An automatic guidance and management method and system for fishing boats entering the port based on intelligent recognition

Abstract

The present invention discloses a method and system for automatic guidance management of fishing boats entering the port based on intelligent recognition, including: dividing the port into several sub-regions and conducting functional analysis, defining the functional attributes of each region, and setting up a ship entering-port monitoring array; tracking and monitoring the ships entering the port in real time through the monitoring array, and using edge computing to conduct intelligent recognition of the ships. Matching the recognition information with a preset database to determine whether the target ship is an identified ship. If it is an identified ship, obtain the real-time berthing area information, calculate the matching berth in combination with the real-time position and ship identification information, generate a berthing guidance plan, and provide route and speed suggestions; if it is not an identified ship, generate a notification instruction based on a preset response template, notify it to handle the registration through a broadcast array, and at the same time, formulate a registration guidance plan and push it in combination with the real-time berthing data of the registration area. Optimize the fishing port entering-port management method and system to improve the entering-port efficiency and safety.

Description

Technical Field

[0001] The present invention relates to the technical field of fishing vessel port entry management, and in particular, to an automatic guidance management method and system for fishing vessel port entry based on intelligent recognition. Background Art

[0002] Intelligent recognition refers to the recognition by an electronic checkpoint camera. Specifically, a set of cameras with different functions and algorithms are installed at the fishing port entry position as an electronic checkpoint. The cameras are adjusted to a specified angle according to the needs of the recognition range (usually covering the entry and exit channels). The passing ships in the recognition area are tracked and photographed, and evidence is captured. At the same time, the edge computing ability is mobilized to quickly identify the specific information of the ships from the video, such as the MMSI number, ship name, size, etc.

[0003] The automatic guidance for fishing vessel port entry based on intelligent recognition refers to the process of data fusion, comparison, and extraction of the intelligent recognition results with the ship data in the background system of the fishing vessel port entry automatic guidance, and finally sending an automatic guidance message to the recognized ship. First, the intelligent recognition results (including ship information and positioning information) are uploaded to the background system of the fishing vessel port entry automatic guidance, compared with the ships already included in the management in the system, the ship is screened out, and the account number and mobile phone number of the ship in the system are obtained, so as to be able to send an automatic guidance message to the mobile phone number and mini-program account number of the ship. At the same time, for ships entering the port for the first time or not registered in the system, the process of ship registration is completed by means of broadcast notification and ship code scanning to send an automatic guidance message. This method is to perform data linkage and processing on the four methods of electronic checkpoint, fishing vessel port entry automatic guidance management system, broadcast, and two-dimensional code recognition with data, realize automatic collection and acquisition of the information of the entering ships, realize the automatic guidance of fishing vessel port entry, and optimize the port entry management of the fishing port.

[0004] In the existing fishing port entry management, some may also install electronic checkpoints, but the use of electronic checkpoints is often only reflected in abnormal monitoring, such as video recognition of abnormal port entry or capturing evidence of abnormalities. There are also fishing port entry guides, but most of them are through manual offline intervention. It is impossible to conduct data intercommunication between the electronic checkpoint recognition results and the fishing port management system, and there is no further use of the data. Summary of the Invention

[0005] The present invention overcomes the defects of the prior art and provides an automatic guidance management method and system for fishing vessel port entry based on intelligent recognition. An important purpose is to optimize the fishing port entry management method and system and improve the port entry efficiency and safety.

[0006] To achieve the above object, the first aspect of the present invention provides an automatic guidance management method for fishing vessel port entry based on intelligent recognition, including:

[0007] Obtain the map of the target port area, divide the target port into several sub-areas, conduct a functional analysis on each sub-area, define the functional attributes of each area of the target port based on the results of the functional analysis, and set up a ship-in-port monitoring array;

[0008] Monitor the incoming ships through the set ship-in-port monitoring array, monitor the real-time incoming ships and track the ship movements, and conduct intelligent identification of the target incoming ships based on edge computing to obtain ship intelligent identification information;

[0009] Retrieve in the preset ship information database according to the ship intelligent identification information, determine whether the target incoming ship is an identified ship, and give guidance to the target incoming ship based on the judgment result;

[0010] If the target incoming ship is an identified ship, obtain the real-time ship berthing area information, calculate and match the berth by combining the real-time incoming position and ship identification information of the target incoming ship, and formulate a berthing guidance plan to provide the berthing route and speed;

[0011] If the target incoming ship is not an identified ship, generate a notification instruction based on the preset response template and send it to the broadcast array to notify the target ship to conduct ship registration, obtain the real-time berthing status data of the ship registration area, and formulate a ship registration guidance plan for pushing.

[0012] In this solution, the obtaining the map of the target port area, dividing the target port into several sub-areas, conducting a functional analysis on each sub-area, defining the functional attributes of each area of the target port based on the results of the functional analysis, and setting up a ship-in-port monitoring array specifically includes:

[0013] Obtain the map of the target port area based on data retrieval, divide the target port into several sub-areas according to the map of the target port area, extract the regional environmental characteristics of each sub-area to obtain the sub-area environmental characteristic information;

[0014] Calculate the cosine similarity values between each sub-area according to the sub-area environmental characteristic information, merge the sub-areas with cosine similarity values greater than the preset threshold, and conduct category integration on each sub-area to obtain the merged sub-area information;

[0015] Extract the environmental attribute characteristics, building attribute characteristics and layout attribute characteristics within each merged sub-area according to the map of the target port area, conduct regional functional analysis through the attributes of each merged sub-area to obtain functional analysis information;

[0016] Assign function labels to each merged sub-area according to the functional analysis information, conduct regional function definition on each merged sub-area, extract the ship-in-port area of the target port area based on the regional kinetic energy definition result, and set up a ship-in-port monitoring array.

[0017] In this solution, the incoming ships are monitored by a set ship incoming port monitoring array, and the real-time incoming ships are monitored and the ship movement is tracked. Specifically, it includes:

[0018] Based on the ship incoming port monitoring array set in the ship incoming port area of the target port, the incoming ships entering the target port in real time are monitored. The YOLOv5 is used to build a target detection model, and the data collected in real time by the camera array in the ship incoming port area is input into the target monitoring model for ship detection;

[0019] The backbone network of the target detection model is built using CSPDarkNet. Through the built backbone network, the features of the real-time collected images input into the model are extracted, and multi-level features of the real-time collected images are obtained to form the original feature map;

[0020] The attention mechanism is introduced to calculate the attention scores of different-level features and generate the attention feature map. The FPN and PAN structures are used to fuse the attention feature map with the original feature map, and target detection is performed through the fused feature map to obtain the target incoming ship detection information;

[0021] The target incoming ship detection information is imported into the Kalman filter to estimate the motion state of the target incoming ship to predict the motion state at the next moment, and the joint matching method is used to analyze whether the predicted motion state matches the motion state observation value at the next moment;

[0022] If it does not match the motion state observation value at the next moment, the historical matching motion state observation value is extracted as training data and imported into the Bayesian inference network for training to obtain the transient posterior distribution;

[0023] Based on the obtained transient posterior distribution, the motion state at the moment before the moment when the mismatched motion state appears is inferred in the Bayesian network to obtain the inferred motion state, and a prediction trajectory is generated according to the inferred motion state. The target incoming ship is locked using the prediction trajectory to obtain the target incoming ship tracking information.

[0024] In this solution, the target incoming ship is intelligently identified based on edge computing to obtain ship intelligent identification information. Specifically, it includes:

[0025] The target incoming ship tracking information is obtained, and the monitoring video stream data of the target incoming ship per unit time is extracted based on the target incoming ship tracking information;

[0026] An intelligent identification model is built based on a dual-channel convolutional neural network. The intelligent identification model includes a first channel and a second channel, and the monitoring video stream data of the target incoming ship per unit time is input into the intelligent identification model for analysis;

[0027] Extract the appearance features and motion features of the input monitored video stream data through the first channel, and extract the hull text features of the target incoming ship in the monitored video stream data through the second channel;

[0028] Input the extracted ship appearance features and hull text features into the fully connected layer for feature fusion by means of joint embedding to obtain fusion features, and analyze the type, position, size and hull text semantics of the target incoming ship based on the fusion features to obtain ship intelligent recognition information.

[0029] In this solution, retrieve in the preset ship information database according to the ship intelligent recognition information, determine whether the target incoming ship is an identified ship, and guide the target incoming ship based on the judgment result, specifically including:

[0030] Obtain the ship intelligent recognition information, extract the MMSI number of the target incoming ship according to the ship intelligent recognition information, form the first retrieval label, and retrieve in the preset ship information database based on the first retrieval label;

[0031] If there is matching data for the first retrieval label in the preset ship information database, it means that the target incoming ship is an identified ship in the current port, then mark the target incoming ship as an identified ship;

[0032] If there is no matching data for the first retrieval label in the preset ship information database, it means that the target incoming ship is an unrepresented ship in the current port, then mark the target incoming ship as a non-identified ship;

[0033] If the MMSI number of the target incoming ship cannot be extracted through the ship intelligent recognition information, then extract the appearance features and hull text semantic features of the target ship to generate a feature portrait of the target incoming ship;

[0034] Based on the fuzzy matching method, calculate the similarity between the feature portrait of the target incoming ship and the ship data of each identified ship stored in the ship information database, and obtain the identified ship data with the highest similarity to the feature portrait of the target incoming ship;

[0035] Judge the similarity value of the identified ship data with the highest similarity to the feature portrait of the target incoming ship with a preset threshold. If it is greater than the preset threshold, it means that the target incoming ship is an identified ship. If it is less than the preset threshold, it means that the target incoming ship is a non-identified ship.

[0036] In this solution, if the target incoming ship is an identified ship, real-time ship docking area information is obtained, and a matching berth is calculated by combining the real-time incoming position of the target incoming ship and the ship identification information, and a docking guidance plan is formulated to provide a docking route and speed, specifically including:

[0037] If, after retrieving in the preset ship information database, there is a matching ship, which means the target incoming ship has been registered and filed as an identified ship in the system, then the incoming upload data of the target incoming ship is obtained;

[0038] Check whether there are differences between the incoming upload data of the target incoming ship and the ship intelligent identification information. If there are differences, the target ship is notified using a broadcast array to remind the target ship to re-register;

[0039] If there are no differences, a docking guidance plan is formulated based on the ship intelligent identification information and the incoming upload data of the target incoming ship, and real-time ship docking area information is obtained. The real-time ship docking area information includes the ship docking status within the docking area and the data of ships waiting to dock;

[0040] Based on the real-time ship docking area information, the characteristics of the ships already docked within the docking area and the characteristics of the ships waiting to dock are extracted, and the remaining berths within the target docking area are calculated according to the extracted characteristics to obtain the remaining berth information;

[0041] Analyze the docking requirements of the target incoming ship based on the ship intelligent identification information and the incoming upload data of the target incoming ship, and perform a matching analysis with the remaining berths to determine whether there is a matching berth for the target incoming ship to dock;

[0042] If there is a matching berth, the target port channel information is obtained and the target port channel characteristics are extracted as intermediate nodes, the regional position characteristics of the matching berth are used as the end node, and the real-time position of the target incoming ship is used as the start node. A preset cost function uses the A* algorithm to generate the initial guidance route for the target incoming ship;

[0043] Obtain the real-time ship navigation information in the target port channel, extract the guidance routes of the ships currently traveling on the channel through the real-time ship navigation information, and compare them with the initial guidance route to analyze whether there are overlapping routes;

[0044] If there are overlapping routes, extract the channel position characteristics of the overlapping routes, and analyze whether a detour is possible through the real-time ship navigation information. If a detour is possible, the detour route is merged into the initial guidance route to generate the final docking guidance route;

[0045] If it is not possible to change the route, use the initial guiding route as the final docking guiding route, extract the ship navigation characteristics corresponding to the overlapping route, and calculate the following navigation speed of the target inbound ship on the overlapping route to generate a ship driving suggestion;

[0046] Based on the ship information database, obtain the system account and registered mobile phone number of the target inbound ship, send the final docking guiding route and the ship driving suggestion to the system account of the target inbound ship, and notify through the mobile phone number.

[0047] In this solution, if the target inbound ship is not an identified ship, a notification instruction is generated based on a preset response template and sent to the broadcast array to notify the target ship for ship registration, and the real-time docking status data of the ship registration area is obtained, and a ship registration guiding plan is formulated and pushed, specifically including:

[0048] If the target inbound ship is not found to have a matching ship after being retrieved in the preset ship information database, it means that the ship has not been registered and filed in the system. Then, obtain the ship intelligent identification information, and extract the identification characteristics of the target non-identified ship through the ship intelligent identification information;

[0049] Based on the identification characteristics of the target non-identified ship, combine with the preset response template to generate a notification instruction, and send the notification instruction to the broadcast array to notify the target non-identified ship;

[0050] Obtain the real-time docking status data of the ship registration area, formulate a ship registration guiding plan in combination with the identification characteristics of the target non-identified ship, and push it to the target ship for offline filing and registration.

[0051] The second aspect of the present invention provides an intelligent identification-based automatic guiding management system for fishing boats entering the port. The system includes: a memory and a processor. The memory contains an intelligent identification-based automatic guiding management method program for fishing boats entering the port. When the intelligent identification-based automatic guiding management method program is executed by the processor, the following steps are implemented:

[0052] Obtain the map of the target port area, divide the target port into several sub-areas, conduct a functional analysis on each sub-area, define the functional attributes of each area of the target port based on the results of the functional analysis, and set up a ship entry monitoring array;

[0053] Monitor the inbound ships through the set ship entry monitoring array, monitor the real-time inbound ships and conduct ship movement tracking, and conduct intelligent identification on the target inbound ship based on edge computing to obtain ship intelligent identification information;

[0054] Retrieve according to the intelligent identification information of the ship in the preset ship information database, determine whether the target incoming ship is an identified ship, and guide the target incoming ship based on the judgment result;

[0055] If the target incoming ship is an identified ship, obtain the real-time ship berthing area information, calculate the matching berth by combining the real-time incoming position of the target incoming ship and the ship identification information, and formulate a berthing guidance plan to provide the berthing route and speed;

[0056] If the target incoming ship is not an identified ship, generate a notification instruction based on the preset response template and send it to the broadcast array to notify the target ship to conduct ship registration, obtain the real-time berthing status data of the ship registration area, and formulate a ship registration guidance plan for pushing.

[0057] The present invention discloses an automatic guidance management method and system for fishing boats entering the port based on intelligent identification, including: dividing the port into several sub-regions and conducting functional analysis, defining the functional attributes of each region, and setting up a ship incoming port monitoring array; real-time tracking and monitoring of incoming ships through the monitoring array, and intelligent identification of ships using edge computing. Matching the identification information with the preset database to determine whether the target ship is an identified ship. If it is an identified ship, obtain the real-time berthing area information, combine the real-time position and the ship identification information, calculate the matching berth, generate a berthing guidance plan, and provide route and speed suggestions; if it is not an identified ship, generate a notification instruction based on the preset response template, notify it to handle the registration through the broadcast array, and at the same time, combine the real-time berthing data of the registration area, formulate a registration guidance plan and push it. Optimize the fishing port incoming port management method and system to improve the incoming port efficiency and safety. Brief Description of the Drawings

[0058] In order to more clearly illustrate the technical solutions in the embodiments or exemplifications of the present invention, the following will briefly introduce the drawings required for use in the embodiments or exemplifications. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings shown without creative efforts.

[0059] Figure 1 It is a flowchart of an automatic guidance management method for fishing boats entering the port based on intelligent identification provided by an embodiment of the present invention;

[0060] Figure 2 It is a flowchart of an automatic guidance method for fishing boats to berth in the port provided by an embodiment of the present invention;

[0061] Figure 3 It is a block diagram of an automatic guidance management system for fishing boats entering the port based on intelligent identification provided by an embodiment of the present invention;

[0062] The implementation, functional features, and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed implementation manners

[0063] In order to more clearly understand the above objects, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0064] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0065] Figure 1 It is a flowchart of a method for automatically guiding and managing the entry of fishing boats into a port based on intelligent recognition provided by an embodiment of the present invention;

[0066] As Figure 1 shown, the present invention provides a flowchart of a method for automatically guiding and managing the entry of fishing boats into a port based on intelligent recognition, including:

[0067] S102. Obtain a map of the target port area, divide the target port into several sub-areas, conduct a functional analysis on each sub-area, define the functional attributes of each area of the target port based on the results of the functional analysis, and set up a ship entry monitoring array;

[0068] S104. Monitor the incoming ships through the set ship entry monitoring array, monitor the real-time incoming ships and track the ship movement, and conduct intelligent recognition on the target incoming ships based on edge computing to obtain ship intelligent recognition information;

[0069] S106. Retrieve in a preset ship information database according to the ship intelligent recognition information, judge whether the target incoming ship is an identified ship, and guide the target incoming ship based on the judgment result;

[0070] S108. If the target incoming ship is an identified ship, obtain the real-time ship docking area information, calculate and match a berth in combination with the real-time entry position and ship identification information of the target incoming ship, and formulate a docking guidance plan to provide a docking route and speed;

[0071] S110. If the target incoming ship is not an identified ship, generate a notification instruction based on a preset response template and send it to a broadcast array to notify the target ship to conduct ship registration, obtain the real-time docking status data of the ship registration area, and formulate a ship registration guidance plan for pushing.

[0072] It should be noted that the present invention provides a method and system for automatically guiding the entry of fishing vessels into the port based on intelligent identification. The method obtains the information of the incoming ships by introducing the intelligent identification of the incoming ships in the electronic bayonet video stream, automatically uploads the identification results, and communicates data with the fishing port entry management system to complete data comparison and system account acquisition, thereby sending targeted entry guidance to the incoming ships. At the same time, for ships that enter the port for the first time or ships that are not registered in the system, the ship registration is automatically completed with the assistance of broadcasting and ship code scanning, so that the ships can also receive the entry guidance. The mapping relationship between the fishing vessels and the management system is opened up, which effectively improves the efficiency of fishing vessels entering the port.

[0073] Further, in a preferred embodiment of the present invention, the target port area map is obtained, the target port is divided into a plurality of sub-areas, each sub-area is functionally analyzed, the functional attributes of each area of ​​the target port are defined based on the functional analysis results, and a ship entry monitoring array is set, specifically including:

[0074] Based on data retrieval, a target port area map is obtained, the target port is divided into a plurality of sub-areas according to the target port area map, and feature extraction is performed on each sub-area to obtain regional environmental features of each sub-area, thereby obtaining sub-area environmental feature information;

[0075] Calculate the cosine similarity value between each sub-region according to the sub-region environmental feature information, merge the sub-regions whose cosine similarity value is greater than a preset threshold, and classify each sub-region to obtain merged sub-region information;

[0076] Extracting environmental attribute features, architectural attribute features, and layout attribute features in each merged sub-region according to the target port area map, performing regional functional analysis through the attributes of each merged sub-region, and obtaining functional analysis information;

[0077] According to the functional analysis information, a functional label is assigned to each merged sub-region, and a regional function is defined for each merged sub-region. Based on the regional kinetic energy definition result, the ship entry area of ​​the target port area is extracted, and a ship entry monitoring array is set.

[0078] It should be noted that based on data retrieval, a regional map of the target port is obtained. The target port is divided into several sub-regions according to the map, and feature extraction is performed on these sub-regions. The extracted content includes the environmental feature information of each sub-region, thereby forming an environmental feature dataset of the sub-regions. Then, according to the environmental feature information of the sub-regions, the cosine similarity values between each pair of sub-regions are calculated, and the cosine similarity is used to judge the similarity of the sub-regions. When the cosine similarity value is greater than a preset threshold, the similar sub-regions are merged. Through the merging process, the category integration of the sub-regions of the port is carried out to form merged sub-region data containing similar regions. Based on the integrated sub-regions, the attribute features of each merged sub-region are extracted, including environmental attributes, building attributes, and layout attributes. These attributes constitute the main description information of the merged sub-regions. Through these attribute information, the regional functional analysis of the merged sub-regions is carried out, and the analysis results are used to identify the main functions of each sub-region, such as the berthing area, freight area, or buffer area, etc., thereby generating functional analysis information. According to the functional analysis results, corresponding function labels are assigned to each merged sub-region, and the function labels are attached to the sub-regions, thereby completing the function definition of the region. Finally, based on the results of the regional function definition, the ship arrival area of the target port is extracted. According to the port's requirement for monitoring incoming fishing boats, appropriate points are selected at the arrival position to deploy electronic checkpoint cameras, broadcasts, and two-dimensional code scanning signs.

[0079] Further, in a preferred embodiment of the present invention, the incoming ships are monitored by a set ship arrival monitoring array, and the real-time incoming ships are monitored and the ship movement is tracked, specifically including:

[0080] Based on a ship arrival monitoring array set in the ship arrival area of the target port, the incoming ships entering the target port in real time are monitored. A target detection model is built using YOLOv5, and the data collected in real time by the camera array in the ship arrival area is input into the target monitoring model for ship detection;

[0081] The target detection model uses CSPDarkNet to build the backbone network. Through the built backbone network, feature extraction is performed on the real-time collected images input into the model to obtain multi-level features of the real-time collected images, forming the original feature map;

[0082] The attention mechanism is introduced to calculate the attention scores of different-level features and generate an attention feature map. The FPN and PAN structures are used to fuse the attention feature map with the original feature map, and target detection is performed through the fused feature map to obtain target incoming ship detection information;

[0083] Import the target inbound ship detection information into the Kalman filter to estimate the motion state of the target inbound ship and predict the motion state at the next moment, and analyze whether the predicted motion state matches the motion state observation value at the next moment by means of joint matching;

[0084] If it does not match the motion state observation value at the next moment, extract the historically matched motion state observation values as training data and import them into the Bayesian inference network for training to obtain the transient posterior distribution;

[0085] Based on the obtained transient posterior distribution, infer the motion state at the moment before the moment when the mismatched motion state appears in the Bayesian network to obtain the inferred motion state, generate a predicted trajectory according to the inferred motion state, lock the target inbound ship by using the predicted trajectory, and obtain the target inbound ship tracking information.

[0086] It should be noted that within the ship inbound area of the target port, the ships entering the target port are monitored in real time through the set ship inbound monitoring array, and the YOLOv5 target detection model is used to process the collected data. The camera array will input the real-time collected image data of the ship inbound area into the target detection model for analysis. CSPDarkNet is used as the backbone network, and through this network, the input image data is feature-extracted to generate an original feature map containing multi-level information. The attention mechanism is introduced to calculate the attention scores of different-level features in the original feature map, so as to generate an attention feature map with stronger attention ability, focus on the key target areas, and thus improve the detection accuracy. Then, the feature pyramid network (FPN) and the path aggregation network (PAN) are used to fuse the attention feature map with the original feature map to further optimize the representation ability of the multi-level features. After the target detection is completed, the detection information of the target ship is imported into the Kalman filter to estimate the motion state of the ship, and the possible motion state of the ship at the next moment is predicted through the filtering process. At the same time, the predicted motion state is compared with the actual observation value at the next moment by means of joint matching, and the basis for joint matching is position, speed, and appearance features. If the matching fails, it means that the motion state of the ship has changed unpredictably. At this time, the motion state observation values matching the target are extracted from the historical data and used as training data to be input into the Bayesian inference network, and the network is used to generate the transient posterior distribution. Through the Bayesian inference network, the unmatched state is further inferred to find out the state at the moment before the unmatched motion state, and finally a reasonable motion state prediction result is generated. According to the inferred motion state, a predicted trajectory of the target inbound ship is generated, and the target ship is accurately positioned through this trajectory to obtain the real-time tracking information of the target ship. Thus, the accuracy of the electronic checkpoints in monitoring multi-target inbound is improved.

[0087] Further, in a preferred embodiment of the present invention, the intelligent identification of the target incoming ship based on edge computing to obtain ship intelligent identification information specifically includes:

[0088] Obtain the tracking information of the target incoming ship, and extract the monitoring video stream data of the target incoming ship per unit time based on the tracking information of the target incoming ship;

[0089] Construct an intelligent identification model based on a dual-channel convolutional neural network. The intelligent identification model includes a first channel and a second channel, and input the monitoring video stream data of the target incoming ship per unit time into the intelligent identification model for analysis;

[0090] Extract the appearance features and motion features of the ship from the input monitoring video stream data through the first channel, and extract the hull text features of the target incoming ship in the monitoring video stream data through the second channel;

[0091] Input the extracted ship appearance features and hull text features into the fully connected layer for feature fusion by means of joint embedding to obtain fusion features, and analyze the type, position, size and hull text semantics of the target incoming ship based on the fusion features to obtain ship intelligent identification information.

[0092] It should be noted that the monitoring video stream data of the target ship per unit time is extracted based on the tracking information of the target incoming ship. To achieve the intelligent identification of the ship, an intelligent identification model is constructed based on a dual-channel convolutional neural network. This model contains a first channel and a second channel, which are respectively responsible for different types of feature extraction work. The extracted monitoring video stream data per unit time is used as the input and passed into the intelligent identification model for analysis and processing. First, the input video stream data is processed through the first channel of the model to extract the appearance features and motion features of the ship. The appearance features include the shape, color and surface structure information of the ship, while the motion features include the movement trajectory, speed and dynamic behavior pattern of the ship between video frames, etc. These features can provide the overall visual and behavioral feature information of the ship for identification. At the same time, the hull text features in the video data are extracted through the second channel, including the ship name, registration number (such as MMSI number) and other identifying text information displayed on the ship. After the feature extraction of the two channels is completed, the ship appearance and motion features obtained from the first channel and the hull text features extracted from the second channel are jointly input into the fully connected layer. In the fully connected layer, the features are fused by means of joint embedding to form a fused feature vector. Through this fused feature, the type, position and size of the target ship are analyzed, and the semantic information in the hull text, such as the ship name and registration information, is parsed, providing accurate data support for subsequent ship management and guidance.

[0093] Further, in a preferred embodiment of the present invention, retrieving in a preset ship information database according to the ship intelligent identification information, determining whether the target inbound ship is an identified ship, and guiding the target inbound ship based on the determination result specifically includes:

[0094] Obtain the ship intelligent identification information, extract the MMSI number of the target inbound ship according to the ship intelligent identification information to form a first retrieval tag, and retrieve in the preset ship information database based on the first retrieval tag;

[0095] If there is matching data for the first retrieval tag in the preset ship information database, indicating that the target inbound ship is an identified ship in the current port, then mark the target inbound ship as an identified ship;

[0096] If there is no matching data for the first retrieval tag in the preset ship information database, indicating that the target inbound ship is an unrepresented ship in the current port, then mark the target inbound ship as a non-identified ship;

[0097] If the MMSI number of the target inbound ship cannot be extracted through the ship intelligent identification information, then extract the appearance features and hull text semantic features of the target ship to generate a feature portrait of the target inbound ship;

[0098] Based on the method of fuzzy matching, calculate the similarity between the feature portrait of the target inbound ship and the vessel data of each identified ship stored in the ship information database, and obtain the identified ship data with the highest similarity to the feature portrait of the target inbound ship;

[0099] Judge the similarity value of the identified ship data with the highest similarity to the feature portrait of the target inbound ship against a preset threshold. If it is greater than the preset threshold, it indicates that the target inbound ship is an identified ship; if it is less than the preset threshold, it indicates that the target inbound ship is a non-identified ship.

[0100] It should be noted that through the ship intelligent identification information, the MMSI number of the target inbound ship is extracted and used to form the first retrieval label for retrieval in the preset ship information database. If the retrieval result shows that there is a matching relationship between the first retrieval label and the data in the database, it indicates that the target inbound ship is an identified ship at the current port, and it is marked as an identified ship. On the contrary, if the first retrieval label has no matching data in the ship information database, it indicates that the target inbound ship is an unidentified ship, and the system will mark it as an un-identified ship. In some cases, the MMSI number of the target inbound ship may not be extracted through the ship intelligent identification information. In this regard, the appearance features of the target inbound ship and the text semantic features of its hull are extracted to generate a target ship feature portrait. The feature portrait includes the shape, color, structure, logo of the ship, as well as text semantic features such as signs and number information. Then, the feature portrait of the target inbound ship is compared with all the identified ship data in the ship information database through a fuzzy matching method to calculate the similarity, and the identified ship data with the highest similarity to the feature portrait of the target inbound ship is obtained. Finally, the identification status of the target ship is determined by comparing the similarity value with a preset threshold. If the similarity value is greater than the preset threshold, the target ship is considered an identified ship; if the similarity value is less than the threshold, the target ship is determined to be an un-identified ship. This ensures that even in the case of missing or unclear MMSI numbers, the system can accurately determine the identification status of the target ship through comprehensive feature analysis.

[0101] Further, in a preferred embodiment of the present invention, if the target inbound ship is an identified ship, the real-time ship docking area information is obtained, and the matching berth is calculated by combining the real-time inbound position of the target inbound ship and the ship identification information, and a docking guidance plan is formulated to provide the docking route and speed, specifically including:

[0102] If there is a matching ship after the target inbound ship is retrieved in the preset ship information database, it means that the target inbound ship has been registered and filed as an identified ship in the system, and the inbound upload data of the target inbound ship is obtained;

[0103] Check whether there is a difference between the inbound upload data of the target inbound ship and the ship intelligent identification information. If there is a difference, the target ship is notified using a broadcast array to remind the target ship to re-register;

[0104] If there is no difference, a docking guidance plan is formulated based on the ship intelligent identification information and the inbound upload data of the target inbound ship, and the real-time ship docking area information is obtained. The real-time ship docking area information includes the docking status of the ships in the docking area and the data of the ships waiting to dock;

[0105] Extract the characteristics of the vessels already docked and the vessels waiting to dock within the docking area based on the real-time vessel docking area information, and calculate the remaining berths within the target docking area according to the extracted characteristics to obtain the remaining berth information;

[0106] Analyze the docking requirements of the target inbound vessel based on the vessel intelligent identification information and the inbound upload data of the target inbound vessel, and perform a matching analysis with the remaining berths to determine whether there is a matching berth for the target inbound vessel to dock;

[0107] If there is a matching berth, obtain the target port waterway information and extract the target port waterway characteristics as intermediate nodes, use the matching berth area location characteristics as the end node, and use the real-time position of the target inbound vessel as the start node. Preset a cost function and use the A* algorithm to generate the initial guidance route for the target inbound vessel;

[0108] Obtain the real-time vessel navigation information in the target port waterway, extract the guidance routes of the vessels currently traveling on the waterway through the real-time vessel navigation information, and compare them with the initial guidance route to analyze whether there are overlapping routes;

[0109] If there are overlapping routes, extract the waterway position characteristics of the overlapping routes, analyze whether it is possible to change the route through the real-time vessel navigation information. If it is possible to change the route and travel, merge the changed route into the initial guidance route to generate the final docking guidance route;

[0110] If it is not possible to change the route and travel, use the initial guidance route as the final docking guidance route, extract the vessel navigation characteristics corresponding to the overlapping route, and calculate the following speed of the target inbound vessel when on the overlapping route to generate a vessel travel suggestion;

[0111] Based on the vessel information database, obtain the system account number and registered mobile phone number of the target inbound vessel, and send the final docking guidance route and the vessel travel suggestion to the system account number of the target inbound vessel and notify through the mobile phone number.

[0112] It should be noted that when a target incoming ship retrieves in the preset ship information database and finds a matching ship record, this indicates that the target ship has completed the system registration and filing and has been marked as an identified ship. At this time, relevant information is extracted from the incoming port upload data of the ship, and this data is compared with the ship intelligent identification information to detect whether there are differences between the two. If differences are detected, a notification is sent through the broadcast array to remind the target ship to re-register to update its registration information. If there are no differences, a docking guidance plan is formulated for the target ship by integrating the ship intelligent identification information and the incoming port upload data. Subsequently, the real-time ship docking information in the port area is obtained, including the status of the ships currently docked in the docking area and the data of the ships waiting to dock. According to this real-time information, the characteristics of the docked ships and the characteristics of the ships waiting to dock are extracted from the docking area to calculate the remaining berths in the target docking area and generate the remaining berth information. The docking requirements of the target ship are analyzed and matched with the remaining berths to determine whether there are suitable berths for the target ship to dock. If there are matching berths, the system obtains the channel information of the target port and extracts its channel characteristics, takes the position characteristics of the berth area as the target node, takes the current real-time position of the ship as the starting node, and uses the A* algorithm and the preset cost function to generate an initial docking guidance route. After generating the initial route, the real-time ship navigation information in the target port channel is further obtained, the guidance routes of other ships currently sailing in the channel are extracted, and compared with the initial guidance route of the target ship to analyze whether there are overlapping routes. If overlapping navigation routes are found, the channel position characteristics of the overlapping route are extracted, and it is analyzed through the real-time ship navigation information whether it is possible to change the course. If it is possible to change the course, the changed route is merged with the initial route to generate the final docking guidance route. If it is not possible to change the course, the original guidance route is maintained as the final route, and based on the ship navigation characteristics of the overlapping route, the following speed of the target ship in the overlapping route section is calculated and corresponding driving suggestions are generated. Finally, the system obtains the system account of the target incoming ship and the mobile phone number at the time of registration through the ship information database, sends the final docking guidance route and driving suggestions to the system account of the target ship, and pushes them to its registered mobile phone number in the form of a notification to ensure accurate information transmission and guide it to complete the port docking process.

[0113] Further, in a preferred embodiment of the present invention, if the target incoming ship is not an identified ship, a notification instruction is generated based on a preset response template and sent to the broadcast array to notify the target ship to conduct ship registration, and the real-time docking status data of the ship registration area is obtained, and a ship registration guidance plan is formulated and pushed, specifically including:

[0114] If there is no matching ship after the target incoming ship is retrieved from the preset ship information database, it means that the ship has not been registered and filed in the system. Then, obtain the ship intelligent identification information, and extract the identification features of the target non-identified ship through the ship intelligent identification information;

[0115] Based on the identification features of the target non-identified ship, generate a notification instruction in combination with the preset response template, and send the notification instruction to the broadcast array to notify the target non-identified ship;

[0116] Obtain the real-time docking status data of the ship registration area, formulate a ship registration guidance plan in combination with the identification features of the target non-identified ship, and push it to the target ship for offline filing and registration.

[0117] It should be noted that when the target incoming ship fails to find a matching ship record after being retrieved from the preset ship information database, it means that the ship has not completed the system registration and filing and belongs to a non-identified ship. In this case, first extract the ship intelligent identification information, and use this information to extract the unique identification features of the target ship from the target ship. These identification features may include the appearance features of the ship, size information, hull text features, etc. Based on the extracted identification features, generate corresponding notification instructions in combination with the preset response template. These instructions are usually used to remind unregistered ships to perform necessary filing operations. Subsequently, the generated notification instructions are sent to the target ship through the broadcast array to ensure that the ship can receive relevant notifications in a timely manner. At the same time, obtain the real-time docking status data of the ship registration area, analyze the berth distribution, current available berths and dynamic information of other docked ships in this area. Combine the identification features of the target non-identified ship to formulate a ship registration guidance plan. Finally, this guidance plan will be pushed to the target non-identified ship to guide it to go to the offline designated registration area to complete the filing process. By reminding to scan the code at the checkpoint for preliminary filing and obtain the guidance to enter the port before entering the port, and adopting manual-assisted offline processing until the ship completes the system filing, the efficiency of ship incoming port management is improved.

[0118] Figure 2 The flowchart of an automatic guidance method for fishing boats to enter and berth in the port provided by an embodiment of the present invention;

[0119] As Figure 2 shown, the present invention provides a flowchart of an automatic guidance method for fishing boats to enter and berth in the port, including:

[0120] S202, if there is a matching ship after the target incoming ship is retrieved from the preset ship information database, it means that the target incoming ship has been registered and filed in the system as an identified ship, then obtain the incoming port upload data of the target incoming ship;

[0121] S204. Check whether there are differences between the incoming port upload data of the target incoming ship and the ship intelligent identification information. If there are differences, use the broadcast array to notify the target ship and remind the target ship to re-register;

[0122] S206. If there are no differences, formulate a berthing guidance plan based on the ship intelligent identification information and the incoming port upload data of the target incoming ship, calculate the remaining berths in the berthing area, analyze the berthing requirements of the target incoming ship, and perform matching analysis with the remaining berths;

[0123] S208. If there are matching berths, use the position characteristics of the matching berth area as the termination node, use the real-time position of the target incoming ship as the starting node, and use the preset cost function to generate the initial guidance route of the target incoming ship using the A* algorithm;

[0124] S210. Extract the guidance routes of the ships currently traveling on the waterway and analyze whether there are overlapping routes. If there are overlapping routes, extract the waterway position characteristics of the overlapping routes and analyze whether a detour is possible. If a detour is possible, merge the detour route into the initial guidance route to generate the final berthing guidance route;

[0125] S212. If a detour is not possible, use the initial guidance route as the final berthing guidance route, extract the ship navigation characteristics corresponding to the overlapping route, and calculate the following speed of the target incoming ship when on the overlapping route to generate a ship travel suggestion.

[0126] It should be noted that when a target inbound ship finds a matching ship after retrieving in the preset ship information database, it indicates that the ship has completed the system registration and filing and belongs to an identified ship. In this case, first, obtain the inbound upload data of the target inbound ship and compare these data with the ship intelligent identification information. If a difference is found between the two, the system will send a notice to the target ship through the broadcast array to remind it to re-register to ensure data consistency in the system, avoid intentional misreporting of information by inbound ships, and improve the safety of port management. If the comparison result shows no difference, a berthing guidance plan will be formulated. By analyzing the berthing requirements of the target ship and the real-time berth conditions in the berthing area, calculate the remaining berths in the area and make a match. If a berth is successfully matched, the specific location of the matched berth will be used as the termination node, and the current location of the target ship will be used as the starting node, and an initial guidance route will be generated using the A* algorithm through a preset cost function. After generating the initial route, extract the guidance routes of other ships currently navigating in the current port channel and compare them with the initial route. If an overlapping part is found between the two routes, further analyze the channel location characteristics corresponding to the overlapping route to determine whether a detour is possible. If a detour is possible, fuse the detoured route with the initial route to generate the final berthing guidance route. If a detour is not possible, directly use the initial route as the final berthing guidance route, and at the same time extract the navigation characteristics of the ships on the overlapping route, and generate a driving suggestion for it by calculating the following speed of the target ship. Ensure that the ship can reach the berthing area efficiently and safely in a complex channel environment, improve port operation efficiency, reduce human errors, and ensure the safety and smoothness of ship entry into the port.

[0127] Figure 3 A fishing boat inbound automatic guidance management system 3 based on intelligent identification provided by an embodiment of the present invention, the system includes: a memory 31, a processor 32, the memory 31 contains a fishing boat inbound automatic guidance management method program based on intelligent identification, and when the fishing boat inbound automatic guidance management method program based on intelligent identification is executed by the processor 32, the following steps are implemented:

[0128] Obtain a map of the target port area, divide the target port into several sub-areas, conduct a functional analysis on each sub-area, define the functional attributes of each area of the target port based on the results of the functional analysis, and set up a ship inbound monitoring array;

[0129] Monitor the inbound ships through the set ship inbound monitoring array, monitor the real-time inbound ships and conduct ship movement tracking, and conduct intelligent identification on the target inbound ship based on edge computing to obtain ship intelligent identification information;

[0130] Retrieve in a preset ship information database according to the intelligent ship identification information, determine whether the target incoming ship is an identified ship, and guide the target incoming ship based on the judgment result;

[0131] If the target incoming ship is an identified ship, obtain the real-time ship docking area information, calculate the matching berth by combining the real-time incoming position of the target incoming ship and the ship identification information, and formulate a docking guidance plan to provide the docking route and speed;

[0132] If the target incoming ship is not an identified ship, generate a notification instruction based on a preset response template and send it to the broadcast array to notify the target ship to conduct ship registration, obtain the real-time docking status data of the ship registration area, and formulate a ship registration guidance plan for pushing.

[0133] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only 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 can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed with each other can be through some interfaces. The indirect coupling or communication connection of the devices or units can be electrical, mechanical, or other forms.

[0134] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units; they can be located in one place or 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.

[0135] In addition, each functional unit in the embodiments of the present invention can be all integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in a unit; the above integrated units can be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0136] Those of ordinary skill in the art will understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including those of the above method embodiments; and the foregoing storage medium includes: various media such as removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0137] Alternatively, if the above integrated units are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solutions of the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. And the foregoing storage medium includes: various media such as removable storage devices, ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0138] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for automatically guiding fishing boats into port based on intelligent identification, characterized in that: include: Obtain a map of the target port area, divide the target port into several sub-areas, perform functional analysis on each sub-area, define the functional attributes of each area of ​​the target port based on the functional analysis results, and set up a ship entry monitoring array; The ship entering the port is monitored through the set ship entry monitoring array, the real-time incoming ships are monitored and the ship movement is tracked. The target incoming ships are intelligently identified based on edge computing to obtain the ship intelligent identification information; Searching a preset ship information database according to the ship intelligent identification information, determining whether the target ship entering the port is an identified ship, and guiding the target ship entering the port based on the determination result; If the target incoming ship is an identified ship, the real-time ship berthing area information is obtained, the matching berth is calculated based on the real-time port position of the target incoming ship and the ship identification information, and a berthing guidance plan is formulated to provide the berthing route and speed; If the target ship entering the port is not an identified ship, a notification instruction is generated based on the preset response template and sent to the broadcast array to notify the target ship to register the ship, obtain the real-time docking status data of the ship registration area, and formulate a ship registration guidance plan for push; The process of monitoring incoming ships by setting a ship entry monitoring array, monitoring incoming ships in real time and tracking their movement specifically includes: Based on the ship entry monitoring array set in the ship entry area of ​​the target port, the incoming ships entering the target port in real time are monitored, and the target detection model is built using YOLOv5, and the real-time data collected by the camera array of the ship entry area is input into the target detection model for ship detection; The target detection model uses CSPDarkNet to build a backbone network, and extracts features from the real-time collected images of the input model through the built backbone network to obtain multi-level features of the real-time collected images to form an original feature map; The attention mechanism is introduced to calculate the attention scores of features at different levels and generate an attention feature map. The attention feature map is fused with the original feature map using the FPN and PAN structures. The target is detected by the fused feature map to obtain the target ship entering the port detection information. The target ship entering the port detection information is imported into the Kalman filter to estimate the motion state of the target ship entering the port to predict the motion state at the next moment, and a joint matching method is used to analyze whether the predicted motion state matches the motion state observation value at the next moment; If it does not match the motion state observation value at the next moment, the motion state observation value that matches the history is extracted as training data and imported into the Bayesian inference network for training to obtain the transient posterior distribution; Based on the acquired transient posterior distribution, the motion state before the mismatched motion state occurs is inferred in the Bayesian network to obtain the inferred motion state, and a predicted trajectory is generated according to the inferred motion state. The predicted trajectory is used to lock the target ship entering the port and obtain the tracking information of the target ship entering the port.

2. According to claim 1, a method for automatically guiding fishing boats into port based on intelligent identification is characterized in that: The method of obtaining a target port area map, dividing the target port into several sub-areas, performing functional analysis on each sub-area, defining functional attributes of each area of ​​the target port based on the functional analysis results, and setting a ship entry monitoring array specifically includes: Acquire a target port area map based on data retrieval, divide the target port into several sub-areas according to the target port area map, extract features from each sub-area to obtain regional environmental features of each sub-area, and obtain sub-area environmental feature information; Calculate the cosine similarity value between each sub-region according to the sub-region environmental feature information, merge the sub-regions whose cosine similarity value is greater than a preset threshold, and classify each sub-region to obtain merged sub-region information; Extracting environmental attribute features, architectural attribute features, and layout attribute features in each merged sub-region according to the target port area map, performing regional functional analysis through the attributes of each merged sub-region, and obtaining functional analysis information; According to the functional analysis information, a functional label is assigned to each merged sub-region, and a regional function is defined for each merged sub-region. Based on the regional kinetic energy definition result, the ship entry area of ​​the target port area is extracted, and a ship entry monitoring array is set.

3. The method for automatically guiding fishing boats entering a port based on intelligent identification according to claim 1 is characterized in that: The intelligent identification of the target ship entering the port based on edge computing to obtain the ship intelligent identification information specifically includes: Acquire tracking information of a target ship entering the port, and extract monitoring video stream data of the target ship entering the port within a unit time based on the tracking information of the target ship entering the port; An intelligent recognition model is constructed based on a dual-channel convolutional neural network, wherein the intelligent recognition model includes a first channel and a second channel, and the monitoring video stream data of the target ship entering the port within the unit time is input into the intelligent recognition model for analysis; The first channel is used to extract the appearance features and motion features of the ship from the input monitoring video stream data, and the second channel is used to extract the hull text features of the target incoming ship from the monitoring video stream data; The extracted ship appearance features and hull text features are input into the fully connected layer and the joint embedding method is used to fuse the features to obtain the fused features. Based on the fused features, the type, position, size and hull text semantics of the target ship entering the port are analyzed to obtain the ship intelligent recognition information.

4. According to the method of claim 1, the automatic guidance management method for fishing boats entering the port based on intelligent identification is characterized in that: The searching in a preset ship information database according to the ship intelligent identification information, determining whether the target ship entering the port is an identified ship, and guiding the target ship entering the port based on the determination result specifically includes: Acquire ship intelligent identification information, extract the MMSI number of the target ship entering the port according to the ship intelligent identification information, form a first search tag, and search in a preset ship information database based on the first search tag; If the first search tag has matching data in the preset ship information database, it means that the target port-entering ship is an identified ship of the current port, and the target port-entering ship is marked as an identified ship; If the first search tag does not have matching data in the preset ship information database, it means that the target ship entering the port is an unidentified ship of the current port, and the target ship entering the port is marked as a non-identified ship; If the MMSI number of the target ship entering the port cannot be extracted through the ship intelligent identification information, the appearance features and hull text semantic features of the target ship are extracted to generate a feature portrait of the target ship entering the port; Based on the fuzzy matching method, the characteristic portrait of the target ship entering the port is used to calculate the similarity with the ship data of each identified ship stored in the ship information database, and the identified ship data with the highest similarity to the characteristic portrait of the target ship entering the port is obtained; The similarity value of the identified ship data with the highest similarity to the characteristic portrait of the target ship entering the port is judged with the preset threshold. If it is greater than the preset threshold, it means that the target ship entering the port is an identified ship. If it is less than the preset threshold, it means that the target ship entering the port is a non-identified ship.

5. The method for automatically guiding fishing boats entering a port based on intelligent identification according to claim 1 is characterized in that: If the target incoming ship is an identified ship, the real-time ship berthing area information is obtained, the matching berth is calculated based on the real-time port position of the target incoming ship and the ship identification information, and a berthing guidance plan is formulated to provide the berthing route and speed, specifically including: If a matching ship exists after searching the preset ship information database for the target ship entering the port, it means that the target ship entering the port has been registered as an identified ship in the system, and the port upload data of the target ship entering the port is obtained; Check whether there is a discrepancy between the target ship's incoming upload data and the ship's intelligent identification information. If there is a discrepancy, use the broadcast array to notify the target ship and remind it to re-register; If there is no difference, a docking guidance plan is formulated based on the ship intelligent identification information and the target incoming ship's incoming port upload data to obtain real-time ship docking area information, which includes the docking status of ships in the docking area and data on ships to be docked; Extracting features of ships already docked in the docking area and features of ships to be docked based on the real-time ship docking area information, and calculating remaining berths in the target docking area according to the extracted features to obtain remaining berth information; Analyze the berthing demand of the target incoming ship based on the ship intelligent identification information and the incoming uploaded data of the target incoming ship, and perform matching analysis with the remaining berths to determine whether there is a matching berth for the target incoming ship to dock; If there is a matching berth, the target port channel information is obtained and the target port channel characteristics are extracted as the intermediate node, the matching berth area location characteristics are used as the end node, the real-time position of the target ship entering the port is used as the start node, and the preset cost function uses the A* algorithm to generate the initial guidance route for the target ship entering the port; Acquire real-time ship navigation information in the target port channel, extract the guidance route of the ship currently traveling on the channel through the real-time ship navigation information, compare it with the initial guidance route, and analyze whether there is an overlapping route; If there is an overlapping route, the channel position features of the overlapping route are extracted, and whether a diversion is possible is analyzed through the real-time ship navigation information. If a diversion is possible, the diversion route is merged into the initial guidance route to generate a final docking guidance route; If it is not possible to change the route, the initial guidance route is used as the final docking guidance route, and the ship navigation characteristics corresponding to the overlapping route are extracted, and the following speed of the target incoming ship on the overlapping route is calculated to generate ship navigation suggestions; Based on the ship information database, the system account and registered mobile phone number of the target ship entering the port are obtained, and the final docking guidance route and ship driving suggestions are sent to the system account of the target ship entering the port and notified via the mobile phone number.

6. The method for automatically guiding fishing boats entering a port based on intelligent identification according to claim 1 is characterized in that: If the target ship entering the port is not an identified ship, a notification instruction is generated based on a preset response template and sent to the broadcast array to notify the target ship to register the ship, obtain the real-time docking status data of the ship registration area, and formulate a ship registration guidance plan for push, which specifically includes: If no matching ship exists after searching the preset ship information database for the target ship entering the port, it means that the ship has not been registered in the system, then the ship intelligent identification information is obtained, and the identification features of the target non-identified ship are extracted through the ship intelligent identification information; Based on the identification features of the target non-identified ship, a notification instruction is generated in combination with a preset response template, and the notification instruction is sent to the broadcast array to notify the target non-identified ship; Obtain real-time docking status data in the ship registration area, formulate a ship registration guidance plan based on the identification characteristics of the target non-identified ship, and push it to the target ship for offline filing and registration.

7. An automatic guidance management system for fishing boats entering a port based on intelligent identification, characterized in that: The system comprises: a memory and a processor, wherein the memory contains a method program for automatically guiding a fishing vessel into a port based on intelligent identification, and when the method program for automatically guiding a fishing vessel into a port based on intelligent identification is executed by the processor, the following steps are implemented: Obtain a map of the target port area, divide the target port into several sub-areas, perform functional analysis on each sub-area, define the functional attributes of each area of ​​the target port based on the functional analysis results, and set up a ship entry monitoring array; The ship entering the port is monitored through the set ship entry monitoring array, the real-time incoming ships are monitored and the ship movement is tracked. The target incoming ships are intelligently identified based on edge computing to obtain the ship intelligent identification information; Searching a preset ship information database according to the ship intelligent identification information, determining whether the target ship entering the port is an identified ship, and guiding the target ship entering the port based on the determination result; If the target incoming ship is an identified ship, the real-time ship berthing area information is obtained, the matching berth is calculated based on the real-time port position of the target incoming ship and the ship identification information, and a berthing guidance plan is formulated to provide the berthing route and speed; If the target ship entering the port is not an identified ship, a notification instruction is generated based on the preset response template and sent to the broadcast array to notify the target ship to register the ship, obtain the real-time docking status data of the ship registration area, and formulate a ship registration guidance plan for push; The process of monitoring incoming ships by setting a ship entry monitoring array, monitoring incoming ships in real time and tracking their movement specifically includes: Based on the ship entry monitoring array set in the ship entry area of ​​the target port, the incoming ships entering the target port in real time are monitored, and the target detection model is built using YOLOv5, and the real-time data collected by the camera array of the ship entry area is input into the target detection model for ship detection; The target detection model uses CSPDarkNet to build a backbone network, and extracts features from the real-time collected images of the input model through the built backbone network to obtain multi-level features of the real-time collected images to form an original feature map; The attention mechanism is introduced to calculate the attention scores of features at different levels and generate an attention feature map. The attention feature map is fused with the original feature map using the FPN and PAN structures. The target is detected by the fused feature map to obtain the target ship entering the port detection information. The target ship entering the port detection information is imported into the Kalman filter to estimate the motion state of the target ship entering the port to predict the motion state at the next moment, and a joint matching method is used to analyze whether the predicted motion state matches the motion state observation value at the next moment; If it does not match the motion state observation value at the next moment, the motion state observation value that matches the history is extracted as training data and imported into the Bayesian inference network for training to obtain the transient posterior distribution; Based on the acquired transient posterior distribution, the motion state before the mismatched motion state occurs is inferred in the Bayesian network to obtain the inferred motion state, and a predicted trajectory is generated according to the inferred motion state. The predicted trajectory is used to lock the target ship entering the port and obtain the tracking information of the target ship entering the port.

Images