Goods flow statistical method and system for sensing driving preference based on transportation network

Abstract

The invention provides a cargo flow statistical method and system for sensing driving preference based on a transportation network, and the method comprises the following steps: obtaining the spatial geographic information of a target water area, and constructing a transportation network diagram of the target water area; obtaining ship lockage record data of a target water area, extracting lockage records of a target ship, and segmenting the lockage records into independent single-voyage lockage record sequences according to time continuity and spatial logic of adjacent lockage records; extracting all nodes in each single-voyage lockage record sequence as a lockage ship driving preference node set; taking nodes in the lockage ship driving preference node set as necessary anchor points, searching and complementing an optimal path between the adjacent necessary anchor points, splicing the optimal path and the necessary anchor points according to a time sequence, and generating a complete ship transportation voyage number route; and carrying out cargo transportation volume accumulation on transportation voyage times passing through the same node and edge according to cargo types, and carrying out statistics on cargo flow distribution of each leg and node in the target water area transportation network.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of waterway transportation, and particularly relates to a freight flow statistical method and system based on transportation network perception of driving preferences. BACKGROUND

[0002] In the research of shipping big data, most of the current researches are large-scale cross-country ocean transportation analysis based on automatic identification system (AIS) data or maritime report data, and the research on inland freight transportation is still relatively less. Although the International Convention for the Safety of Life at Sea (SOLAS) and the Technical Regulations for Inland Ship Inspection are mandatory to equip and continuously turn on AIS for ships at sea, the professional quality of inland ship crews is uneven, and some crews artificially turn off the equipment to avoid supervision or save energy, which makes it difficult to achieve real-time and full-coverage supervision of ship dynamics in inland waters. In recent years, the importance of inland shipping network freight transportation has gradually increased, and the granularity analysis of shipping network freight volume data obtained by investigation and port throughput statistics has been unable to meet the demand.

[0003] At present, for the analysis of inland ship transportation path and freight flow, the following methods are mainly used: 1. Taking AIS data as the object, the inland transportation is speculated by analyzing the ship dynamic data of coastal ports or international trade routes; 2. Relied on artificial statistics or a small amount of sensing equipment to estimate, this method is mainly used to estimate the preliminary information of inland ship transportation path and freight flow; 3. Concentrated on using search algorithm and subjective and objective weight method to estimate the transportation path, this method estimates the actual sailing route of the ship and the distribution of freight flow through mathematical model and algorithm optimization.

[0004] However, these existing technologies have many deficiencies. On the one hand, the speculation of the transportation path is not accurate enough, and the actual preferences of the driver and the actual sailing behavior of the ship are not fully considered, resulting in deviation between the speculation result and the actual situation; on the other hand, the freight flow statistics lack objective basis, and it is difficult to truly reflect the dynamic changes and distribution rules of the freight transportation in the inland shipping network. SUMMARY

[0005] The application provides a freight flow statistical method and system based on transportation network perception of driving preferences, which solves the problem of inaccurate inland ship transportation path speculation in the prior art, resulting in lack of objective basis for freight flow statistics, by mining the transportation space-time relationship of the ship lock records of the numerous traffic scheduling ships distributed in the inland river.

[0006] To solve the above technical problems, the application provides a freight flow statistical method based on transportation network perception of driving preferences, comprising the following steps: Step S1: Obtain the spatial geographic information of the target water area, and construct a transportation network graph of the target water area; Step S2: Obtain the ship lock record data of the target water area, extract the full lock record of the target ship according to the ship identification code, and divide the full lock record into several independent single voyage lock record sequences according to the time continuity, spatial adjacency relationship and heading logic of adjacent lock records. Step S3: Extract the starting point, ending point and passing nodes of each single voyage lock record sequence to form a set of lock ship driving preference nodes. Step S4: In the transportation network graph of the target water area, take the nodes in the set of lock ship driving preference nodes as the necessary anchor points, search and complete the shortest path between adjacent necessary anchor points as the optimal path, splice all necessary anchor points and optimal paths in time sequence to generate a complete ship transportation voyage route. Step S5: Map the complete ship transportation voyage route to the transportation network graph of the target water area, and accumulate the calculation of the cargo type and cargo volume of the transportation voyage passing through the same node and edge to count the cargo flow distribution of each section and node in the transportation network of the target water area.

[0007] Preferably, the step S1 of constructing the transportation network graph of the target water area comprises the following steps: Step S11: Obtain the remote sensing image data of the target water area, and extract the traffic scheduling ship lock, channel branching and intersection point, channel starting and ending point and port as the water area element based on the remote sensing image data; Step S12: Define the water area element as a node of the transportation network graph, and connect two nodes by the actual physical channel as the edge of the transportation network graph according to the geographical spatial distribution of each water area element; Step S13: Extract the channel center line based on the remote sensing image data, calculate the length of the actual physical channel as the weight of the edge, and construct an undirected weighted graph as the transportation network graph of the target water area.

[0008] Preferably, the step S2 of dividing the full lock record into several independent single voyage lock record sequences comprises the following steps: Step S21: Define each lock record of the target ship as a data point , and the data point is a 6-tuple including lock time, shipping scheduling lock name, voyage starting point, voyage ending point, cargo type and cargo volume. The full lock record of the target ship is sorted in chronological order to form a lock record sequence; Step S22: Initialize the current voyage recorded shipping scheduling lock name set to an empty set, traverse the sorted full lock record, and take the current traversed lock record point judging the voyage attribution; Step S23: judging whether the shipping scheduling lock name of already exists in , and judging whether the voyage starting point, the voyage ending point, the cargo type and the cargo volume of are all the same; if the shipping scheduling lock name of does not exist in , and the voyage starting point, the voyage ending point, the cargo type and the cargo volume of are all the same, it is determined that belongs to the current shipping network voyage, and the shipping scheduling lock name of is added to ; otherwise, it is determined that is a starting record point of a new shipping network voyage, the full-quantity lock record is segmented, and is emptied and the shipping scheduling lock name of is added to the emptied set; Step S24: repeating step S23 until the full-quantity lock record is traversed.

[0009] Preferably, the searching and completing the shortest path between adjacent mandatory anchor points as the optimal path in step S4 comprises the following steps: Step S41: selecting two nodes adjacent in time sequence in the lock ship driving preference node set as the starting anchor point and the target anchor point of the current search; Step S42: judging whether and exist directly connected edges in the transportation network graph of the target water area; if yes, selecting the directly connected edges as the path connecting and ; if no, searching the shortest path between and , and taking the searched shortest path as the completed path between and ; Step S43: splicing each path completed based on adjacent preference nodes to obtain the complete path of the target ship under the current voyage.

[0010] Preferably, the cumulative calculation of the cargo type and the cargo volume of the transportation voyage passing through the same node and edge in step S5 comprises the following steps: Step S51: initializing the cargo volume attribute value of each edge in the transportation network graph of the target water area as zero; Step S52: traversing each complete ship transportation voyage route, obtaining the cargo volume value corresponding to each complete ship transportation voyage route, and accumulating the cargo volume value into the cargo volume attribute value of each edge passed by the corresponding route.

[0011] The application also provides a cargo flow statistical system based on transportation network perception driving preference, which is realized based on the above-mentioned cargo flow statistical method based on transportation network perception driving preference. The network construction module: extracts water area elements based on remote sensing image data, and constructs a transportation network graph of the target water area. The voyage extraction module: extracts ship single-voyage lock record sequences based on ship lock record data, extracts the voyage starting point, voyage ending point and passing nodes in each single-voyage lock record sequence, and forms a lock ship driving preference node set. The path completion module: selects two nodes adjacent in time sequence in the lock ship driving preference node set as the starting anchor point and the target anchor point for searching, searches the shortest path between the starting anchor point and the target anchor point in the transportation network graph of the target water area, completes the missing passing points in the ship transportation single-voyage route, and generates a complete ship transportation voyage route. The cargo flow statistical module: extracts cargo volume information and cargo type information from the single-voyage lock record sequence, assigns the cargo volume information and cargo type information to the corresponding complete ship transportation voyage route, accumulatively calculates the cargo type and cargo volume passing through the same node and edge, and statistically calculates the cargo flow distribution of each section and node.

[0012] Preferably, the network construction module comprises an element extraction unit and a graph network generation unit. The element extraction unit: extracts traffic scheduling lock, channel branching and intersection point, channel starting and ending point and port as water area elements based on remote sensing image data. The graph network generation unit: defines the water area elements as nodes of the transportation network graph of the target water area, takes the connection relationship of the actual channel as the edge of the transportation network graph of the target water area according to the geographical spatial distribution of each water area element, takes the actual length of the channel as the weight of the edge, and constructs an undirected weighted graph as the transportation network graph of the target water area.

[0013] Preferably, the voyage extraction module comprises a sequence sorting unit, a voyage judgment unit, a voyage segmentation unit and a path splicing unit. The sequence sorting unit: based on ship lock record data, extracting full-quantity lock record of a target ship through a ship identification code, and sorting the full-quantity lock record of the target ship according to the order of lock time; The voyage judgment unit: traversing the sorted full-quantity lock record, judging whether adjacent data points belong to the same ship transportation voyage by comparing the information of adjacent data points; The voyage segmentation unit: aggregating data points belonging to the same ship transportation voyage, and segmenting data points not belonging to the same ship transportation voyage, to form a ship single-voyage lock record sequence; The path splicing unit: extracting the voyage starting point, shipping dispatching ship lock name and voyage ending point from the ship single-voyage lock record sequence, and splicing to form a ship transportation single-voyage route.

[0014] Preferably, the path completion module comprises a neighbor judgment unit, a path search unit and a path generation unit; The neighbor judgment unit: traversing adjacent node pairs in the set of lock ship driving preference nodes, judging whether each pair of adjacent nodes is a neighbor node of each other in the transportation network graph of the target water area; The path search unit: when the two nodes are not neighbor nodes of each other, taking the edge weight of the transportation network graph of the target water area as the cost, searching for the shortest path between the two nodes; The path generation unit: supplementing the searched way points between adjacent lock driving preference nodes to generate a complete ship transportation single-voyage route.

[0015] Preferably, the cargo flow statistics module comprises a path input unit, a cumulative calculation unit and a distribution output unit; The path input unit: mapping the complete ship transportation voyage route to the transportation network graph of the target water area; The cumulative calculation unit: making the nodes and edges passed by the complete ship transportation voyage route inherit the cargo type and cargo volume of the current voyage, and cumulatively calculating the cargo type and cargo volume passing through the same nodes and edges; The distribution output unit: outputting the cargo flow distribution of each section and node.

[0016] The beneficial effects of the present application at least include: 1. By obtaining ship lock record data, the actual behavior data of the ship is directly used, ensuring that the basic data for analysis has high objectivity and authenticity, and compared with the traditional method relying on artificial statistics or subjective weight, human error and subjective bias are reduced; 2. The continuity of time and spatial logical segmentation of adjacent lock records can accurately restore the actual sailing path of the ship, avoiding errors caused by incomplete or incorrect data. 3. By extracting the starting point, end point and key nodes of the voyage, and searching for the optimal path with these nodes as the necessary anchor points, the sailing route of the ship can be completely restored, not only considering the actual preference of the ship, but also ensuring the integrity of the path through the optimal path completion. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The method flowchart of the embodiment of the present application is shown. Figure 2 The water area element extracted based on the remote sensing image of a certain province in the embodiment of the present application is shown. Figure 3 The graph model based on a certain inland waterway transportation network in the embodiment of the present application is shown. Figure 4 The flowchart of extracting single voyage records of lock passing ships in the embodiment of the present application is shown. Figure 5 The flowchart of extracting lock passing ship voyage path in the embodiment of the present application is shown. Figure 6 The shortest path search principle based on lock driving preference in the embodiment of the present application is shown. Figure 7 The freight volume cumulative distribution diagram of a certain inland waterway in the embodiment of the present application is shown. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0019] As shown in the drawings, Figure 1 The embodiment of the present application faces inland shipping network, and based on the multi-source fusion of lock networking charging data and remote sensing image, proposes a search implementation of voyage path reconstruction based on driving preference constraint, and carries out freight flow statistics and traffic distribution for bottleneck section identification and shipping scheduling optimization. Specifically, a freight flow statistics method based on transportation network perception driving preference is provided, including the following steps: Step S1: Obtain the spatial geographic information of the target water area, and construct the transportation network graph of the target water area.

[0020] Specifically, to achieve network representation of inland waterways, this embodiment of the invention first extracts water features from the target water area based on remote sensing image data. Major water bodies are identified through high-resolution remote sensing imagery, and the main shipping networks, traffic control lock facilities, and major port nodes along the route are extracted. The image recognition results are then corrected and verified by combining existing water system vector data and shipping network information to ensure the accuracy of water area boundaries and node information.

[0021] The main water features extracted from remote sensing imagery include: traffic control locks, waterways, and ports. The attributes and extraction methods for these water features are explained in detail below: (1) Traffic scheduling locks: As cross-river structures, traffic scheduling locks are distributed in the shipping network. Their main function in inland waterway transportation is to improve the navigation environment, regulate water level differences, and connect the shipping network. This embodiment of the invention mainly extracts the geographical coordinates of the traffic scheduling locks, namely latitude and longitude, and collects ship passage data.

[0022] (2) Waterways: Waterways consist of winding rivers navigable by ships. They are the carriers of ship transportation and also the main framework of the inland waterway transportation network. To simplify the structure of the inland waterway transportation network, this paper only extracts the important locations in the spatial distribution of the shipping network that have actual coupling relationships, namely the starting and ending points of the waterways, and the points where the waterways branch and intersect. The starting and ending points of the waterways are the points where the waterways end within the target water area, the points where the waterways branch off are the points where a river splits, and the points where the waterways intersect are the points where two or more waterways meet. In this embodiment of the invention, the geographical coordinates of the starting and ending points of the waterways, and the points where they branch and intersect are mainly extracted, namely latitude and longitude.

[0023] (3) Port: The port is the place where transportation tasks are carried out, and it is responsible for loading, unloading and transshipment of goods. The embodiments of the present invention mainly extract the geographical coordinates of the port, that is, latitude and longitude.

[0024] Taking a certain province as an example, the water features extracted from the remote sensing images are as follows: Figure 2 As shown.

[0025] Step S2: Obtain the vessel passage record data of the locks throughout the target waters. Extract the full passage record of the target vessel based on the vessel identification code. Based on the temporal continuity, spatial adjacency, and heading logic of adjacent passage records, divide the full passage record into several independent single-voyage passage record sequences.

[0026] Step S3: Extract the voyage start point, voyage end point, and nodes passed through in each single voyage lock passage record sequence to form a set of lock passage vessel driving preference nodes.

[0027] Specifically, after extracting the water features, a graph model of the inland waterway transportation network needs to be constructed to map the real geographical features and complex transportation relationships to the network. The extracted features, including navigation locks, channel bifurcation and intersection points, channel start and end points, and the locations of important ports, are defined as nodes in the graph. For edge settings, the actual geographic spatial distribution of each feature needs to be considered, with real channels, navigation locks, and port connections used as edges in the graph.

[0028] The transportation network diagram of the target waterway constructed in this embodiment of the invention is an undirected weighted graph: ; In the formula, This represents a set of nodes, which include: shipping scheduling locks, waterway intersections and junctions, shipping network start and end points, and ports. Represents the set of edges, each edge Corresponding to a segment of the shipping network; This represents the set of edge weights, which are primarily determined by the length of the shipping network in that segment.

[0029] Right now: ; In the formula, For nodes and nodes The length of the shipping network between them is obtained by subtracting the shipping network mileage.

[0030] The resulting network adjacency matrix can be represented as: .

[0031] Taking the northern part of a certain province as an example, the inland waterway transportation network model is as follows: Figure 3 As shown.

[0032] Vessel passage data originates from the traffic control lock network toll database. Vessels passing through locks are required to report relevant static information to the local transportation department. Simultaneously, sensors and related reporting equipment systems distributed around the shipping control locks record the dynamic information of the reporting vessels. This includes: Maritime Mobile Service Identifier (MMSI), vessel identification number, upstream / downstream direction, name of the traffic control lock, reporting time, port of registry, vessel type, relevant static parameters of the vessel, type of cargo transported, voyage origin, voyage destination, and actual cargo tonnage. The shipping control lock name indicates the specific shipping control lock being passed through. Relevant static parameters of the vessel include overall length, overall width, gross tonnage, deadweight tonnage, and molded depth.

[0033] In areas rich in river resources, locks are typically widely distributed at the choke points where rivers or waterways connect. These locks record detailed information about passing vessels in real time. A series of vessel passage records form a traceable spatiotemporal data stream, making it possible to simulate and model inferences about inland waterway transportation and trace inland waterway shipping routes.

[0034] Ship lock passage data may contain a series of problems such as duplicate data, missing data, data anomalies, and data disorder due to human error or system failure. Therefore, before data analysis, it is necessary to clean the collected data, such as removing duplicate data, filling in missing data, and unifying data format to ensure data quality and experimental rigor.

[0035] Vessel lock passage data consists of information records that need to be entered each time a vessel passes through a lock in the current traffic scheduling schedule. Typically, a vessel will pass through at least one lock during a voyage and report voyage information. When a large number of these records are scattered across the shipping scheduling lock statistics, it is difficult to explore the voyage information and lock passage patterns of a vessel through simple observation. Therefore, in order to effectively monitor the complete historical record left by vessels throughout their voyages, it is necessary to extract each voyage from a large vessel lock passage database.

[0036] To address the characteristics of vessel lock passage data, this invention presents a complete voyage extraction process. The extraction process for a single voyage record of a vessel passing through the lock is as follows: Figure 4 As shown, the process involves retrieving all lock passage records for a single vessel from the lock passage database, then determining, in chronological order, whether information such as the lock name, origin and destination, passage direction, type of transported cargo, and cargo volume of a particular record belongs to the same voyage. Finally, records from different voyages are segmented and stored in chronological order.

[0037] When retrieving records for a single voyage, the following two points are worth noting: 1. When a ship's journey is short, only a single shipping control lock may record the ship's passage. Such voyages are usually fixed transport voyages between two regions, and there are multiple lock records of the same ship passing through the same shipping control lock within a short period of time. 2. When a ship's journey is long, the same ship may have multiple lock passage records from different shipping locks within a certain period of time. These records together constitute the journey information for this voyage.

[0038] The specific steps are as follows: Step S31: Extract data from the ship lock passage database using MMSI. All gate passage records, a single data point Expressed as a 6-tuple, this 6-tuple is the first record of the ship passing through the lock. The information includes the lock passage time at each point, the name of the lock in the shipping scheduling system, the starting point of the voyage, the ending point of the voyage, the type of cargo, and the composition of the cargo volume. , express The number of data points recorded in the gate passage record. The sequence of gate passage records consists of data points arranged in chronological order of gate passage time, represented as... ; Step S32: Segment the extracted single-ship lock passage record sequence into single-voyage segments. For In the records of the same voyage, all data points arranged in chronological order should have identical information except for the names of the traffic control locks. Therefore, it is necessary to iterate through... By comparing the information of adjacent data points in the lock passage record sequence, it can be determined whether each sub-sequence belongs to the same voyage. The determination is based on... as follows: ; In the formula, This is the set of lock names for the current voyage that have been recorded for shipping scheduling. , , , , They are respectively The shipping schedule includes the name of the lock, the starting point of the voyage, the ending point of the voyage, the type of cargo, and the cargo volume. , , , They are respectively The voyage's origin, destination, type of cargo, and cargo volume; When judged as hour, This belongs to the current shipping network voyage, and Add the name of the shipping scheduling lock to the list ;otherwise To establish a new shipping network voyage starting point, the entire lock passage record is segmented and cleared. And The names of the shipping locks used for scheduling are added to the cleared set; Step S33: Repeat step S32 until... The sequence has been segmented.

[0039] The above steps can be used to segment different voyages of the same vessel.

[0040] For the segmented single-voyage sequences, further information extraction is required to form single-voyage paths. The segmented single-voyage sequences record the vessel's lock-passing declaration information for that voyage in chronological order. All data points within a single voyage sequence are identical except for the lock-passing time and the name of the traffic control lock. For example... Figure 5 As shown, by extracting the starting point, the name of the shipping control lock, and the destination, a single voyage sequence can be pieced together into a complete transportation path for that voyage, such as "starting point → traffic control lock 1 → traffic control lock 2 → … → destination". The pieced-together path passes through at least one traffic control lock.

[0041] The single-voyage route of a ship records the originating port, the destination port, and the traffic control locks it passes through, but it lacks specific channel indications. Therefore, it is necessary to supplement the single-voyage route with factors actually considered by the driver. This invention proposes a shortest path search model for inland waterway transportation networks based on lock-passing driving preferences. It uses a constructed transportation network graph G of the target waterway as the basis, and the points passed through in the existing single-voyage route as the lock-passing driving preference node set, employing a heuristic search. The algorithm searches for and completes the shortest path between preferred nodes.

[0042] Step S4: In the transportation network map of the target waters, take the nodes in the set of navigation preference nodes of the lock-passing vessels as the necessary anchor points, search and complete the shortest path between adjacent necessary anchor points as the optimal path, and splice all necessary anchor points and optimal paths in time order to generate a complete ship transportation voyage route.

[0043] Specifically, such as Figure 6 As shown, The algorithm primarily searches for the optimal path between preferred nodes, with the weight of the optimal path determined by the weight of the transportation network graph G in the target waterway. If two preferred nodes are neighbors in G, no additional waypoints are needed, and the two nodes are directly connected. If two preferred nodes are not directly connected in G, the optimal path is searched based on the edge weights in G, and the found optimal path is then inserted between the two nodes to form a complete navigation path. The state transition function of the algorithm is defined as: ; ; In the formula, For the evaluation function, it represents the process via node n The estimated total cost of reaching the destination; From the previous preference node to the current node n The cumulative actual cost; This represents the estimated cost from the current node to the next preferred node.

[0044] The process involves traversing the set of lock-crossing driving preference nodes, supplementing the waypoints between each pair of nodes, until the destination is reached. Through this process, the transportation route for a single voyage can be effectively represented in G. Considering the path finding based on lock-crossing driving preferences can greatly reconstruct the actual navigation route taken by the driver at that time, providing strong support for the statistics of cargo flow in various shipping networks.

[0045] Step S5: Map the complete shipping route to the transportation network graph of the target waters, accumulate the cargo types and cargo volumes of shipping voyages that pass through the same nodes and edges, and statistically analyze the cargo flow distribution of each segment and node in the transportation network of the target waters.

[0046] Specifically, based on complete voyage routes, each voyage carries the same static characteristics such as cargo type and deadweight tonnage. Therefore, by inputting the complete voyage route into the transportation network diagram of the target waterway, the nodes and edges along the route inherit these characteristics. When different voyages are input into the transportation network diagram of the target waterway, these static characteristics must be cumulatively calculated when passing through the same nodes and edges. After inputting all voyage data into the transportation network diagram of the target waterway, the cargo distribution of each segment and node can be understood by viewing the cargo type information and cargo volume statistics of each edge.

[0047] The specific implementation includes the following steps: Step S51: Initialize the freight volume attribute value of each edge in the transportation network graph of the target water area to zero.

[0048] Step S52: Traverse each complete shipping route, obtain the cargo volume value corresponding to each complete shipping route, and add the cargo volume value to the cargo volume attribute value of each edge traversed by the corresponding route.

[0049] Taking the waterways in the northern part of a certain province as an example, the province's inland waterway network has a total length of over 24,000 kilometers, of which more than 2,400 kilometers are at or above the third level. It also boasts 101 navigation locks supporting the efficient operation of the waterway transportation network. In one region alone, 34 interconnected navigation locks record over 80% of the cargo flow in the entire region's transportation network.

[0050] By collecting coal transport data from vessels passing through the locks in this region in January 2024, a total of 28,735 records were collected, including 5,654 upstream records and 23,081 downstream records. The upstream coal transport voyages were input into the region's inland waterway transport network, and the cumulative cargo volume distribution across each shipping network is shown below. Figure 7 As shown.

[0051] This invention also provides a freight flow statistics system based on transportation network-aware driving preferences. The system includes: a network construction module, a voyage extraction module, a route completion module, and a freight flow statistics module.

[0052] The network construction module extracts water features from remote sensing image data and constructs a transportation network map of the target water area. This module includes: a feature extraction unit and a graph network generation unit.

[0053] Feature extraction unit: Based on remote sensing image data, traffic control locks, channel bifurcation and intersection points, channel start and end points, and ports are extracted as water features.

[0054] Graph network generation unit: Defines water features as nodes in the transportation network graph of the target water area; uses the connectivity of waterways as edges in the transportation network graph of the target water area based on the geographic spatial distribution of each water feature; and constructs an undirected weighted graph as the transportation network graph of the target water area using the actual length of the waterway as the weight of the edge.

[0055] The voyage extraction module extracts the route for a single voyage of a ship based on the ship's lock passage data and obtains a set of lock passage driving preference nodes. This module includes: a sequence sorting unit, a voyage judgment unit, a voyage segmentation unit, and a path splicing unit.

[0056] Sequence sorting unit: Based on the ship lock passage data, the full lock passage records of the target ship are extracted through the ship identification code, and the full lock passage records of the target ship are sorted according to the order of the lock passage time.

[0057] Voyage Judgment Unit: Traverse all sorted lock passage records and determine whether adjacent data points belong to the same shipping network voyage by comparing the information of adjacent data points.

[0058] Voyage segmentation unit: Data points belonging to the same voyage are aggregated, while data points not belonging to the same voyage are segmented to form a single voyage sequence of ship transportation.

[0059] Path splicing unit: Extract the voyage start point, traffic control lock name and voyage end point from the single voyage sequence of shipping and splice them to form the single voyage route of shipping.

[0060] The path completion module selects two adjacent nodes in the lock passage driving preference node set in chronological order as the starting and target anchor points for the search. It then searches the transportation network map of the target waterway for the shortest path between the starting and target anchor points, completing the missing waypoints in the single voyage route of the vessel and generating a complete transportation voyage route. This module includes: a neighbor determination unit, a path search unit, and a path generation unit.

[0061] Neighbor determination unit: Traverse the adjacent node pairs in the set of gate driving preference nodes and determine whether each pair of adjacent nodes is a neighbor node in the transportation network graph of the target water area.

[0062] Path search unit: When two nodes are not neighbors, the shortest path between the two nodes is searched at the cost of the edge weights of the transportation network graph of the target water area.

[0063] Path generation unit: The searched waypoints are added between adjacent lock passage driving preference nodes to generate a complete ship transport voyage route.

[0064] The cargo flow statistics module extracts cargo volume and cargo type information from the single-voyage lock passage record sequence, assigns the cargo volume and cargo type information to the corresponding complete ship transport voyage route, accumulates and calculates the cargo type and cargo volume for those passing through the same nodes and edges, and statistically analyzes the cargo flow distribution of each segment and node. This module includes: a path input unit, an accumulation calculation unit, and a distribution output unit.

[0065] Path Input Unit: Maps the complete shipping voyage route to the transport network map of the target waters.

[0066] Cumulative Calculation Unit: This unit allows nodes and edges along a complete shipping voyage to inherit the cargo types and cargo volumes of the current voyage, and performs cumulative calculations on the cargo types and cargo volumes that pass through the same nodes and edges.

[0067] Distributed output unit: Outputs the cargo flow distribution of each flight segment and node.

[0068] This invention employs recorded data of vessel lock passage for spatiotemporal modeling, which to some extent fills the research gap in the analysis of inland waterway vessel transport routes and cargo flow distribution in the hinterland. The proposed shortest path modeling method based on driver preferences can effectively reconstruct the actual transport routes of vessels based on objective factual data, better reflecting actual operational patterns and improving the authenticity and reliability of cargo flow statistics. By combining remote sensing imagery with waterway transport network topology modeling, multi-source heterogeneous elements such as ports, shipping networks, and shipping scheduling locks are uniformly mapped into the network model, achieving a high-precision digital representation of complex inland waterway systems and contributing to the formation of a computationally computable inland waterway cargo transport network. Through the input and cumulative calculation of multiple inland waterway vessel transport voyage routes, this invention can achieve dynamic and fine-grained quantitative results in the statistics of vessel transport cargo types and volumes.

[0069] The methods and systems provided by the embodiments of the present invention can effectively achieve accurate statistics on cargo flow in inland waterway shipping networks, providing data support for the identification of regional bottleneck waterways and the optimization of traffic organization.

[0070] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0071] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A freight flow statistics method based on perceived driving preferences in a transportation network, characterized in that, Includes the following steps: Step S1: Obtain spatial geographic information of the target water area and construct a transportation network map of the target water area; Step S2: Obtain the vessel lock passage record data of the entire lock in the target water area, extract the full lock passage record of the target vessel according to the vessel identification code, and divide the full lock passage record into several independent single-voyage lock passage record sequences according to the temporal continuity, spatial adjacency relationship and heading logic of adjacent lock passage records. Step S3: Extract the voyage start point, voyage end point, and nodes passed through in each single voyage lock passage record sequence to form a set of lock passage vessel driving preference nodes; Step S4: In the transportation network map of the target water area, take the nodes in the set of lock-passing vessel driving preference nodes as the necessary anchor points, search and complete the shortest path between adjacent necessary anchor points as the optimal path, and splice all necessary anchor points and the optimal path in time order to generate a complete vessel transportation voyage route. Step S5: Map the complete ship transport voyage route to the transport network map of the target water area, accumulate and calculate the cargo types and cargo volumes of transport voyages that pass through the same nodes and edges, and statistically analyze the cargo flow distribution of each segment and node in the transport network of the target water area.

2. The freight flow statistics method based on perceived driving preferences in a transportation network according to claim 1, characterized in that: The construction of the transportation network map of the target waterway in step S1 includes the following steps: Step S11: Obtain remote sensing image data of the target water area, and extract traffic control locks, channel bifurcation and intersection points, channel start and end points and ports of the target water area as water area elements based on the remote sensing image data. Step S12: Define the water features as nodes of the transportation network graph, and according to the geographic spatial distribution of each water feature, use the actual physical waterway connecting two nodes as the edge of the transportation network graph. Step S13: Extract the centerline of the waterway based on remote sensing image data, calculate the length of the actual physical waterway as the weight of the edge, and construct an undirected weighted graph as the transportation network graph of the target water area.

3. The freight flow statistics method based on perceived driving preferences in a transportation network according to claim 1, characterized in that: Step S2 involves dividing the full lock passage record into several independent single-voyage lock passage record sequences, including the following steps: Step S21: Define each lock passage record of the target vessel as a data point. data points The lock passage records of the target vessel are sorted in chronological order according to the lock passage time, including the lock passage time, lock name, voyage start point, voyage end point, cargo type and cargo volume, forming a lock passage record sequence. Step S22: Initialize the set of lock names recorded for the current voyage's shipping scheduling. Given an empty set, iterate through all sorted gate passage records, and then select the gate passage record currently being traversed. Determine the voyage affiliation; Step S23: Determine Does the name of the shipping control lock already exist? In the middle, simultaneously judge and Are the voyage origin, voyage destination, cargo type, and cargo volume all the same? If so... The name of the shipping control lock does not exist. ,and and If the voyage origin, voyage destination, cargo type, and cargo volume are all the same, then it is determined that... This belongs to the current shipping network voyage and will Add the name of the shipping scheduling lock to the list Otherwise, determine To mark the starting point of a new shipping network voyage, the entire lock passage record is segmented and cleared. And The names of the shipping locks used for scheduling are added to the cleared set; Step S24: Repeat step S23 until all the gate passage records have been traversed.

4. The freight flow statistics method based on perceived driving preferences in a transportation network according to claim 1, characterized in that: Step S4, which involves searching for and completing the shortest path between adjacent necessary anchor points as the optimal path, includes the following steps: Step S41: Select two adjacent nodes in the set of lock-passing vessel driving preference nodes according to time order as the starting anchor point for the current search. and target anchor point ; Step S42: Determine and Does the transportation network diagram of the target water area contain directly connected edges? If so, select the directly connected edges as the connection. and The path; if it does not exist, then search. and The shortest path between, the shortest path found is used as and Complete the path between them; Step S43: Piece together the paths completed based on adjacent preference nodes to obtain the complete path of the target ship in the current voyage.

5. A freight flow statistics method based on perceived driving preferences in a transportation network, as described in claim 1, is characterized in that... Step S5, which involves cumulatively calculating the types and volumes of goods transported on transport voyages that pass through the same nodes and edges, includes the following steps: Step S51: Initialize the freight volume attribute value of each edge in the transportation network graph of the target water area to zero; Step S52: Traverse each complete shipping route, obtain the cargo volume value corresponding to each complete shipping route, and add the cargo volume value to the cargo volume attribute value of each edge traversed by the corresponding route.

6. A freight flow statistics system based on transportation network-aware driving preferences, implemented based on the freight flow statistics method based on transportation network-aware driving preferences as described in any one of claims 1-5, characterized in that, The system includes: a network construction module, a voyage extraction module, a route completion module, and a cargo flow statistics module; The network construction module extracts water features based on remote sensing image data and constructs a transportation network map of the target water area. The voyage extraction module extracts a single voyage lock passage record sequence based on the ship lock passage record data, and extracts the voyage start point, voyage end point and the nodes passed through in each single voyage lock passage record sequence to form a set of lock passage ship driving preference nodes; The path completion module selects two adjacent nodes in the set of navigation preference nodes for passing through the lock as the starting anchor point and the target anchor point for the search. It searches for the shortest path between the starting anchor point and the target anchor point in the transportation network map of the target water area, completes the missing waypoints in the single voyage route of the ship transportation, and generates a complete ship transportation voyage route. The cargo flow statistics module extracts cargo volume and cargo type information from the single voyage lock passage record sequence, assigns the cargo volume and cargo type information to the corresponding complete ship transport voyage route, accumulates and calculates the cargo type and cargo volume passing through the same nodes and edges, and statistically analyzes the cargo flow distribution of each voyage segment and node.

7. A freight flow statistics system based on perceived driving preferences in a transportation network, as described in claim 6, characterized in that: The network construction module includes: an element extraction unit and a graph network generation unit; The element extraction unit extracts traffic control locks, channel bifurcation and intersection points, channel start and end points, and ports as waterway elements based on remote sensing image data. The graph network generation unit: defines the water features as nodes of the transportation network graph of the target water area; based on the geographic spatial distribution of each water feature, uses the actual waterway connections as edges of the transportation network graph of the target water area; and uses the actual length of the waterway as the weight of the edges to construct an undirected weighted graph as the transportation network graph of the target water area.

8. A freight flow statistics system based on perceived driving preferences in a transportation network, as described in claim 6, characterized in that: The voyage extraction module includes: a sequence sorting unit, a voyage judgment unit, a voyage segmentation unit, and a path splicing unit; The sequence sorting unit: Based on the ship lock passage record data, it extracts the full lock passage records of the target ship through the ship identification code, and sorts the full lock passage records of the target ship according to the order of the lock passage time; The voyage determination unit: traverses the sorted full lock passage records and determines whether adjacent data points belong to the same ship transport voyage by comparing the information of adjacent data points; The voyage segmentation unit aggregates data points belonging to the same ship voyage and segments data points that do not belong to the same ship voyage to form a single voyage lock passage record sequence. The path splicing unit extracts the voyage start point, shipping scheduling lock name, and voyage end point from the single voyage lock passage record sequence of the ship, and splices them to form a single voyage route for ship transportation.

9. A freight flow statistics system based on perceived driving preferences in a transportation network, as described in claim 6, is characterized in that... The path completion module includes: a neighbor judgment unit, a path search unit, and a path generation unit; The neighbor determination unit: traverses the adjacent node pairs in the set of navigation preference nodes for vessels passing through the lock, and determines whether each pair of adjacent nodes is a neighbor node in the transportation network diagram of the target water area; The path search unit: when two nodes are not neighbors, it searches for the shortest path between the two nodes at the cost of the edge weights of the transportation network graph of the target water area. The path generation unit: fills the searched waypoints between adjacent lock passage driving preference nodes to generate a complete single voyage route for ship transportation.

10. A freight flow statistics system based on perceived driving preferences in a transportation network, as described in claim 6, is characterized in that... The cargo flow statistics module includes: a path input unit, a cumulative calculation unit, and a distributed output unit; The path input unit maps the complete ship transport voyage route to the transport network map of the target waterway; The cumulative calculation unit: enables the nodes and edges along the complete ship transport voyage route to inherit the cargo types and cargo volumes of the current voyage, and performs cumulative calculations on the cargo types and cargo volumes that pass through the same nodes and edges; The distributed output unit outputs the cargo flow distribution of each flight segment and node.

Images