to the river direction; extending along the river direction with each boundary point as the center to obtain the intersection point with the adjacent lock chamber; obtaining the region of interest where the lock is suspected based on all the intersection points on the boundary lines of the adjacent lock chambers that are approximately perpendicular to the river direction; and using the initial channel or river range and the target channel or river range as masks to filter out at least one interfering factor from the water body and non-river range from the region of interest to obtain the lock extraction result.

[0073] As one possible implementation, embodiments of this application can obtain the boundary points at both ends of each lock chamber approximately perpendicular to the river direction. Using each boundary point as the center, a survey line is drawn along the river direction to obtain the intersection points with adjacent lock chambers. By drawing the minimum bounding rectangle of all intersection points on the boundary lines of adjacent lock chambers approximately perpendicular to the river direction, the region of interest (RoI) where the suspected lock exists can be obtained.

[0074] Furthermore, in this embodiment of the application, the initial channel or river range and the target channel or river range obtained in step S102 can be used as masks to filter out interference factors (such as land parts between rivers) in the water body and non-river range in the RoI, respectively, to obtain the final lock extraction result.

[0075] Combination Figures 2 to 8 As shown, an embodiment is used to illustrate in detail the working principle of the lock extraction method based on high-resolution remote sensing imagery of this application.

[0076] like Figure 2 As shown, the embodiments of this application can be divided into a river extraction stage and a lock extraction stage. Specifically, the embodiments of this application may include the following steps:

[0077] Step S1: Extract water bodies from high-resolution remote sensing images using an improved fuzzy classification algorithm. The extraction results can be as follows: Figure 3 As shown.

[0078] 1) High-resolution remote sensing images are classified using the Enhanced Fuzzy C-Means (EnFCM) algorithm, and binary classification images for each category are obtained.

[0079] 1.1) By preprocessing the original image through a linear weighted summation of pixels and their neighboring pixels, an average image of the original image and its local neighboring pixels is formed:

[0080]

[0081] Where, ξ k Let x represent the grayscale value of the k-th pixel in image ξ. j x represents k The neighboring pixels, N k Indicates x k The set of pixels within the center window. x represents k The average gray value of the surrounding neighborhood, where α represents a parameter used to control the influence of the neighborhood.

[0082] 1.2) Define the objective function as J s For any l, when satisfying Under the constraints, find J s Minimum value:

[0083]

[0084] Among them, v l Let u represent the i-th cluster center. il γ represents the fuzzy membership degree of gray value l relative to cluster i, c represents the number of clusters, q represents the number of gray levels in the given image, and γ l This represents the number of pixels with a grayscale value equal to l, where l = 1, ..., q.

[0085] 2) Calculate the Normalized Difference Water Index (NDWI) of the high-resolution image and perform binary segmentation to obtain a coarse water body extraction image, i.e., the initial water body extraction image. The formula for calculating the Normalized Difference Water Index (NDWI) is as follows:

[0086]

[0087] In this context, GREEN represents the green light band of the image, and NIR represents the near-infrared band of the image.

[0088] 3) The difference between the binary images of each EnFCM category and the coarse water extraction image is calculated. The image with the lowest difference is determined as the accurate water extraction image, i.e., the water image. The mean absolute error (MAE) is used as the difference measure, and its calculation formula is as follows:

[0089]

[0090] Where m represents the number of pixels, x i This represents the actual observed value of the i-th pixel. This represents the predicted value of the i-th pixel.

[0091] Step S2: Extract the target channel or river extent using DW (Dynamic World) water body probability data. The extraction results for the target channel or river extent in each experimental area can be as follows: Figure 4 As shown.

[0092] 1) Based on the knowledge of those skilled in the art, relevant thresholds are set to divide the DW data into water body and non-water body regions.

[0093] 2) The 8-direction neighborhood search algorithm is applied to divide the DW water body area into different connected domains. The area of ​​each connected domain is calculated and filtered by area. The area with the larger area is retained as the rough channel or river range, i.e. the initial channel or river range.

[0094] 3) Using the rough channel or river range as a mask, the precise river range, i.e. the target channel or river range, is selected from the precise water body image obtained in step S1.

[0095] Step S3: Based on the prior knowledge that lock chambers are typically approximately rectangular and their area is usually much smaller than the river section area between other water-crossing structures, potential lock chamber areas are obtained through area and shape filtering. According to this prior knowledge, an empirical threshold is set to filter the target channel or river range by shape, selecting smaller river sections as potential lock chamber areas. The lock chamber extraction effect in the experimental area can be as follows: Figure 5 As shown

[0096] Step S4: Based on prior knowledge of the spectral characteristics of the lock and the water body, and the fact that locks typically divide rivers into several sections, the locks are accurately extracted.

[0097] This application embodiment can obtain the boundary points at both ends of each lock chamber approximately perpendicular to the river direction. Using each boundary point as the center, a survey line is drawn along the river direction to obtain the intersection points with adjacent lock chambers. By drawing the minimum bounding rectangle of all intersection points on the boundary lines of adjacent lock chambers approximately perpendicular to the river direction, the region of interest (RoI) where the suspected lock exists can be obtained.

[0098] Furthermore, in this embodiment, the initial channel or river range and the target channel or river range obtained in step S1 can be used as masks to filter out interference factors (such as land portions between rivers) in the RoI, respectively, from water bodies and non-river ranges, to obtain, as shown below. Figure 7 The final lock extraction result is shown.

[0099] Step S5: Randomly generate verification points from the lock extraction results of each study area, such as... Figure 8 As shown in the embodiment of this application, the verification point type in the lock extraction result can be used as the test value, and the actual land cover type of each verification point determined by the manual visual interpretation method can be used as the true value. The overall accuracy and Kappa coefficient are calculated through the confusion matrix to quantitatively evaluate the accuracy of the algorithm.

[0100] In summary, the embodiments of this application first use a fuzzy classification algorithm to perform unsupervised classification of the study area image, and accurately identify the water body category by calculating the similarity between each category and the water body range of the Normalized Difference Water Index (NDWI) threshold. Subsequently, by fully utilizing the prior knowledge that the lock chambers are approximately rectangular and have relatively small areas, suspected lock chamber areas are identified through shape filtering and area threshold screening. At the same time, based on prior knowledge such as locks usually appearing in pairs, significant differences between spectral features and water areas, and relatively regular shapes of lock chambers (channel sections between two locks), an intelligent algorithm is further designed based on the lock chamber extraction results to achieve accurate lock segmentation.

[0101] The lock extraction method based on high-resolution remote sensing imagery proposed in this application can extract water body images from high-resolution remote sensing images of the target area, divide the water body probability data of the target area using a preset threshold to obtain the water body region, and determine the target channel or river range in the target area based on the water body region and the water body image. Combined with prior knowledge of the lock chamber shape, lock chamber regions that meet preset suspicion conditions are screened out. Combining the spectral characteristics of the lock and prior knowledge of the water body, lock extraction is performed in the lock chamber regions to obtain lock extraction results that meet preset accuracy conditions. Lock extraction using satellite remote sensing can greatly reduce the workload and cost of manual operations, help the daily dynamic maintenance of lock hubs, and provide more comprehensive data support for the construction and management of water transport infrastructure. On this basis, by combining fuzzy theory and prior knowledge of locks, lock targets can be extracted from high-resolution remote sensing images, thereby effectively extracting small lock targets in the images. This solves the technical problems in related technologies, such as the high image resolution requirements when recognizing irregularly shaped objects, the difficulty in separating them from other building backgrounds in complex scenes, and the uncertainty in remote sensing image classification, which leads to low accuracy of recognition results.

[0102] Next, referring to the accompanying drawings, a lock extraction device based on high-resolution remote sensing images, according to an embodiment of this application, is described.

[0103] Figure 9 This is a block diagram of a lock extraction device based on high-resolution remote sensing imagery, according to an embodiment of this application.

[0104] like Figure 9 As shown, the lock extraction device 10 based on high-resolution remote sensing imagery includes: a first extraction module 100, a determination module 200, a filtering module 300, and a second extraction module 400.

[0105] Specifically, the first extraction module 100 is used to extract water body images from high-resolution remote sensing images of the target area.

[0106] The determination module 200 is used to divide the water body probability data of the target area using a preset threshold to obtain the water body area, and to determine the range of the target channel or river in the target area based on the water body area and the water body image.

[0107] The filtering module 300 is used to filter out lock chamber areas that meet preset suspected conditions from the target channel or river range based on prior knowledge of the shape of the lock chamber.

[0108] The second extraction module 400 is used to combine the spectral characteristics of the lock and prior knowledge of the water body to extract the lock in the lock chamber area and obtain the lock extraction result that meets the preset accuracy conditions.

[0109] Optionally, in one embodiment of this application, the first extraction module 100 includes: a classification unit, a first calculation unit, and a second calculation unit.

[0110] The classification unit is used to classify high-resolution remote sensing images to obtain binary classification images for each category.

[0111] The first calculation unit is used to calculate the normalized water index of the high-resolution remote sensing image and perform binary segmentation on the normalized water index to obtain the initial water body extraction image.

[0112] The second calculation unit is used to calculate the difference between the binary classification images of each category and the initial water body extraction image, obtain the calculation results, and obtain the water body image based on the calculation results.

[0113] Optionally, in one embodiment of this application, the determining module 200 includes: a third calculation unit, a first filtering unit, and a second filtering unit.

[0114] The third calculation unit is used to divide the water body area into multiple connected domains and calculate the area of ​​each connected domain.

[0115] The first filtering unit is used to filter out connected regions with an area greater than a first preset area threshold as the initial channel or river range.

[0116] The second filtering unit is used to filter out the target channel or river range from the water body image by using the initial channel or river range as a mask.

[0117] Optionally, in one embodiment of this application, the screening module 300 includes a determining unit and a third screening unit.

[0118] The determining unit is used to determine the empirical threshold based on prior knowledge of shape.

[0119] The third screening unit is used to perform shape filtering on the target channel or river range using an empirical threshold, and to select river sections with an area smaller than the second preset area threshold as gate chamber areas that meet the preset suspected conditions.

[0120] Optionally, in one embodiment of this application, the second extraction module 400 includes: an acquisition unit, a fourth calculation unit, and a filtering unit.

[0121] The acquisition unit is used to acquire the boundary points at both ends of each gate chamber in the gate chamber area that are approximately perpendicular to the river direction, and to obtain the intersection point with the adjacent gate chamber by extending along the river direction with each boundary point as the center.

[0122] The fourth calculation unit is used to obtain the region of interest where the suspected lock exists based on all intersections on the boundary line of the adjacent lock chambers that are approximately perpendicular to the river direction.

[0123] The filtering unit is used to filter out at least one interfering factor from the water body and non-river range from the region of interest, respectively, using the initial channel or river range and the target channel or river range as masks, to obtain the lock extraction result. It should be noted that the foregoing explanation of the lock extraction method embodiment based on high-resolution remote sensing imagery also applies to the lock extraction device based on high-resolution remote sensing imagery in this embodiment, and will not be repeated here.

[0124] The lock extraction device based on high-resolution remote sensing imagery proposed in this application can extract water body images from high-resolution remote sensing images of the target area, divide the water body probability data of the target area using a preset threshold to obtain the water body area, and determine the target channel or river range in the target area based on the water body area and the water body image. Combining the prior knowledge of the lock chamber shape, lock chamber areas that meet preset suspicion conditions are screened out. Combining the spectral characteristics of the lock and the prior knowledge of the water body, lock extraction is performed in the lock chamber areas to obtain lock extraction results that meet preset accuracy conditions. Lock extraction using satellite remote sensing can greatly reduce the workload and cost of manual operations, help the daily dynamic maintenance of lock hubs, and provide more comprehensive data support for the construction and management of water transport infrastructure. On this basis, by combining fuzzy theory and prior knowledge of locks, lock targets can be extracted from high-resolution remote sensing images, thereby effectively extracting small lock targets in the images. This solves the technical problems in related technologies, such as the high image resolution requirements when recognizing irregularly shaped objects, the difficulty in separating them from other building backgrounds in complex scenes, and the uncertainty in remote sensing image classification, which leads to low accuracy of recognition results.

[0125] Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0126] The memory 1001, the processor 1002, and the computer program stored on the memory 1001 and capable of running on the processor 1002.

[0127] When the processor 1002 executes the program, it implements the lock extraction method based on high-resolution remote sensing images provided in the above embodiments.

[0128] Furthermore, electronic devices also include:

[0129] Communication interface 1003 is used for communication between memory 1001 and processor 1002.

[0130] The memory 1001 is used to store computer programs that can run on the processor 1002.

[0131] The memory 1001 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0132] If the memory 1001, processor 1002, and communication interface 1003 are implemented independently, then the communication interface 1003, memory 1001, and processor 1002 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized into address buses, data buses, control buses, etc. For ease of representation, Figure 10 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0133] Optionally, in a specific implementation, if the memory 1001, processor 1002, and communication interface 1003 are integrated on a single chip, then the memory 1001, processor 1002, and communication interface 1003 can communicate with each other through an internal interface.

[0134] The processor 1002 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0135] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for extracting locks based on high-resolution remote sensing images.

[0136] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the lock extraction method based on high-resolution remote sensing imagery provided in this embodiment of the invention.

[0137] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0138] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0139] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0140] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0141] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0142] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0143] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0144] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for extracting ship locks based on high-resolution remote sensing imagery, characterized in that, Includes the following steps: Extract water body images from high-resolution remote sensing images of the target area; The water body probability data of the target area is divided using a preset threshold to obtain the water body area, and the range of the target channel or river in the target area is determined based on the water body area and the water body image. Based on prior knowledge of the shape of the lock chamber, lock chamber areas that meet preset suspected conditions are selected from the target channel or river range; By combining the spectral characteristics of the lock with prior knowledge of the water body, lock extraction is performed in the lock chamber area to obtain lock extraction results that meet the preset accuracy conditions.

2. The method according to claim 1, characterized in that, The extraction of water body images from high-resolution remote sensing images of the target area includes: The high-resolution remote sensing images are classified to obtain binary classification images for each category; The normalized water index of the high-resolution remote sensing image is calculated, and the normalized water index is subjected to binary segmentation to obtain the initial water body extraction image. The difference between the binary classification images of each category and the initial water body extraction image is calculated to obtain the calculation results, and the water body image is obtained based on the calculation results.

3. The method according to claim 1, characterized in that, The step of dividing the water body probability data of the target area using a preset threshold to obtain the water body area, and determining the range of the target channel or river in the target area based on the water body area and the water body image, includes: The water body region is divided into multiple connected domains, and the area of ​​each of the multiple connected domains is calculated; Select connected components with an area greater than a first preset area threshold as the initial channel or river range; Using the initial channel or river range as a mask, the target channel or river range is selected from the water body image.

4. The method according to claim 1, characterized in that, The method of selecting lock chamber areas that meet preset potential conditions from the target channel or river range based on prior knowledge of the lock chamber shape includes: Determine the empirical threshold based on the prior knowledge of the shape; The target channel or river range is shape filtered using the empirical threshold, and river sections with an area smaller than a second preset area threshold are selected as the gate chamber areas that meet the preset suspected conditions.

5. The method according to claim 3, characterized in that, The method combines the spectral characteristics of the lock with prior knowledge of the water body to extract the lock within the lock chamber area, obtaining lock extraction results that meet preset accuracy conditions, including: Obtain the boundary points at both ends of each lock chamber in the lock chamber area that are approximately perpendicular to the river direction, and extend along the river direction with each boundary point as the center to obtain the intersection point with the adjacent lock chamber; The region of interest where the suspected lock exists is obtained by identifying all intersections on the boundary line of adjacent lock chambers that are approximately perpendicular to the river direction. Using the initial channel or river range and the target channel or river range as masks, at least one interfering factor in the water body and non-river range is filtered out from the region of interest to obtain the lock extraction result.

6. A lock extraction device based on high-resolution remote sensing imagery, characterized in that, include: The first extraction module is used to extract water body images from high-resolution remote sensing images of the target area; The determination module is used to divide the water body probability data of the target area using a preset threshold to obtain the water body area, and to determine the range of the target channel or river in the target area based on the water body area and the water body image; The filtering module is used to filter out lock chamber areas that meet preset suspected conditions from the target channel or river range based on prior knowledge of the shape of the lock chamber. The second extraction module is used to combine the spectral characteristics of the lock and prior knowledge of the water body to extract the lock in the lock chamber area and obtain the lock extraction result that meets the preset accuracy conditions.

7. The apparatus according to claim 6, characterized in that, The first extraction module includes: A classification unit is used to classify the high-resolution remote sensing image to obtain a binary classification image for each category; The first calculation unit is used to calculate the normalized water index of the high-resolution remote sensing image and perform binary segmentation on the normalized water index to obtain an initial water body extraction image. The second calculation unit is used to calculate the difference between the binary classification images of each category and the initial water body extraction image, obtain the calculation results, and obtain the water body image based on the calculation results.

8. The apparatus according to claim 6, characterized in that, The determining module includes: The third calculation unit is used to divide the water body region into multiple connected domains and calculate the area of ​​each of the multiple connected domains; The first filtering unit is used to filter out connected regions with an area greater than a first preset area threshold as the initial channel or river range. The second filtering unit is used to filter out the target channel or river range from the water image by using the initial channel or river range as a mask.

9. An electronic device, characterized in that, include: The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the lock extraction method based on high-resolution remote sensing imagery as described in any one of claims 1-5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the lock extraction method based on high-resolution remote sensing imagery as described in any one of claims 1-5.

Images