Shoreline wharf utilization state monitoring method and device

Abstract

The invention discloses a shoreline wharf utilization state monitoring method and device, and belongs to the field of artificial intelligence of shoreline resource supervision, and the method comprises the steps: obtaining comprehensive geographic space data of a target region, extracting wharf data from the comprehensive geographic space data through a deep learning prediction model, and carrying out the multi-scale detection, the method comprises the steps of obtaining wharf position information, obtaining ship AIS track data of a target area, analyzing a staying behavior of a ship motion track according to the ship AIS track data, extracting ship staying points, establishing a buffer area based on the wharf position information, and counting ship staying parameters corresponding to the ship staying points falling into the buffer area; and performing statistical analysis and comprehensive duration comparison on the ship parking parameters to obtain a shoreline wharf utilization state of the target area. The use condition of the wharf is dynamically monitored according to the real-time comprehensive geographic space data, powerful support is provided for wharf planning and management, and the accuracy and efficiency of wharf state supervision are improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence for shoreline resource supervision, and more specifically, to a method and device for monitoring the utilization status of shoreline docks. Background Art

[0002] As an important part of the ecosystem, the shoreline can improve the intensive and economical utilization degree of the shoreline, which is of great significance for the spatial control of the shoreline space of rivers and lakes and the restoration of the ecological environment of rivers and lakes. The overall planning and layout, as well as the reasonable development and utilization of shoreline docks, play an important role in promoting the sustainable development of shoreline resources.

[0003] In the current existing technologies, there are mainly two types of methods to manage shoreline docks. On the one hand, high-resolution satellite remote sensing images are used to identify ports. On the other hand, radar images are used, and the hierarchical change detection method is used to achieve the detection of ports and targets within the port.

[0004] Although these methods can accurately identify the spatial pattern and infrastructure of ports to a certain extent. However, since the river and lake shorelines usually mainly consist of floating island docks, not only do they show a smaller area in remote sensing images, but they are also easily misclassified with ships traveling on rivers and lakes. More importantly, due to the complexity and dynamics of river and lake shorelines, as well as the diversity of the usage conditions of docks or ports, it is difficult to dynamically monitor the utilization status of docks (such as the frequency of ships docking at the dock) only based on remote sensing image data. Summary of the Invention

[0005] Embodiments of the present invention provide a method and device for monitoring the utilization status of shoreline docks to solve the technical defects of insufficient supervision of shoreline resources and lack of effective information in the prior art.

[0006] In a first aspect, a method for monitoring the utilization status of shoreline docks disclosed in this application includes:

[0007] Obtain the comprehensive geospatial data of the target area, and use a deep learning prediction model to extract dock data from the comprehensive geospatial data for multi-scale detection to obtain dock location information;

[0008] Obtain the AIS trajectory data of ships in the target area, and analyze the stay behavior of the ship movement trajectory according to the AIS trajectory data of the ships to extract ship stay points;

[0009] Establish a buffer zone based on the dock location information, and count the ship mooring parameters corresponding to the ship stay points falling into the buffer zone;

[0010] Perform statistical analysis and comprehensive duration comparison on the ship mooring parameters to obtain the utilization status of the shoreline docks in the target area.

[0011] Further, the method for extracting dock data from the comprehensive geospatial data by using the deep learning prediction model for multi-scale detection to obtain dock location information includes:

[0012] Performing dock interpretation based on the comprehensive geospatial data to obtain an interpretation result and using the interpretation result to train a deep learning model;

[0013] Inputting the comprehensive geospatial data into the trained deep learning model to obtain dock data and performing multi-scale detection to obtain dock location information.

[0014] Further, the method for performing dock interpretation based on the comprehensive geospatial data to obtain an interpretation result and using the interpretation result to train a deep learning model includes:

[0015] Based on dock label identifications, screening out target vector data corresponding to docks from the comprehensive geospatial data;

[0016] Taking the center point of the target vector data as the geometric center to establish a first regular grid with a preset scale size of M×M;

[0017] For each first regular grid, obtaining a satellite remote sensing image covering the grid and performing dock interpretation and recognition based on the satellite remote sensing image;

[0018] Within the dock area identified by interpretation, determining the minimum bounding rectangle of the dock and constructing a training label for the dock based on the position information of the minimum bounding rectangle;

[0019] Constructing a training sample based on the dock training label and the corresponding satellite remote sensing image data, and training a deep learning model based on the training sample.

[0020] Further, the method for inputting the comprehensive geospatial data into the trained deep learning model to obtain dock data and performing multi-scale detection to obtain dock location information includes:

[0021] Establishing multiple first regular grids with a preset scale size of M×M within the target area, using the trained deep learning model to identify and extract dock data within each first regular grid to obtain a preliminary extracted dock set;

[0022] Taking each first regular grid as the center, establishing a second regular grid with a preset scale size of 3M×3M, and using the trained deep learning model to identify and extract dock data within the second regular grid to obtain an extracted dock set in the overlapping area between the second regular grid and the first regular grid;

[0023] Calculate the similarity between the initially extracted quay set and the quay set extracted from the overlapping area of the second regular grid and the first regular grid, and verify the result to determine the quay recognition result and obtain the quay location information.

[0024] Further, the calculating the similarity between the initially extracted quay set and the quay set extracted from the overlapping area of the second regular grid and the first regular grid, and verifying the result to determine the quay recognition result and obtain the quay location information includes:

[0025] Calculate the intersection area between the initially extracted quay set and the quay set extracted from the overlapping area of the two grids, and divide it by the area of the initially extracted quay set to obtain the quay extraction similarity.

[0026] When it is determined that the quay extraction similarity is greater than the preset similarity threshold, take the extraction result in the initially extracted quay set as the final quay recognition result; otherwise, discard the extraction result in the initially extracted quay set.

[0027] Further, the obtaining the AIS trajectory data of ships in the target area, analyzing the staying behavior of the ship movement trajectory according to the AIS trajectory data of the ships, and extracting the ship staying points includes:

[0028] Based on the AIS trajectory data of the ships, traverse the movement trajectories of each ship in the area.

[0029] During the traversal, use the time interval and position change between adjacent trajectory points to calculate the continuous movement speed of the ship and the number of trajectory points of the movement trajectory, and regard the trajectory segment with a continuous movement speed less than the preset speed threshold and the number of movement trajectory points greater than the preset point threshold as the ship staying behavior.

[0030] For the AIS trajectory segment of each ship staying behavior, take the centroid point of all trajectory (space) points in the trajectory segment as the ship staying point.

[0031] Further, the statistically analyzing and comprehensively comparing the ship mooring parameters to obtain the shoreline quay utilization status of the target area includes:

[0032] The quay utilization status includes the quay utilization frequency and the quay utilization rate.

[0033] Statistically calculate the total number of staying points of the quay in the unit time T in the ship mooring parameters, and take it as the quay utilization frequency.

[0034] Statistically calculate the total residence duration of the ship residence behaviors corresponding to all stop points in the dock buffer area within the unit time T, and the total business hours of the dock. Based on the comparison calculation of the total residence duration and the total business hours, the dock utilization rate is obtained.

[0035] In a second aspect, the present application also discloses a shoreline dock utilization status monitoring device, including:

[0036] A dock identification module, configured to obtain comprehensive geospatial data of a target area, and extract dock data from the comprehensive geospatial data by using a deep learning prediction model for multi-scale detection to obtain dock location information;

[0037] A ship stop point analysis module, configured to obtain ship AIS trajectory data of a target area, and analyze the stop behaviors of the ship movement trajectory according to the ship AIS trajectory data to extract ship stop points;

[0038] A dock buffer area analysis module, configured to establish a buffer area based on the dock location information, and statistically calculate the ship berthing parameters corresponding to the ship stop points falling into the buffer area;

[0039] A status statistics and monitoring module, configured to perform statistical analysis and comprehensive duration comparison on the ship berthing parameters to obtain the utilization status of the shoreline dock in the target area.

[0040] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of any one of the above-mentioned remote sensing change information extraction methods based on zero-shot learning are implemented.

[0041] In a fourth aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above-mentioned remote sensing change information extraction methods based on zero-shot learning are implemented.

[0042] Compared with the prior art, a shoreline dock utilization status monitoring method and device provided by the present application have the following technical effects:

[0043] 1) This application obtains the comprehensive geospatial data of the target area, uses a deep learning prediction model to extract dock data from the comprehensive geospatial data for multi-scale detection to obtain dock location information, then obtains the ship AIS trajectory data of the target area, analyzes the stay behavior of the ship movement trajectory based on the ship AIS trajectory data, extracts ship stay points, and then establishes a buffer zone based on the dock location information, and counts the ship berthing parameters corresponding to the ship stay points falling into the buffer zone, so as to conduct statistical analysis and comprehensive duration comparison of the ship berthing parameters, and obtain the utilization status of the shoreline dock in the target area.

[0044] By predicting the dock location through a deep learning algorithm and combining it with multi-scale detection, the dock location can be automatically identified according to the comprehensive geospatial data, and then a buffer zone is established based on the identified dock location. Furthermore, by counting the number and stay time of ship stay points falling into the buffer zone, the analysis of ship activities can be further refined. This method helps to reveal the spatial relationship and time characteristics between ships and docks, can dynamically monitor the usage of ports or docks (such as the frequency of ships docking at ports) according to real-time comprehensive geospatial data, provides strong support for the planning and management of docks, and greatly improves the accuracy and efficiency of dock status supervision.

[0045] 2) By dividing the target area into regular grids of different scales and combining with a deep learning algorithm for prediction, the situation where the bow or stern is misidentified as a dock can be further avoided, and the robustness of the identification can be improved;

[0046] 3) Using the obtained ship AIS trajectory data, by identifying the trajectory stay behavior of each ship, ship stay points are extracted. Compared with the traditional method of determining the ship stay area based on clustering algorithms (such as DBSCAN, etc.), the time complexity of the algorithm of the present invention is reduced from O(n2) to O(n), which can significantly improve the processing efficiency of AIS trajectory data. Brief Description of the Drawings

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0048] Figure 1 It is a method flow chart of a method for monitoring the utilization status of a shoreline dock provided by an embodiment of the present invention;

[0049] Figure 2 It is a schematic diagram of obtaining the range of remote sensing image data based on OSM dock vector data provided by an embodiment of the present invention;

[0050] Figure 3 It is a schematic diagram of multi-scale detection of docks provided by an embodiment of the present invention.

[0051] Figure 4 It is a schematic diagram of obtaining ship stay points based on AIS trajectory data provided by an embodiment of the present invention;

[0052] Figure 5 It is a schematic diagram of calculating the number of stay points within the dock buffer range provided by an embodiment of the present invention;

[0053] Figure 6 It is a schematic diagram of the calculation result of the utilization frequency of docks in a certain main stream area of the Yangtze River provided by an embodiment of the present invention;

[0054] Figure 7 It is a schematic diagram of the calculation result of the utilization rate of docks in a certain main stream area of the Yangtze River provided by an embodiment of the present invention;

[0055] Figure 8 It is a schematic diagram of the structure of a shoreline dock utilization status monitoring device provided by an embodiment of the present invention.

[0056] Figure 9 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0057] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0058] The terms "first", "second", "third", etc. in the specification, claims and drawings of this application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally further include steps or units not listed, or may optionally further include other steps or units inherent to these processes, methods, products or devices.

[0059] Next, example embodiments according to the present application will be described in detail with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the example embodiments described herein.

[0060] Application overview

[0061] As an important part of the ecosystem, the shoreline can improve the intensive and economical utilization degree of the shoreline, which is of great significance for the spatial control of the water area shoreline of rivers and lakes and the restoration of the ecological environment of rivers and lakes. The overall planning and layout, as well as the reasonable development and utilization of shoreline docks, play an important role in promoting the sustainable development of shoreline resources.

[0062] Currently, some scholars have identified ports based on high-resolution satellite remote sensing images. For example, Li et al. established an object-oriented remote sensing detection method for the spatial pattern of ports; Bovolo et al. used radar images and implemented the detection of ports and targets within the ports by using a hierarchical change detection method.

[0063] Although these methods can accurately identify the spatial pattern and infrastructure of ports to a certain extent. However, since the river and lake shorelines usually mainly consist of floating island docks, which not only show a smaller area on remote sensing images but are also prone to misclassification with the ships sailing on the rivers and lakes. More importantly, due to the complexity and dynamics of the river and lake shorelines, as well as the diversity of the usage conditions of docks or ports, it is difficult to dynamically monitor the utilization status of docks (such as the frequency of ships docking at the docks) only based on remote sensing image data.

[0064] Therefore, it is necessary to comprehensively utilize high-resolution remote sensing images, AIS trajectory data, and other auxiliary data sources and technical methods in view of the limitations of the existing technologies to achieve real-time and accurate monitoring of the utilization rate of shoreline docks.

[0065] In view of the above technical problems, the concept of this application is to propose a method for monitoring the utilization status of shoreline docks, a device for monitoring the utilization status of shoreline docks, and an electronic device. By using a deep learning algorithm to predict the dock location and combining it with multi-scale detection, after automatically identifying the dock location based on comprehensive geospatial data, a buffer zone is established based on the identified dock location. Furthermore, by counting the number of ship stop points and the stop time falling within the buffer zone, the analysis of ship activities can be further refined. This method helps to reveal the spatial relationship and time characteristics between ships and docks, dynamically monitor the usage conditions of ports or docks (such as the frequency of ships docking at the port) according to real-time comprehensive geospatial data, thereby providing strong support for the planning and management of docks and greatly improving the accuracy and efficiency of dock status supervision.

[0066] After introducing the basic principle of this application, the following will specifically introduce various non-limiting embodiments of this application with reference to the accompanying drawings.

[0067] Exemplary Method

[0068] Figure 1 The flowchart of the method for monitoring the utilization status of shoreline docks according to an embodiment of this application is illustrated

[0069] As shown Figure 1 in the figure, the shoreline wharf utilization status monitoring method according to the embodiment of the present application includes:

[0070] S1. Obtain the comprehensive geospatial data of the target area, extract wharf data from the comprehensive geospatial data by using a deep learning prediction model for multi-scale detection, and obtain wharf location information;

[0071] S2. Obtain the ship AIS trajectory data of the target area, analyze the stay behavior of the ship movement trajectory according to the ship AIS trajectory data, and extract ship stay points;

[0072] S3. Establish a buffer based on the wharf location information, and count the ship mooring parameters corresponding to the ship stay points falling into the buffer;

[0073] S4. Conduct statistical analysis and comprehensive duration comparison on the ship mooring parameters to obtain the utilization status of the shoreline wharf in the target area.

[0074] Among them, the comprehensive geospatial data is a data set that integrates multi-source information and is used to accurately describe the geographical features and spatial distribution status of the earth's surface, such as high-resolution satellite remote sensing image data and OpenStreetMap (OSM) data.

[0075] Next, each step will be described in detail.

[0076] In step S1, obtain the comprehensive geospatial data of the target area, extract wharf data from the comprehensive geospatial data by using a deep learning prediction model for multi-scale detection, and obtain wharf location information.

[0077] Specifically, first obtain the sub-meter high-resolution satellite remote sensing image data of the target area, and ensure that the resolution of the obtained remote sensing image is not lower than 1 meter.

[0078] In this embodiment, the OSM data of the target area is also obtained. Among them, the OSM data contains various geographical elements, such as streets, buildings, rivers, lakes, mountains, public facilities, etc. These data are all composed of three basic elements: nodes, paths, and relationships, and the attributes of geographical objects are described through tags, such as road types, building uses, etc.

[0079] Furthermore, use a deep learning prediction model to extract wharf data from the comprehensive geospatial data for multi-scale detection to obtain wharf location information.

[0080] Specifically, first screen out the target vector data representing the wharf from the OSM data through wharf label identification (such as the value of "value = pier").

[0081] After that, taking the center point of the target vector data as the center, a first regular grid with a size of M×M is established (please refer to Figure 2 ), and the high-resolution satellite remote sensing images covered within the M×M grid are interpreted manually by visual inspection, and the minimum bounding rectangle of the dock area is marked as the training label.

[0082] Furthermore, this application will also construct training samples for the deep learning model in combination with the training label and the corresponding satellite remote sensing image data, and use the trained deep learning model to perform dock recognition respectively within each first regular grid and within a second regular grid with a size of 3M×3M established with each first regular grid as the center (please refer to Figure 3 ), and obtain the corresponding results I and result II.

[0083] Furthermore, in order to improve the accuracy and reliability of dock recognition, this application will also extract the similarity between result I and result II, that is, the value obtained by dividing the overlapping area of the two extraction results by the extraction area of result I.

[0084] And when it is determined that the similarity is greater than the preset similarity threshold, the extraction result in result I is retained as the final dock recognition result; otherwise, the extraction result in result I is discarded.

[0085] In step S2, the ship AIS trajectory data of the target area is obtained, and the stay behavior of the ship movement trajectory is analyzed according to the ship AIS trajectory data, and the ship stay points are extracted.

[0086] Specifically, please refer to Figure 4 , this application will obtain the ship AIS trajectory data of the target area,

[0087] wherein, the ship AIS trajectory data records key information such as the movement path, timestamp, speed, and heading of the ship in the target area.

[0088] After that, the movement trajectories of each ship in the area are traversed in turn, and the trajectory segments with a continuous movement speed less than the preset speed threshold V and the number of trajectory points greater than the preset point threshold P are extracted as the ship stay behavior. Please refer to Figure 4 , the AIS trajectory segment of the ship stay behavior records the centroid point of the trajectory segment of the ship moving at a low speed continuously in the target area and the distribution of these trajectory segments.

[0089] In order to accurately identify the stay position and stay duration of the ship in the target area, this application will determine the centroid point of all trajectory points within the AIS trajectory segment of the ship stay behavior through geometric operations or statistical analysis methods, and use it as the ship stay point of the trajectory segment, and count the start time and end time of the trajectory segment.

[0090] In a specific implementation case, in order to better identify and analyze the staying behavior of ships in a specific area, the present invention extracts a trajectory segment with a continuous movement speed less than 1 byte and a trajectory point number greater than 5 as the ship staying behavior. For the AIS trajectory segment of each ship staying behavior, the centroid point of all trajectory (space) points within the trajectory segment is used as the ship staying point.

[0091] Step S3: Establish a buffer zone based on the dock location information, and count the ship mooring parameters corresponding to the ship staying points falling within the buffer zone.

[0092] The ship mooring parameters are specific parameters that reflect the state of ship staying, the number of staying points, and the staying time, etc. For example, the number of ship staying points and the corresponding staying time.

[0093] Specifically, please refer to Figure 5 , this application will establish a circular buffer zone with a radius of 100 m around the dock center point. By analyzing the centroid points of all trajectory points within the AIS trajectory segment of the ship staying behavior, the number of ship staying points and the staying time falling within the buffer zone are counted. Specifically, first, the centroid points of all trajectory points within the AIS trajectory segment of each ship staying behavior are determined, and then it is judged whether these centroid points are located within the buffer zone. Among them, for the centroid points located within the buffer zone, the corresponding number of ship staying points and the staying time will be further recorded, so as to obtain the statistical situation of ship activities within the buffer zone.

[0094] In step S4, statistical analysis and comprehensive duration comparison are performed on the ship mooring parameters to obtain the utilization status of the shoreline dock in the target area.

[0095] The utilization status of the shoreline dock refers to the usage situation and efficiency of the dock's occupied shoreline resources within a certain period, such as the dock utilization frequency, the dock utilization rate, and the dock capacity ratio, etc.

[0096] The dock utilization frequency refers to the number of times the dock is used by ships within a specific time period. It reflects the busyness and business activity of the dock. A higher utilization frequency usually means that there are more ships berthing and operating at the dock, which may be due to reasons such as the superior geographical location of the dock, large cargo throughput, or high service level. For example, a container dock located at a transportation hub may have dozens of ships berthing every day, and its utilization frequency is relatively high.

[0097] The dock utilization rate is usually expressed by the ratio of the actual throughput of the dock to the designed throughput. It reflects the degree to which the dock exerts its designed capacity in actual operation. The higher the dock utilization rate, the more fully the dock resources are utilized and the higher the operation efficiency. For example, if the designed annual throughput of a dock is 10 million tons and the actual annual throughput reaches 8 million tons, then its dock utilization rate is 80%.

[0098] The quay capacity ratio, which is the ratio of the actual annual throughput of the quay to the designed throughput capacity, reflects the actual utilization efficiency of the quay. Generally, when the actual throughput to the designed capacity is 1:1.2, the fitness is the best. When the capacity ratio is lower than 0.83, it belongs to the state of unsaturated throughput, and the quay efficiency and benefits are not fully exerted. When the capacity ratio is greater than 1, the quay throughput exceeds the designed throughput, which belongs to the full-load operation state, and it may be necessary to consider expansion or optimization and upgrading.

[0099] In an optional embodiment, in the method for monitoring the utilization state of a shoreline quay provided in this embodiment, the statistical analysis and comprehensive duration comparison of the ship berthing parameters to obtain the utilization state of the shoreline quay in the target area include:

[0100] The quay utilization state includes the quay utilization frequency and the quay utilization rate;

[0101] Count the total number of stop points of the quay in the unit time T in the ship berthing parameters and use it as the quay utilization frequency;

[0102] Count the total stay duration of the ship stay behaviors corresponding to all stop points in the quay buffer area in the unit time T in the ship berthing parameters and the total business hours of the quay, and perform a comparison calculation based on the total stay duration and the total business hours to obtain the quay utilization rate.

[0103] Specifically, first, according to the counted number of ship stop points, it can be calculated how many ships have stopped at the quay in a specific time period (such as one day, one week or one month), so as to obtain the quay utilization frequency. This frequency reflects the frequency of use of the quay. Second, in order to calculate the quay utilization rate, this application will add up the total stay time of all ships at the quay and then divide it by the total time that the quay can be open (i.e., the business hours, such as 24 hours a day or the total time of one week). The obtained proportional value is the quay utilization rate, which reflects the comparison relationship between the actual use situation and the potential use capacity of the quay in a specific time period.

[0104] As can be seen from the above, a method for monitoring the utilization status of a shoreline wharf disclosed in this application, through a method combining deep learning prediction and multi-scale detection, can further avoid the situation where the bow or stern of a ship is misidentified as a wharf, improving the robustness of identification; by using the obtained ship AIS trajectory data, by identifying the trajectory stay behavior of each ship, the ship stay points are extracted. Compared with the traditional method of determining the ship stay area based on a clustering algorithm (such as DBSCAN, etc.), the time complexity of the algorithm of the present invention is reduced from O(n2) to O(n), which can significantly improve the processing efficiency of AIS trajectory data; a buffer zone is established based on the identified wharf position, and the number and stay time of ship stay points falling into the buffer zone are counted, which can further refine the analysis of ship activities. This method helps to reveal the spatial relationship and time characteristics between ships and wharves, providing scientific support for the planning and management of wharves.

[0105] In one embodiment, in step S1, when using the deep learning prediction model to extract wharf data from the comprehensive geospatial data for multi-scale detection to obtain wharf position information, it includes:

[0106] S21, performing wharf interpretation based on the comprehensive geospatial data, obtaining an interpretation result and using the interpretation result to train a deep learning model;

[0107] Specifically, based on the comprehensive geospatial data, wharf interpretation is performed through an artificial visual interpretation method to obtain the position information of the wharf. Among them, the artificial visual interpretation method refers to using high-resolution satellite remote sensing image data and OSM data, combining professional knowledge and experience, and through visual observation and judgment by the naked eye, identifying and extracting the spatial information such as the position, shape, size, and direction of the wharf and its related facilities, as well as the characteristics of the surrounding environment of the wharf, such as water area, land, buildings, transportation network, etc.

[0108] In this embodiment, a deep learning object monitoring model based on Yolov5 is adopted. The training label of the minimum bounding rectangle corresponding to the marked wharf range and the corresponding high-resolution satellite remote sensing image data are used as training samples and input into the Yolov5 model for training. By using the powerful feature extraction ability and object detection accuracy of Yolov5, the position of the wharf in the image is accurately identified.

[0109] S22, inputting the comprehensive geospatial data into the trained deep learning model, obtaining wharf data and performing multi-scale detection to obtain wharf position information.

[0110] Specifically, in this application, the trained deep learning model is used to identify piers within regular grids of different sizes. Specifically, first, this application establishes an initial grid of a preset scale size of M×M, and uses the trained deep learning model to extract a preliminary set of piers from it; then, the scope of the initial grid is expanded to establish an extended grid of a preset scale size of 3M×3M (specifically, reference can be made to Figure 3 ), which covers and partially overlaps with the initial grid, and piers in the overlapping area are extracted to obtain the set of piers extracted from the overlapping area between the second regular grid and the first regular grid. Finally, the final pier recognition result is determined through similarity calculation and result verification.

[0111] In the above embodiment, based on the establishment of the regular initial grid of M×M, in order to further avoid the situation where the bow or stern is misidentified as a pier, an idea of multi-scale detection is proposed, that is, detecting targets in the same area at different scales (M×M and 3M×3M) to improve the detection accuracy.

[0112] In one of the embodiments, in step S21, the pier interpretation is performed based on the comprehensive geospatial data to obtain an interpretation result and use the interpretation result to train the deep learning model, including:

[0113] Step S211, based on the pier label identification, filter out the target vector data corresponding to the piers from the comprehensive geospatial data.

[0114] Specifically, this application filters out the target vector data representing piers from the OSM data based on the pier label identification corresponding to "value = pier". It should be noted that the pier information in the OSM data may exist in different label or attribute value forms, and "value = pier" is one of the common representation methods.

[0115] Step S212, with the center point of the target vector data as the geometric center, establish a first regular grid with a preset scale size of M×M.

[0116] Specifically, please refer to Figure 2 , this application will use the center point of the target vector data (i.e., the position marked as "pier center point" in the figure) as the benchmark, and establish a square first regular grid according to the size of M×M.

[0117] Step S213, for each first regular grid, obtain the satellite remote sensing image covering the grid, and perform pier interpretation and recognition based on the satellite remote sensing image.

[0118] Step S214, within the pier area identified by interpretation, determine the minimum bounding rectangle of the pier, and construct the training label of the pier based on the position information of the minimum bounding rectangle.

[0119] Specifically, the minimum bounding rectangle constructed in the current embodiment can tightly wrap the dock area while minimizing unnecessary blank areas as much as possible. This minimum bounding rectangle not only provides the precise position and range of the dock in the image but also serves as the basis for constructing the dock training label. Specifically, the training label constructed in this application will include the upper left coordinate and the lower right coordinate of this minimum bounding rectangle to adapt to Yolov5.

[0120] Step S215: Construct training samples based on the dock training label and the corresponding satellite remote sensing image data, and train a deep learning model based on the training samples.

[0121] Specifically, this application will construct training samples suitable for Yolov5 based on the dock training label and the corresponding satellite remote sensing image data. By combining the dock training label with the satellite remote sensing image data, training samples that contain both target position information and rich visual information can be constructed, thereby effectively training the Yolov5 model to enable it to accurately identify docks.

[0122] In one embodiment, in step S22, inputting the comprehensive geospatial data into the trained deep learning model to obtain dock data and performing multi-scale detection to obtain dock position information includes:

[0123] Step S221: Establish multiple first regular grids of a preset scale size M×M within the target area, and use the trained deep learning model to identify and extract the dock data within each first regular grid to obtain a preliminary extracted dock set.

[0124] Specifically, this application will divide multiple regular grids according to the preset scale size M×M, and these grids evenly cover the target area. Among them, each grid is regarded as an independent recognition unit, and the trained deep learning model is used to process the image data within each grid to identify and extract potential dock features. Finally, the identified dock objects are collected to form a preliminary extracted dock set. This set contains all objects initially judged as docks within the target area.

[0125] Step S222: Centering on each first regular grid, establish a second regular grid with a preset scale size of 3M×3M, and use the trained deep learning model to identify and extract the dock data within the second regular grid to obtain an extracted dock set in the overlapping area between the second regular grid and the first regular grid.

[0126] Specifically, in this application, each first regular grid (with a size of M×M) is used as the center, and a larger second regular grid is expanded outward, with a preset size of 3M×3M. This expansion process ensures that each first regular grid is completely contained within the corresponding second regular grid. Subsequently, the trained deep learning model is used to perform detailed identification and analysis on the image data within each second regular grid to extract the dock features contained therein.

[0127] It should be noted that since the size of the second regular grid is three times that of the first regular grid, there will be a large amount of duplicate identification in their overlapping areas, which helps to reduce the situation where the bow or stern is misidentified as a dock, thereby enhancing the robustness and accuracy of the identification.

[0128] Step S223: Calculate the similarity and verify the results based on the preliminary extracted dock set and the dock set extracted from the overlapping area between the second regular grid and the first regular grid to determine the dock identification result and obtain the dock position information.

[0129] Specifically, the above similarity calculation formula includes:

[0130]

[0131] Among them, St represents the similarity of the two detection results at the M×M scale and the 3M×3M scale, S 提取结果I represents the dock area determined based on the preliminary extracted dock set at the M×M scale, S 提起结果II represents the dock area determined based on the dock set extracted from the overlapping extended area at the 3M×3M scale, S 提取结果I∩提起结果II represents the overlapping area between the preliminary extracted dock set and the dock set extracted from the overlapping extended area.

[0132] In one embodiment, the preset scale size is 500 meters × 500 meters.

[0133] Specifically, taking a certain main stream area of the Yangtze River as the target area, the preset scale sizes are set to 300 meters × 300 meters, 500 meters × 500 meters, and 700 meters × 700 meters respectively, the similarity thresholds for the second detection are set to 0.3, 0.5, and 0.7 respectively, and the precision rate, recall rate, and F1 score are used to evaluate the dock detection effects of different test thresholds as shown in Tables 1 and 2:

[0134] Table 1 Evaluation results of single-scale (first time) detection

[0135] Grid size Precision (%) Recall (%) F1 score 300m 72.50％ 87.60％ 0.793 500m 83.50％ 92.80％ 0.879 700m 55.90％ 92.70％ 0.698

[0136] Table 2 Evaluation results of multi-scale (second time) detection

[0137]

[0138] As can be seen from Table 1 and Table 2, first of all, under the comparison of different grid sizes, the multi-scale detection is better than the single-scale detection in terms of evaluation results. Further, in the combination of different grid sizes and similarity thresholds, when the grid size M is set to 500 meters × 500 meters and the similarity S is set to 0.5, the comprehensive performance of dock detection is the most excellent. This combination not only maintains a high level in terms of precision, effectively reducing the misdetection situation, but also achieves good results in terms of recall rate, ensuring that as many docks as possible are correctly identified. More importantly, through the comprehensive consideration of the F1 score, the superiority of this combination in balancing precision and recall rate is further verified.

[0139] In one embodiment, in step S223, the similarity calculation and result verification are performed based on the initially extracted dock set and the extracted dock set in the overlapping area between the second regular grid and the first regular grid to determine the dock recognition result, including:

[0140] Step S2231, calculate the overlapping area between the initially extracted dock set and the extracted dock set in the overlapping area of the extended region, and divide the value by the area of the initially extracted dock set to obtain the dock extraction similarity.

[0141] Specifically, this application will determine the specific positions and shapes of each dock area in the initially extracted dock set and the extracted dock set in the overlapping area of the extended region. Then, through the spatial intersection operation on these two sets, the overlapping area of the dock areas in the two sets is determined.

[0142] Further, after obtaining the overlapping area, it is necessary to calculate the ratio between this overlapping area and the total area of the initially extracted dock set. This ratio reflects the overlapping degree between the initially extracted result and the extracted result of the extended region, that is, the dock extraction similarity. The specific calculation method is to divide the overlapping area by the total area of the initially extracted dock set to obtain a proportional value between 0 and 1. Among them, the closer this value is to 1, the higher the overlapping degree between the two sets, and the higher the similarity of dock extraction.

[0143] Step S2232, when it is determined that the dock extraction similarity is greater than the preset similarity threshold, take the extraction result in the initially extracted dock set as the final dock recognition result; otherwise, discard the extraction result in the initially extracted dock set.

[0144] Specifically, when it is determined that the similarity of the extracted docks is greater than the preset similarity threshold, it indicates that there is a high degree of overlap and consistency between the initially extracted dock set and the dock set extracted from the overlapping area of the extended region. Therefore, the extraction result in the initially extracted dock set can be used as the final dock recognition result. This decision is based on the reliability of the similarity calculation, ensuring the accuracy and credibility of the recognition result.

[0145] In one embodiment, the value of the similarity threshold is 0.5.

[0146] Specifically, the selection value of the similarity threshold can refer to the analysis result of Table 1 mentioned above, which will not be elaborated here.

[0147] In one embodiment, in step S3, for the pre-acquired target area ship AIS trajectory data, analyze the staying behavior of the ship's movement trajectory and extract the ship's staying points, including:

[0148] Step S31: Based on the ship AIS trajectory data, traverse the movement trajectories of each ship in the area.

[0149] S32: During the traversal process, use the time interval and position change between adjacent trajectory points to calculate the continuous movement speed of the ship and the number of trajectory points of the movement trajectory, and regard the trajectory segment with a continuous movement speed less than the preset speed threshold and the number of movement trajectory points greater than the preset point threshold as the ship staying behavior;

[0150] Specifically, first, it is necessary to traverse the obtained ship AIS trajectory data and process the trajectory point information of each ship one by one. Then, for adjacent trajectory points, use the time interval and position change between them to calculate the continuous movement speed of the ship. At the same time, it is also necessary to count the trajectory points of each ship to count the total number of points of its movement trajectory. Finally, screen the trajectory segments according to the preset speed threshold and point threshold, that is, when it is determined that the continuous movement speed of a certain trajectory segment is less than the preset speed threshold and the number of its movement trajectory points is greater than the preset point threshold, it is considered that the trajectory segment has significant staying or low-speed movement characteristics and can be retained as the AIS trajectory segment.

[0151] S33: For the AIS trajectory segment of each ship's staying behavior, use the centroid point of all trajectory (space) points in the trajectory segment as the ship's staying point.

[0152] It should be noted that the corresponding centroid point is determined based on the positions of all the trajectory points within the segment as the stop point of the ship within this trajectory segment. This is because the centroid point can better reflect the average position of the ship within this area. Especially when the ship is sailing at a low speed or staying for a long time, the centroid point can provide a relatively accurate and stable reference point. Through such processing, a corresponding stop point can be determined for each AIS trajectory segment, which is convenient for subsequent calculation and analysis of the utilization frequency of the terminal and the terminal utilization rate. At the same time, since the centroid point is calculated based on the positions of all the trajectory points, it can also reduce the influence of the error of a single trajectory point on the final result to a certain extent, and improve the accuracy and reliability of stop point recognition.

[0153] In one embodiment, in step S4, the statistical analysis and comprehensive duration comparison of the ship mooring parameters are performed to obtain the shoreline terminal utilization status of the target area, including:

[0154] The terminal utilization status includes the terminal utilization frequency and the terminal utilization rate;

[0155] The total number of stop points of the terminal within the unit time T in the ship mooring parameters is statistically counted and used as the terminal utilization frequency;

[0156] The total stay duration of the ship stay behaviors corresponding to all the stop points within the terminal buffer area within the unit time T in the ship mooring parameters and the total business hours of the terminal are statistically counted. Based on the comparison and calculation of the total stay duration and the total business hours, the terminal utilization rate is obtained.

[0157] Step S41, statistically count the total number of stop points within the unit time T and use it as the terminal utilization frequency.

[0158] Step S42, statistically count the total stay duration of the ship stay behaviors corresponding to all the stop points within the terminal buffer area within the unit time T and the total business hours of the terminal. Based on the comparison and calculation of the total stay duration and the total business hours, the terminal utilization rate is obtained.

[0159] Specifically, based on steps S41 to S42, it should be noted that this application will statistically count the number and stay time of the ship stop points falling within the buffer area, and calculate the number of stop points within the unit time T (such as one month or one year), and use it as the terminal utilization frequency (such as number / month, number / year, times / month). Taking a certain main stream area of the Yangtze River as the target area, please refer to Figure 6 , the spatial distribution of the terminal utilization frequency can be intuitively displayed through the annotation of different patterns. Such as Figure 6As shown, this application divides the frequency of dock utilization in the area into three levels, namely: 0-50 times / month, 51-100 times / month and 101-427 times / month. These levels are marked with different symbols to distinguish the active degree of dock utilization in different areas. It should be noted that this spatial distribution map not only shows the order of magnitude of dock utilization frequency, but also reveals the hot spots and relatively deserted areas of dock utilization through geographical location information, which can provide important value reference for port management.

[0160] Specifically, based on the docking time and business hours of the terminal, this application calculates the terminal utilization rate through the following formula:

[0161]

[0162] Where N represents the number of days in the unit time T (such as a month or a year), 〖T(i)〗_Dwell time (≤24h) represents the union of the dwell time of all trajectory segments in the terminal buffer corresponding to the i-th day, that is, the total dwell time of all trajectory segments in the terminal buffer, and 〖T(i)〗_Business hours represents the business hours of the corresponding terminal on the i-th day.

[0163] In the current embodiment, based on the ratio of the total stay time and the total business hours, the actual utilization efficiency of the terminal within a given time is reflected through the ratio.

[0164] Taking a Yangtze River main basin as the target area, please refer to Figure 7 , the utilization rate of the wharf presents different distribution characteristics in space. Specifically, Figure 7 As shown in the figure, this application divides the terminal utilization rate into three intervals: 0.00-0.25, 0.25-0.75 and 0.75-1.00. These intervals are marked with different symbols to intuitively show the differences in terminal utilization rates. Through this figure, we can clearly see the changes in terminal utilization rates in different geographical locations, which is of great significance for understanding terminal operation efficiency and optimizing resource allocation.

[0165] In a specific implementation case, taking a main Yangtze River basin as the target area, a shoreline terminal utilization status monitoring disclosed in this application specifically includes the following steps:

[0166] (1) Obtain Google's high-resolution satellite remote sensing image data and use the trained Yolov5 deep learning model to detect the location of the dock. The specific steps are as follows:

[0167] a) Obtain Google satellite remote sensing image data with a resolution of 0.1 meters and the corresponding OSM data of the city, and use "value = pier" to extract vector data representing the pier from the OSM data;

[0168] b) Establish a regular grid with a size of 500m × 500m centered on the center point of the wharf vector data;

[0169] c) Visually interpret the actual wharves in the Google satellite remote sensing image data within the current 500m × 500m unit grid (as Figure 2 shown), and take the minimum bounding rectangle of the wharf range as the training label of the wharf;

[0170] d) Use the training label and the corresponding Google satellite remote sensing image data as training samples to train the Yolov5 deep learning model;

[0171] e) Establish a regular grid with a size of 500m × 500m in the target area, and use the trained Yolov5 model to extract the wharves within each regular grid as Result I;

[0172] f) Then, establish a regular grid with a size of 1500m × 1500m centered on each regular grid (as Figure 3 shown), use the trained Yolov5 model to extract all the wharves within the 1500m × 1500m grid, and take the wharves within the center of the 500m × 500m grid as Result II;

[0173] g) Calculate the similarity of the extraction results I and II, that is, divide the overlapping area of the two extraction results by the extraction area of Result I. If the similarity is greater than 0.5, keep the extraction result in Result I; otherwise, discard the extraction result in Result I.

[0174] (2) Based on the obtained AIS trajectory data, traverse the movement trajectories of each ship; consider the trajectory segments with a continuous movement speed less than 1 byte and a trajectory point count greater than 5 as ship staying behaviors. For each AIS trajectory segment of the ship staying behavior, take the centroid point of all the trajectory (space) points within the trajectory segment as the ship staying point. The specific steps are as follows:

[0175] a) Obtain the ship AIS trajectory data of the main Yangtze River basin within the target area;

[0176] b) Traverse the movement trajectories of each ship in turn, and extract the trajectory segments with a continuous movement speed less than 1 byte and a trajectory point count greater than 5 as ship staying behaviors;

[0177] c) Take the centroid point of all the trajectory (space) points within the AIS trajectory segment of each ship staying behavior as the ship staying point of this trajectory segment, and count the start time and end time of the trajectory segment.

[0178] (3) Establish a buffer based on the identified dock data, and count the number of ship stay points and the stay time falling within the buffer, so as to calculate the dock utilization frequency and the dock utilization rate. The specific steps are as follows:

[0179] a) Based on the dock center points identified from the high-resolution satellite remote sensing image data in step (1), establish a buffer with a radius of 100 m;

[0180] b) Count the number of ship stay points and the stay time falling within the buffer, calculate the total number of stay points within one month, take it as the dock utilization frequency, and calculate the dock utilization rate within one month according to the aforementioned dock utilization rate calculation formula.

[0181] In summary, the present application proposes an innovative method for monitoring the utilization status of shoreline docks. This method combines high-resolution satellite remote sensing image data and AIS trajectory data to achieve precise monitoring of the utilization of docks in a certain main stream of the Yangtze River. By training the Yolov5 deep learning model to automatically identify and extract the dock locations, the monitoring efficiency and accuracy are greatly improved. At the same time, by analyzing the ship movement trajectories using AIS trajectory data, the ship stay points are accurately identified, providing a reliable data basis for calculating the dock utilization frequency and utilization rate. This method is not only applicable to a certain main stream of the Yangtze River, but also can be extended to other similar waters, providing strong technical support for fields such as port management, resource optimization, and environmental protection.

[0182] According to another aspect of the embodiments of the present application, as Figure 8 shown, there is also provided a monitoring device for implementing the above-mentioned method for monitoring the utilization status of shoreline docks. The device may include:

[0183] A dock identification module, configured to obtain the comprehensive geospatial data of the target area, and extract dock data from the comprehensive geospatial data for multi-scale detection using a deep learning prediction model to obtain dock location information;

[0184] A ship stay point analysis module, configured to obtain the ship AIS trajectory data of the target area, and analyze the stay behavior of the ship movement trajectory according to the ship AIS trajectory data to extract ship stay points;

[0185] A dock buffer analysis module, configured to establish a buffer based on the dock location information, and count the ship berthing parameters corresponding to the ship stay points falling within the buffer;

[0186] A status statistics and monitoring module, configured to perform statistical analysis and comprehensive duration comparison on the ship berthing parameters to obtain the utilization status of the shoreline docks in the target area.

[0187] It should be noted here that the implementation examples and scenarios of the above modules and corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should be noted that the above modules, as part of the device, can run in a hardware environment, can be implemented by software, or can be implemented by hardware, where the hardware environment includes a network environment.

[0188] According to another aspect of the embodiments of the present application, a storage medium is also provided. Optionally, in this embodiment, the above storage medium can be used to execute the program code of any one of the above shoreline wharf utilization status monitoring methods in the embodiments of the present application.

[0189] Optionally, in this embodiment, the storage medium is set to store program code for executing the following steps:

[0190] S1, obtain the comprehensive geospatial data of the target area, use the deep learning prediction model to extract wharf data from the comprehensive geospatial data for multi-scale detection, and obtain the wharf location information;

[0191] S2, obtain the ship AIS trajectory data of the target area, and analyze the staying behavior of the ship movement trajectory according to the ship AIS trajectory data, and extract the ship staying points;

[0192] S3, establish a buffer based on the wharf location information, and count the ship berthing parameters corresponding to the ship staying points falling into the buffer;

[0193] S4, perform statistical analysis and comprehensive duration comparison on the ship berthing parameters to obtain the utilization status of the shoreline wharf in the target area.

[0194] Optionally, the specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be elaborated here in this embodiment.

[0195] Among them, the computer-readable storage medium can include but is not limited to any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, micro drives, and magneto-optical disks, ROM, RAM, EPROM, EEPROM, DRAM, VRAM, flash memory devices, magnetic cards or optical cards, nano-systems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0196] According to another aspect of the embodiments of the present application, an electronic device for implementing the above-mentioned shoreline wharf utilization status monitoring method's grabbing method is also provided. The electronic device can be a server, a terminal, or a combination thereof.

[0197] Figure 9 is a schematic structural diagram of an optional electronic device according to the embodiments of the present application, as Figure 9As shown, it includes a processor 902, a communication interface 904, a memory 906, and a communication bus 908. Among them, the processor 902, the communication interface 904, and the memory 906 communicate with each other through the communication bus 908. Among them,

[0198] The memory 906 is used to store computer programs;

[0199] The processor 902, when executing the computer program stored on the memory 906, implements the following steps:

[0200] S1. Obtain the comprehensive geospatial data of the target area, and use the deep learning prediction model to extract dock data from the comprehensive geospatial data for multi-scale detection to obtain the dock location information;

[0201] S2. Obtain the ship AIS trajectory data of the target area, and analyze the stay behavior of the ship movement trajectory according to the ship AIS trajectory data to extract the ship stay points;

[0202] S3. Establish a buffer based on the dock location information, and count the ship berthing parameters corresponding to the ship stay points falling into the buffer;

[0203] S4. Perform statistical analysis and comprehensive duration comparison on the ship berthing parameters to obtain the utilization status of the shoreline docks in the target area.

[0204] Optionally, the communication bus can be a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 9 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used for communication between the above-mentioned electronic device and other devices.

[0205] The memory can include RAM, and can also include non-volatile memory, for example, at least one disk memory. Optionally, the memory can also be at least one storage device located far from the aforementioned processor.

[0206] As an example, the above-mentioned memory 906 includes, but is not limited to, other module units in the above-mentioned device for monitoring the utilization status of the shoreline docks, which will not be elaborated in this example.

[0207] The above-mentioned processor may be a general-purpose processor, including but not limited to: CPU (Central Processing Unit, central processing unit), NP (Network Processor, network processor), etc.; it may also be a DSP (Digital Signal Processing, digital signal processor), ASIC (Application Specific Integrated Circuit, application-specific integrated circuit), FPGA (Field-Programmable Gate Array, field programmable gate array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0208] Optionally, specific examples in this embodiment may refer to the examples described in the above-mentioned embodiment, and details are not described herein again.

[0209] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, some steps may be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0210] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0211] In several embodiments provided by this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some service interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0212] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0213] In addition, in each embodiment of the present application, the functional units may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional unit.

[0214] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned memory includes: USB flash drives, read-only memory (ROM), random access memory (RAM), mobile hard disks, magnetic disks, or optical discs, etc., which can store program codes.

[0215] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. This program can be stored in a computer-readable memory, and the memory can include: flash drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs, etc.

[0216] The above are only exemplary embodiments of the present disclosure, and the scope of the present disclosure cannot be limited thereby. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure still fall within the scope covered by the present disclosure. After considering the specification and practicing the present disclosure herein, those skilled in the art will readily think of other embodiments of the present disclosure. The present application aims to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the technical field not described in the present disclosure. The specification and embodiments are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

[0217] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope described in this specification.

[0218] Those skilled in the art can easily understand that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for monitoring the utilization status of a shoreline wharf, characterized in that Including: Obtain the comprehensive geospatial data of the target area, and use a deep learning prediction model to extract quay data from the comprehensive geospatial data for multi-scale detection to obtain quay location information; Obtain the vessel AIS trajectory data of the target area, and analyze the staying behavior of the vessel movement trajectory according to the vessel AIS trajectory data to extract vessel staying points; Establish a buffer zone based on the quay location information, and count the vessel berthing parameters corresponding to the vessel staying points falling within the buffer zone; Conduct statistical analysis and comprehensive duration comparison on the vessel berthing parameters to obtain the utilization status of the shoreline quay in the target area.

2. The method for monitoring the utilization status of the shoreline quay according to claim 1, characterized in that The step of using a deep learning prediction model to extract quay data from the comprehensive geospatial data for multi-scale detection to obtain quay location information includes: Perform quay interpretation based on the comprehensive geospatial data to obtain an interpretation result and use the interpretation result to train a deep learning model; Input the comprehensive geospatial data into the trained deep learning model to obtain quay data and perform multi-scale detection to obtain quay location information.

3. The method for monitoring the utilization status of the shoreline quay according to claim 2, characterized in that The step of performing quay interpretation based on the comprehensive geospatial data to obtain an interpretation result and using the interpretation result to train a deep learning model includes: Based on the quay label identification, screen out the target vector data corresponding to the quay from the comprehensive geospatial data; Taking the center point of the target vector data as the geometric center, establish a first regular grid with a preset scale size of M×M; For each first regular grid, obtain the satellite remote sensing image covering the grid, and perform quay interpretation and identification based on the satellite remote sensing image; Within the quay area identified by interpretation, determine the minimum bounding rectangle of the quay, and construct a training label for the quay based on the position information of the minimum bounding rectangle; Construct a training sample based on the quay training label and the corresponding satellite remote sensing image data, and train a deep learning model based on the training sample.

4. The method for monitoring the utilization status of the shoreline quay according to claim 2, characterized in that The step of inputting the comprehensive geospatial data into the trained deep learning model to obtain quay data and perform multi-scale detection to obtain quay location information includes: Establish multiple first regular grids with a preset scale size of M×M in the target area, use the trained deep learning model to identify and extract the quay data within each first regular grid to obtain a preliminary extracted quay set; Taking each first regular grid as the center, establish a second regular grid with a preset scale size of 3M×3M, and use the trained deep learning model to identify and extract the quay data within the second regular grid to obtain an extracted quay set in the overlapping area between the second regular grid and the first regular grid; Perform similarity calculation and result verification based on the preliminary extracted quay set and the extracted quay set in the overlapping area between the second regular grid and the first regular grid to determine the quay recognition result and obtain the quay location information.

5. The shoreline wharf utilization status monitoring method according to claim 4, wherein the similarity calculation and result verification are performed based on the initially extracted wharf set and the wharf set extracted from the overlapping area between the second regular grid and the first regular grid to determine the wharf recognition result and obtain the wharf location information, including: calculating the intersection area between the initially extracted wharf set and the wharf set extracted from the overlapping area of the two grids, and dividing the value by the area of the initially extracted wharf set to obtain the wharf extraction similarity; when it is determined that the wharf extraction similarity is greater than the preset similarity threshold, taking the extraction result in the initially extracted wharf set as the final wharf recognition result; otherwise, discarding the extraction result in the initially extracted wharf set.

6. The shoreline wharf utilization status monitoring method according to any one of claims 1-5, wherein acquiring the ship AIS trajectory data of the target area, and analyzing the staying behavior of the ship movement trajectory according to the ship AIS trajectory data to extract the ship staying points, including: traversing the movement trajectories of each ship in the area based on the ship AIS trajectory data; during the traversal, using the time interval and position change between adjacent trajectory points to calculate the continuous movement speed of the ship and the number of trajectory points of the movement trajectory, and regarding the trajectory segment with the continuous movement speed less than the preset speed threshold and the number of movement trajectory points greater than the preset point threshold as the ship staying behavior; for the AIS trajectory segment of each ship staying behavior, taking the centroid point of all trajectory (space) points in the trajectory segment as the ship staying point.

7. The shoreline wharf utilization status monitoring method according to any one of claims 1-5, wherein performing statistical analysis and comprehensive duration comparison on the ship berthing parameters to obtain the shoreline wharf utilization status of the target area, including: the wharf utilization status includes the wharf utilization frequency and the wharf utilization rate; counting the total number of staying points of the wharf within the unit time T in the ship berthing parameters and taking it as the wharf utilization frequency; counting the total staying duration of the ship staying behaviors corresponding to all staying points in the wharf buffer area within the unit time T in the ship berthing parameters and the total business hours of the wharf, and performing a comparison calculation based on the total staying duration and the total business hours to obtain the wharf utilization rate.

8. A shoreline wharf utilization status monitoring device, characterized in that, including: a wharf recognition module, configured to acquire the comprehensive geospatial data of the target area, and use a deep learning prediction model to extract wharf data from the comprehensive geospatial data for multi-scale detection to obtain the wharf location information; a ship staying point analysis module, configured to acquire the ship AIS trajectory data of the target area, and analyze the staying behavior of the ship movement trajectory according to the ship AIS trajectory data to extract the ship staying points; a wharf buffer area analysis module, configured to establish a buffer area based on the wharf location information and count the ship berthing parameters corresponding to the ship staying points falling into the buffer area; a status statistical monitoring module, configured to perform statistical analysis and comprehensive duration comparison on the ship berthing parameters to obtain the shoreline wharf utilization status of the target area.

9. An electronic device, comprising: a processor; and a memory in which computer program instructions are stored, and when the computer program instructions are run by the processor, the processor is caused to execute the shoreline terminal utilization status monitoring method according to any one of claims 1-7.

10. A computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are run by a processor, the processor is caused to execute the shoreline terminal utilization status monitoring method according to any one of claims 1-7.

Images