Harbor dangerous cargo safety supervision method

Abstract

The invention relates to a harbor dangerous cargo safety supervision method, and relates to the technical field of dangerous cargo management, a state acquisition device and a radio frequency positioning tag are arranged in a dangerous cargo container, and the state acquisition device is used for acquiring internal state information of the dangerous cargo container; according to the port storage positions of the multiple dangerous cargo containers and the environment information of the current period, a safety supervision patrol path is generated; the patrol trolley is controlled to patrol based on the safety supervision patrol path, in the patrol process, internal state information, collected by the state collection device, of the dangerous cargo container is received, position information of the dangerous cargo container is obtained based on the radio frequency positioning tag, and an image of the dangerous cargo container is collected; dangerous cargo safety supervision is carried out based on the internal state information, the image and the position information of the dangerous cargo container, and the method has the advantage of improving the quality of harbor dangerous cargo safety supervision.

Description

Technical Field

[0001] The present invention relates to the technical field of dangerous goods management, and in particular to a method for safely supervising dangerous goods in ports. Background Art

[0002] Ports are important distribution centers for trade goods, with a wide variety of goods frequently transshipped daily, including dangerous goods. Dangerous goods are substances and items with hazardous properties such as explosiveness, flammability, toxicity, infection, corrosion, and radioactivity, which can easily cause personal injury, property damage, or environmental pollution during transportation, storage, production, operation, use, and disposal, and therefore require special protective measures. Water transportation, with its high economic efficiency and large capacity, is the most commonly used mode of transport for dangerous goods. Port terminals, as hubs connecting water and other modes of transportation, carry a large volume of dangerous goods. With the development of port transportation, densely populated areas such as commercial and residential areas have gradually emerged in the surrounding areas. A leak, fire, or explosion of hazardous chemicals transported by ports could cause serious harm to nearby densely populated areas. Therefore, the safety and regulation of dangerous goods in ports is crucial and cannot be ignored.

[0003] In the existing technology, the safety supervision of dangerous goods in ports is mainly carried out manually, which is inefficient and has low real-time performance.

[0004] Therefore, it is necessary to provide a method for safety supervision of dangerous goods in ports to improve the quality of safety supervision of dangerous goods in ports. Summary of the Invention

[0005] The present invention provides a method for safely supervising dangerous goods in a port, comprising: arranging a state acquisition device and a radio frequency positioning tag in a dangerous goods container, wherein the state acquisition device is used to collect internal state information of the dangerous goods container; generating a safety supervision inspection path according to the port storage locations of multiple dangerous goods containers and environmental information of the current cycle; controlling an inspection vehicle to perform inspections based on the safety supervision inspection path, and during the inspection, receiving the internal state information of the dangerous goods container collected by the state acquisition device, obtaining the position information of the dangerous goods container based on the radio frequency positioning tag, and collecting an image of the dangerous goods container; and performing safety supervision of the dangerous goods based on the internal state information, image and position information of the dangerous goods container.

[0006] Furthermore, a status collection device is set in the dangerous goods container, including: for each type of dangerous goods, determining multiple candidate status factors corresponding to the dangerous goods type, obtaining a first test data set of the dangerous goods type, determining the correlation coefficient between any two candidate status factors corresponding to the dangerous goods type based on the first test data set of the dangerous goods type, and determining the status factor corresponding to the dangerous goods type based on the correlation coefficient between any two candidate status factors corresponding to the dangerous goods type; based on the dangerous goods type corresponding to the dangerous goods container and the status factor corresponding to each dangerous goods type, a status collection device is set in the dangerous goods container.

[0007] Furthermore, a safety supervision inspection path is generated based on the port storage locations of multiple dangerous goods containers and the environmental information of the current period, including: constructing a dangerous goods container safety supervision inspection network based on the port storage locations of multiple dangerous goods containers, wherein the dangerous goods container safety supervision inspection network includes a starting point, an end point and an intermediate node, and one intermediate node represents one dangerous goods container; determining the risk value of each intermediate node based on the environmental information of the current period; and generating a safety supervision inspection path based on the dangerous goods container safety supervision inspection network and the risk value of each intermediate node through an improved ant colony algorithm.

[0008] Furthermore, based on the environmental information of the current period, the risk value of each intermediate node is determined, including: for each type of dangerous goods, a second test data set of the dangerous goods type is obtained, based on the second test data set of the dangerous goods type, the correlation coefficient between each environmental factor and the risk value of the dangerous goods type is determined, based on the correlation coefficient between each environmental factor and the risk value of the dangerous goods type, the environmental factors corresponding to the dangerous goods type are determined, and based on the environmental factors corresponding to the dangerous goods type, a risk value prediction model corresponding to the dangerous goods type is established; for each intermediate node, the initial risk value of the intermediate node is determined based on the environmental information of the current period through the risk value prediction model corresponding to the dangerous goods container, and the risk value of the intermediate node is iteratively corrected based on the initial risk values ​​of adjacent intermediate nodes.

[0009] Furthermore, the heuristic information function and pheromone update function of the improved ant colony algorithm are related to the risk value of each intermediate node and the travel distance between any two intermediate nodes.

[0010] Furthermore, obtaining the location information of the dangerous goods container based on the radio frequency positioning tags includes: setting a plurality of auxiliary radio frequency positioning tags in the storage area of ​​the dangerous goods container; determining at least two target auxiliary radio frequency positioning tags from the plurality of auxiliary radio frequency positioning tags based on the real-time position of the patrol vehicle and the setting positions of the plurality of auxiliary radio frequency positioning tags; obtaining the location information of the dangerous goods container based on the at least two target auxiliary radio frequency positioning tags and the radio frequency positioning tag in the dangerous goods container.

[0011] Furthermore, based on the real-time position of the patrol vehicle and the setting positions of multiple auxiliary RF positioning tags, at least two target auxiliary RF positioning tags are determined from the multiple auxiliary RF positioning tags, including: for each auxiliary RF positioning tag, based on the real-time position of the patrol vehicle and the setting position of the auxiliary RF positioning tag, calculating the real-time distance between the patrol vehicle and the auxiliary RF positioning tag; based on the real-time distance between the patrol vehicle and each auxiliary RF positioning tag, determining multiple candidate auxiliary RF positioning tags from the multiple auxiliary RF positioning tags; for each candidate auxiliary RF positioning tag, based on the port storage positions of multiple dangerous goods containers, determining the port storage position of the dangerous goods container on the signal transmission path between the candidate auxiliary RF positioning tag and the RF receiver on the patrol vehicle, and determining the signal attenuation coefficient of the candidate auxiliary RF positioning tag based on the dangerous goods information of the dangerous goods container on the signal transmission path between the candidate auxiliary RF positioning tag and the RF receiver on the patrol vehicle; based on the signal attenuation coefficient of each candidate auxiliary RF positioning tag, determining at least two target auxiliary RF positioning tags from the multiple candidate auxiliary RF positioning tags.

[0012] Furthermore, based on at least two target-assisted radio frequency positioning tags and the radio frequency positioning tag in the dangerous goods container, the location information of the dangerous goods container is obtained, including: for each target-assisted radio frequency positioning tag, based on the signal strength of the target-assisted radio frequency positioning tag received by the radio frequency receiver on the inspection vehicle, determining the distance between the target-assisted radio frequency positioning tag and the radio frequency receiver on the inspection vehicle; based on the signal strength of the radio frequency positioning tag in the dangerous goods container received by the radio frequency receiver on the inspection vehicle, determining the distance between the radio frequency positioning tag in the dangerous goods container and the radio frequency receiver on the inspection vehicle; based on the distance between each target-assisted radio frequency positioning tag and the radio frequency receiver on the inspection vehicle and the distance between the radio frequency positioning tag in the dangerous goods container and the radio frequency receiver on the inspection vehicle, generating the location information of the dangerous goods container.

[0013] Furthermore, based on the internal status information, images and location information of dangerous goods containers, dangerous goods safety supervision is carried out, including: internal stability supervision of dangerous goods containers based on the internal status information of dangerous goods containers; container damage supervision and container isolation distance supervision of dangerous goods containers based on the images of dangerous goods containers; and stacking position supervision of dangerous goods containers based on the location information of dangerous goods containers.

[0014] Furthermore, based on the image of the dangerous goods container, container damage monitoring and container isolation distance monitoring of the dangerous goods container are carried out, including: for each adjacent dangerous goods container of the dangerous goods container, determining the isolation distance between the dangerous goods container and the adjacent dangerous goods container based on the image of the dangerous goods container, determining the isolation distance threshold based on the dangerous goods type of the dangerous goods container and the dangerous goods type of the adjacent dangerous goods container, and conducting container isolation distance monitoring based on the isolation distance between the dangerous goods container and the adjacent dangerous goods container and the isolation distance threshold; determining the surface flatness of the dangerous goods container based on the image of the dangerous goods container, and conducting container damage monitoring of the dangerous goods container based on the surface flatness of the dangerous goods container.

[0015] Compared with the existing technology, the method for safe supervision of dangerous goods in ports provided in this specification has at least the following beneficial effects:

[0016] 1. The status acquisition device can monitor key status information such as temperature, humidity, and pressure inside the container in real time. Once an abnormality is detected, such as excessive temperature or humidity, which may cause the risk of dangerous goods leakage or explosion, the device will immediately issue an early warning, allowing relevant personnel to take quick measures to prevent potential safety accidents.

[0017] Radio frequency positioning tags can accurately record the location information of each container, allowing supervisors to quickly locate any container requiring attention, greatly improving the accuracy and efficiency of supervision. Status acquisition equipment can continuously record the internal status data of the container. This data can be used for subsequent analysis and evaluation, helping supervisors better understand the storage conditions of hazardous materials and formulate more scientific and reasonable safety management measures. The image information collected by the inspection vehicle during the inspection process can be used to monitor the external condition of the container, such as whether there are signs of damage or leakage. Combined with image recognition technology, it can achieve automatic identification and judgment of the container status, further improving the intelligent level of safety management. Based on the port storage location of multiple hazardous goods containers and the current cycle of environmental information, a safety supervision inspection route is generated to ensure that the inspection vehicle conducts inspections along the optimal route, thus saving time and resources. In the event of a safety accident or emergency, supervisors can quickly obtain information at the scene of the accident through radio frequency positioning tags and status acquisition equipment, including the location of the accident container, the internal condition, and the possible degree of danger, so as to make more accurate emergency response decisions and reduce the losses caused by the accident.

[0018] 2. The improved ant colony algorithm can comprehensively consider the storage locations of multiple dangerous goods containers and the environmental information of the current cycle, such as weather and traffic conditions, to generate a more efficient and reasonable inspection path. This helps to reduce the driving distance and time of the inspection vehicle and improve inspection efficiency. By determining the risk value for each intermediate node (that is, each dangerous goods container) and generating an inspection path based on these risk values, it can ensure that the inspection vehicle gives priority to containers with higher risks, thereby improving the accuracy and pertinence of safety supervision. The improved ant colony algorithm is adaptive and can dynamically adjust the inspection path according to changes in environmental information. This means that when the port environment or the type of dangerous goods changes, the algorithm can automatically adjust the strategy to ensure the timeliness and effectiveness of the inspection path;

[0019] 3. Multiple auxiliary RF positioning tags are installed in the hazardous goods container storage area. These tags serve as reference points to more accurately determine the relative position of the inspection vehicle and the hazardous goods containers. By selecting at least two target auxiliary RF positioning tags, a positioning network can be formed, further improving positioning accuracy and reliability. This method takes into account the possible presence of hazardous goods containers in the signal transmission path between the candidate auxiliary RF positioning tags and the RF receiver on the inspection vehicle, as well as the impact of the hazardous goods information in these containers on signal attenuation. This method can more accurately calculate the signal attenuation coefficient, thereby more accurately determining the target auxiliary RF positioning tags. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, like numbers represent like structures, wherein:

[0021] Figure 1 This is a flow chart of a method for safely supervising dangerous goods in a port shown in one embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments.

[0023] Figure 1 This is a flow chart of a method for safely supervising dangerous goods at a port shown in one embodiment of the present application. Figure 1 As shown, a method for safety supervision of dangerous goods at ports may include the following processes.

[0024] Step 101: Set a status collection device and a radio frequency positioning tag in a dangerous goods container.

[0025] Among them, the status collection device is used to collect the internal status information of the dangerous goods container.

[0026] In some embodiments, a status collection device is provided in a dangerous goods container, including:

[0027] For each dangerous goods type, determining multiple candidate state factors corresponding to the dangerous goods type (e.g., temperature inside the container, smoke concentration inside the container, container vibration, pressure inside the container, gas concentration inside the container, humidity inside the container, etc.), obtaining a first test data set for the dangerous goods type, and determining, based on the first test data set for the dangerous goods type, a correlation coefficient between any two candidate state factors corresponding to the dangerous goods type; and determining, based on the correlation coefficient between any two candidate state factors corresponding to the dangerous goods type, a state factor corresponding to the dangerous goods type;

[0028] Based on the dangerous goods type corresponding to the dangerous goods container and the status factors corresponding to each dangerous goods type, a status collection device is set in the dangerous goods container.

[0029] Specifically, the first test data set of the dangerous goods type may include a test state data set of simulated containers in multiple states corresponding to the dangerous goods type, wherein the type of dangerous goods stored in the simulated container is consistent with the dangerous goods type, and the states of dangerous goods stored in simulated containers in different states are different.

[0030] The multiple candidate state factors corresponding to the dangerous goods type may be state factors whose characteristic values ​​of the candidate state factors in the simulated containers in different states have large differences.

[0031] For example, a flammable liquid is stored in the simulated container.

[0032] State 1: This simulates a container containing flammable liquids under normal transport conditions. Temperature, humidity, pressure, and other parameters within the container remain within safe ranges, and smoke and gas concentrations are also at normal levels.

[0033] State 2: The simulated container also stores flammable liquids, but the temperature inside the container is artificially raised to close to the flash point of the flammable liquid.

[0034] State 3: A slight leakage of flammable liquid stored in a simulated container occurs, causing the gas concentration in the container to increase.

[0035] State 4: Simulates the vibration encountered by the container during storage to simulate the impact of vibration on the stability of the container.

[0036] Condition 5: The humidity in the simulated container is artificially adjusted to extreme levels (too high or too low) to evaluate the impact of humidity on cargo stability.

[0037] For any two candidate state factors corresponding to the dangerous goods type, the Pearson correlation coefficient of the two candidate state factors in each state can be calculated based on the characteristic values ​​of the two candidate state factors at multiple test time points in the test state data set of the simulated container in each state, and the Pearson correlation coefficients of the two candidate state factors in each state are weighted and summed to obtain the correlation coefficient between the two candidate state factors corresponding to the dangerous goods type.

[0038] The status factor corresponding to a dangerous goods type can be determined based on the correlation coefficient between any two candidate status factors corresponding to the dangerous goods type according to the following process:

[0039] S1011. For each candidate state factor, based on the correlation coefficients between the candidate state factor and other candidate state factors, calculate the mean of the correlation coefficients corresponding to the candidate state factor;

[0040] S1012, determine whether the mean correlation coefficient corresponding to each candidate state factor is greater than the mean correlation coefficient threshold. If so, complete the screening; if not, execute S1013;

[0041] S1013. Eliminate candidate status factors whose correlation coefficient mean is less than or equal to the correlation coefficient mean threshold, and execute S1014.

[0042] S1014. For each candidate state factor, based on the correlation coefficient between the candidate state factor and other remaining candidate state factors from the previous round, calculate the mean correlation coefficient corresponding to the candidate state factor, and execute S1012.

[0043] It can be understood that by calculating the correlation coefficient between any two candidate status factors corresponding to the dangerous goods types, the status factors corresponding to the dangerous goods types with correlation can be screened out, providing a data basis for the subsequent realization of the internal stability supervision of dangerous goods containers by combining multiple status factors.

[0044] Step 102: Generate a safety supervision inspection route based on the port storage locations of the plurality of dangerous goods containers and the environmental information of the current cycle.

[0045] In some embodiments, step 102 specifically includes:

[0046] Based on the port storage locations of multiple dangerous goods containers, a dangerous goods container safety supervision and inspection network is constructed, wherein the dangerous goods container safety supervision and inspection network includes a starting point, an end point, and intermediate nodes, and each intermediate node represents a dangerous goods container;

[0047] Determine the risk value of each intermediate node based on the current cycle's environmental information;

[0048] Through the improved ant colony algorithm, a safety supervision inspection path is generated based on the dangerous goods container safety supervision inspection network and the risk value of each intermediate node.

[0049] In some embodiments, determining the risk value of each intermediate node based on the environmental information of the current period includes:

[0050] For each type of dangerous goods, obtaining a second test data set of the dangerous goods type, determining a correlation coefficient between each environmental factor and a risk value of the dangerous goods type based on the second test data set of the dangerous goods type, determining an environmental factor corresponding to the dangerous goods type based on the correlation coefficient between each environmental factor (e.g., ambient temperature, ambient humidity, ambient light, ambient wind speed, etc.) and the risk value of the dangerous goods type, and establishing a risk value prediction model corresponding to the dangerous goods type based on the environmental factors corresponding to the dangerous goods type, wherein the risk value prediction model may be a long short-term memory network (LSTM) model;

[0051] For each intermediate node, the risk value prediction model corresponding to the dangerous goods container is used to determine the initial risk value of the intermediate node based on the environmental information of the current period, and the risk value of the intermediate node is iteratively corrected based on the initial risk values ​​of the adjacent intermediate nodes.

[0052] Specifically, the second test data set of the dangerous goods type may include risk values ​​of simulated containers under different test environments, wherein the risk values ​​of the simulated containers under different test environments may be calculated manually or according to preset rules.

[0053] Specifically, the Pearson correlation coefficient between the environmental factor and the risk value of the dangerous goods type can be calculated based on the characteristic value of each environmental factor in each test environment and the risk value of the simulated container in each test environment. Environmental factors with correlation coefficients greater than a correlation coefficient threshold can be used as the environmental factors corresponding to the dangerous goods type.

[0054] For each intermediate node, key environmental features are extracted from the environmental information of the current period based on the environmental factors corresponding to the type of dangerous goods to which the dangerous goods container represented by the intermediate node belongs, wherein the key environmental features may include the characteristic values ​​of the corresponding environmental factors to which the dangerous goods container represented by the intermediate node belongs at multiple time points, and the initial risk value of the intermediate node is determined based on the key environmental features through the risk value prediction model corresponding to the dangerous goods container.

[0055] For each intermediate node, the adjacent intermediate node of the intermediate node may be an intermediate node whose distance to the intermediate node is less than a preset distance threshold.

[0056] The risk value of the intermediate node can be iteratively corrected according to the following process:

[0057] S1021. For each intermediate node, calculate a risk correction value of the intermediate node based on the initial risk values ​​of adjacent intermediate nodes, and correct the initial risk value of the intermediate node based on the risk correction value of the intermediate node to generate a corrected risk value of the intermediate node.

[0058] S1022. For each intermediate node, calculate the corrected difference between the risk value before correction and the risk value after correction in the current iteration;

[0059] S1023: Determine whether the risk difference of each intermediate node is less than the risk difference threshold. If so, perform the correction. If not, execute S1024.

[0060] S1024. For each intermediate node, based on the risk value of the adjacent intermediate node corrected in the previous iteration, calculate the risk correction value of the intermediate node in the current iteration, correct the risk value of the intermediate node corrected in the previous iteration based on the risk correction value of the intermediate node in the current iteration, generate the corrected risk value of the intermediate node in the current iteration, and execute S1022.

[0061] The risk correction value of the intermediate node can be calculated according to the following formula:

[0062]

[0063] Among them, △R (t,i) is the risk correction value of the i-th intermediate node, w (n,i) is the weight corresponding to the nth adjacent intermediate node of the i-th intermediate node, R (n,i,t-1) is the corrected risk value of the nth adjacent intermediate node of the ith intermediate node at the t-1th iteration, N is the total number of adjacent nodes of the ith intermediate node, r (n,i) is the correlation coefficient between the dangerous goods type of the nth adjacent intermediate node of the i-th intermediate node and the dangerous goods type of the i-th intermediate node, D (n,i) is the straight-line distance between the nth adjacent intermediate node of the i-th intermediate node and the i-th intermediate node, and D0 is a preset straight-line distance. For any two types of dangerous goods, a third test data set of the two types of dangerous goods can be obtained, where the third test data set can include risk values ​​of the two types of dangerous goods under multiple test environments. Based on the risk values ​​of the two types of dangerous goods under the multiple test environments, the Pearson correlation coefficient of the two types of dangerous goods is calculated.

[0064] As you can understand, by considering the risk values ​​of adjacent intermediate nodes and weighting them based on their distance from a specific intermediate node and the relevance of dangerous goods types, we can more comprehensively assess the risk of each intermediate node. This helps capture local risk characteristics within the network, thereby improving the accuracy of the overall risk assessment. This multi-dimensional consideration not only considers the risk values ​​of adjacent nodes, but also factors such as the distance between nodes and the relevance of dangerous goods types. This makes the risk assessment more comprehensive and more accurate in reflecting the actual situation.

[0065] In some embodiments, the heuristic information function and the pheromone update function of the improved ant colony algorithm are related to the risk value of each intermediate node and the travel distance between any two intermediate nodes.

[0066] For example, the heuristic information function of the improved ant colony algorithm can be:

[0067]

[0068] Among them, η (i,j) is the heuristic information from the i-th node to the j-th node, R j is the risk value of the jth node, D (i,j) is the driving distance from the i-th node to the j-th node.

[0069] It can be understood that, through the heuristic information function of the improved ant colony algorithm, ants can be guided to be more inclined to choose paths with high risk values ​​and short travel distances.

[0070] The pheromone update function of the improved ant colony algorithm can be:

[0071]

[0072] Among them, τ ij (t+1) is the pheromone concentration from the i-th node to the j-th node at time t+1, ρ is the pheromone volatility coefficient, and its value range is 0 <p<1,τ ij (t) is the pheromone concentration from the i-th node to the j-th node at time t, is the sum of the pheromone increments left by all ants from the i-th node to the j-th node at time t, △τ ij k is the pheromone increment left by the kth ant from the i-th node to the j-th node at time t, is a function related to the path length and risk value. The higher the risk value of the j-th node, the shorter the travel distance from the i-th node to the j-th node. △τ ij k The bigger.

[0073] As you can understand, the pheromone increment is directly related to the node's risk value and the path length. This means that when a path connects nodes with higher risk values ​​and shorter travel distances, the pheromone increment on that path will be greater. This design allows the algorithm to prioritize patrolling nodes with higher risk values ​​and / or closer distances when selecting patrol paths. Through a reasonable pheromone update strategy, the algorithm can gradually converge to an optimal or near-optimal patrol path. Because the pheromone increment is directly related to the path's risk value and length, the algorithm can more quickly identify a better path during the iteration process, thereby increasing the convergence speed. At the same time, this design also helps improve the solution quality, making the final patrol path more in line with actual needs.

[0074] The improved ant colony algorithm can be used to generate a safety supervision inspection path based on the dangerous goods container safety supervision inspection network and the risk value of each intermediate node according to the following process:

[0075] S1025. Initialize the pheromone matrix: initialize the pheromone matrix of the entire network according to the initial pheromone value.

[0076] S1026. Generate ant colony: Randomly place ants in the network, and each ant represents a potential patrol path.

[0077] S1027. Ant Path Selection: Each ant selects the next node to visit based on its current location and the heuristic information function. Roulette wheel selection, probabilistic selection, and other methods can be used to simulate the ant's decision-making process.

[0078] S1028. Update ant positions and pheromones: After each ant moves one step, its position information is updated. The pheromone matrix in the network is updated according to the pheromone update function.

[0079] S1029, determine the termination condition: check whether the maximum number of iterations or the stopping condition (such as finding the optimal path) has been reached. If the termination condition has not been reached, return to step 3 to continue execution.

[0080] S1030, outputting the optimal path: After all ants have completed their patrol, the optimal patrol path is selected based on the pheromone matrix and the heuristic information function.

[0081] It can be understood that through the above process, the improved ant colony algorithm can generate an effective safety supervision inspection path based on the dangerous goods container safety supervision inspection network and the risk value of each intermediate node.

[0082] Step 103: Control the inspection vehicle to conduct an inspection based on the safety supervision inspection route. During the inspection, receive the internal status information of the dangerous goods container collected by the status collection device, obtain the location information of the dangerous goods container based on the radio frequency positioning tag, and collect images of the dangerous goods container.

[0083] In some embodiments, obtaining location information of a dangerous goods container based on a radio frequency positioning tag includes:

[0084] Install multiple auxiliary radio frequency positioning tags in the dangerous goods container storage area;

[0085] Based on the real-time position of the patrol car and the set positions of the multiple auxiliary radio frequency positioning tags, at least two target auxiliary radio frequency positioning tags are determined from the multiple auxiliary radio frequency positioning tags;

[0086] The location information of the dangerous goods container is obtained based on at least two target-assisted radio frequency positioning tags and a radio frequency positioning tag in the dangerous goods container.

[0087] In some embodiments, based on the real-time location of the patrol vehicle and the locations of the plurality of auxiliary radio frequency positioning tags, determining at least two target auxiliary radio frequency positioning tags from the plurality of auxiliary radio frequency positioning tags includes:

[0088] For each auxiliary RF positioning tag, the real-time distance between the patrol car and the auxiliary RF positioning tag is calculated based on the real-time position of the patrol car and the setting position of the auxiliary RF positioning tag;

[0089] Based on the real-time distance between the patrol car and each auxiliary RF positioning tag, multiple candidate auxiliary RF positioning tags are determined from multiple auxiliary RF positioning tags. For example, the auxiliary RF positioning tags whose real-time distance is less than the real-time distance threshold can be used as candidate auxiliary RF positioning tags;

[0090] For each candidate auxiliary radio frequency positioning tag, based on the port storage locations of multiple dangerous goods containers, the port storage location of the dangerous goods container on the signal transmission path between the candidate auxiliary radio frequency positioning tag and the radio frequency receiver on the inspection vehicle is determined; based on the dangerous goods information (e.g., the type and quantity of dangerous goods) of the dangerous goods container on the signal transmission path between the candidate auxiliary radio frequency positioning tag and the radio frequency receiver on the inspection vehicle, the signal attenuation coefficient of the candidate auxiliary radio frequency positioning tag is determined;

[0091] Based on the signal attenuation coefficient of each candidate auxiliary RF positioning tag, at least two target auxiliary RF positioning tags are determined from multiple candidate auxiliary RF positioning tags. For example, the candidate auxiliary RF positioning tags whose signal attenuation coefficient is less than the signal attenuation coefficient threshold can be used as the target auxiliary RF positioning tags.

[0092] Specifically, an attenuation prediction model can be established to determine the signal attenuation coefficient of the candidate auxiliary RF positioning tag based on the dangerous goods information of the dangerous goods container on the signal transmission path between the candidate auxiliary RF positioning tag and the RF receiver on the patrol vehicle. The attenuation prediction model can be a decision tree model.

[0093] In some embodiments, obtaining location information of the dangerous goods container based on at least two target-assisted radio frequency positioning tags and a radio frequency positioning tag in the dangerous goods container includes:

[0094] For each target-assisted radio frequency positioning tag, the distance between the target-assisted radio frequency positioning tag and the radio frequency receiver on the patrol vehicle is determined based on the signal strength of the target-assisted radio frequency positioning tag received by the radio frequency receiver on the patrol vehicle;

[0095] Based on the signal strength of the radio frequency positioning tag in the dangerous goods container received by the radio frequency receiver on the inspection vehicle, the distance between the radio frequency positioning tag in the dangerous goods container and the radio frequency receiver on the inspection vehicle is determined;

[0096] Based on the distance between each target-assisted RF positioning tag and the RF receiver on the inspection vehicle and the distance between the RF positioning tag in the dangerous goods container and the RF receiver on the inspection vehicle, the location information of the dangerous goods container is generated. For example, using the principle of triangulation positioning or multi-point positioning, based on the distance between each target-assisted RF positioning tag and the RF receiver on the inspection vehicle and the distance between the RF positioning tag in the dangerous goods container and the RF receiver on the inspection vehicle, the location information of the dangerous goods container is generated.

[0097] Step 104 : Perform safety supervision of dangerous goods based on the internal state information, image, and location information of the dangerous goods container.

[0098] In some embodiments, step 104 specifically includes:

[0099] Based on the internal status information of dangerous goods containers, internal stability supervision of dangerous goods containers is carried out;

[0100] Based on the images of dangerous goods containers, the system can monitor the damage and isolation distance of dangerous goods containers.

[0101] Based on the location information of dangerous goods containers, the stacking location of dangerous goods containers is supervised.

[0102] Specifically, for dangerous goods types, an internal stability determination model corresponding to the dangerous goods type can be established, and the internal stability value of the dangerous goods container can be determined based on the internal state information of the dangerous goods container through the internal stability determination model, wherein the internal stability determination model can be a long short-term memory network model.

[0103] In some embodiments, based on the image of the dangerous goods container, container damage monitoring and container isolation distance monitoring of the dangerous goods container are performed, including:

[0104] For each adjacent dangerous goods container of a dangerous goods container, determining an isolation distance between the dangerous goods container and the adjacent dangerous goods container based on an image of the dangerous goods container, determining an isolation distance threshold based on the dangerous goods type of the dangerous goods container and the dangerous goods type of the adjacent dangerous goods container, and performing container isolation distance supervision based on the isolation distance between the dangerous goods container and the adjacent dangerous goods container and the isolation distance threshold;

[0105] Based on the image of the dangerous goods container, the surface flatness of the dangerous goods container is determined, and based on the surface flatness of the dangerous goods container, container damage monitoring of the dangerous goods container is performed.

[0106] Specifically, a distance determination model can be used to determine the separation distance between a dangerous goods container and adjacent dangerous goods containers based on an image of the dangerous goods container. The distance determination model can be a convolutional neural network model. A separation distance threshold can be determined based on the Pearson correlation coefficient of the two dangerous goods types. A higher Pearson correlation coefficient indicates a higher separation distance threshold. When the separation distance between a dangerous goods container and an adjacent dangerous goods container is less than the separation distance threshold, the separation distance is determined to be unsatisfactory.

[0107] The surface flatness of the dangerous goods container can be determined based on the image of the dangerous goods container through a flatness detection model. When the surface flatness of the dangerous goods container is less than a surface flatness threshold, it is determined that the dangerous goods container is damaged. The flatness detection model can be a convolutional neural network model.

[0108] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A method for safety supervision of dangerous goods in ports, characterized in that: include: A status collection device and a radio frequency positioning tag are arranged in the dangerous goods container, wherein the status collection device is used to collect internal status information of the dangerous goods container; Generate safety supervision inspection routes based on the port storage locations of multiple dangerous goods containers and the current cycle's environmental information; Control the inspection vehicle to conduct inspections based on the safety supervision inspection route. During the inspection, it receives internal status information of the dangerous goods container collected by the status acquisition equipment, obtains the location information of the dangerous goods container based on the radio frequency positioning tag, and collects images of the dangerous goods container; Dangerous goods safety supervision is carried out based on the internal status information, images and location information of dangerous goods containers.

2. A method for safety supervision of dangerous goods in ports according to claim 1, characterized in that: Status collection equipment is installed in dangerous goods containers, including: For each dangerous goods type, determining multiple candidate status factors corresponding to the dangerous goods type, obtaining a first test data set of the dangerous goods type, determining a correlation coefficient between any two candidate status factors corresponding to the dangerous goods type based on the first test data set of the dangerous goods type, and determining a status factor corresponding to the dangerous goods type based on the correlation coefficient between any two candidate status factors corresponding to the dangerous goods type; Based on the dangerous goods type corresponding to the dangerous goods container and the status factors corresponding to each dangerous goods type, a status collection device is set in the dangerous goods container.

3. A method for safety supervision of dangerous goods in ports according to claim 2, characterized in that: Generate a safety inspection route based on the port storage location of multiple dangerous goods containers and the current cycle's environmental information, including: Constructing a dangerous goods container safety supervision and inspection network based on the port storage locations of multiple dangerous goods containers, wherein the dangerous goods container safety supervision and inspection network includes a starting point, an end point, and intermediate nodes, and each intermediate node represents a dangerous goods container; Determine the risk value of each intermediate node based on the current cycle's environmental information; Through the improved ant colony algorithm, a safety supervision inspection path is generated based on the dangerous goods container safety supervision inspection network and the risk value of each intermediate node.

4. A method for safety supervision of dangerous goods in ports according to claim 3, characterized in that: Based on the current cycle's environmental information, determine the risk value of each intermediate node, including: For each dangerous goods type, obtaining a second test data set of the dangerous goods type, determining a correlation coefficient between each environmental factor and a risk value of the dangerous goods type based on the second test data set of the dangerous goods type, determining an environmental factor corresponding to the dangerous goods type based on the correlation coefficient between each environmental factor and the risk value of the dangerous goods type, and establishing a risk value prediction model corresponding to the dangerous goods type based on the environmental factors corresponding to the dangerous goods type; For each intermediate node, the risk value prediction model corresponding to the dangerous goods container is used to determine the initial risk value of the intermediate node based on the environmental information of the current period, and the risk value of the intermediate node is iteratively corrected based on the initial risk values ​​of the adjacent intermediate nodes.

5. A method for safety supervision of dangerous goods in ports according to claim 4, characterized in that: The heuristic information function and pheromone update function of the improved ant colony algorithm are related to the risk value of each intermediate node and the travel distance between any two intermediate nodes.

6. A method for safely supervising dangerous goods in a port according to any one of claims 1 to 5, characterized in that: Obtaining the location information of dangerous goods containers based on radio frequency positioning tags, including: Install multiple auxiliary radio frequency positioning tags in the dangerous goods container storage area; Based on the real-time position of the patrol car and the set positions of the multiple auxiliary radio frequency positioning tags, at least two target auxiliary radio frequency positioning tags are determined from the multiple auxiliary radio frequency positioning tags; The location information of the dangerous goods container is obtained based on at least two target-assisted radio frequency positioning tags and a radio frequency positioning tag in the dangerous goods container.

7. A method for safety supervision of dangerous goods in ports according to claim 6, characterized in that: Based on the real-time position of the patrol car and the set positions of multiple auxiliary radio frequency positioning tags, at least two target auxiliary radio frequency positioning tags are determined from the multiple auxiliary radio frequency positioning tags, including: For each auxiliary RF positioning tag, the real-time distance between the patrol car and the auxiliary RF positioning tag is calculated based on the real-time position of the patrol car and the setting position of the auxiliary RF positioning tag; Based on the real-time distance between the patrol car and each auxiliary radio frequency positioning tag, multiple candidate auxiliary radio frequency positioning tags are determined from multiple auxiliary radio frequency positioning tags; For each candidate auxiliary radio frequency positioning tag, based on the port storage locations of multiple dangerous goods containers, the port storage location of the dangerous goods container on the signal transmission path between the candidate auxiliary radio frequency positioning tag and the radio frequency receiver on the inspection vehicle is determined; based on the dangerous goods information of the dangerous goods container on the signal transmission path between the candidate auxiliary radio frequency positioning tag and the radio frequency receiver on the inspection vehicle, a signal attenuation coefficient of the candidate auxiliary radio frequency positioning tag is determined; At least two target auxiliary radio frequency positioning tags are determined from a plurality of candidate auxiliary radio frequency positioning tags based on the signal attenuation coefficient of each candidate auxiliary radio frequency positioning tag.

8. A method for safety supervision of dangerous goods in ports according to claim 6, characterized in that: Acquiring location information of the dangerous goods container based on at least two target-assisted radio frequency positioning tags and a radio frequency positioning tag in the dangerous goods container, including: For each target-assisted radio frequency positioning tag, the distance between the target-assisted radio frequency positioning tag and the radio frequency receiver on the patrol vehicle is determined based on the signal strength of the target-assisted radio frequency positioning tag received by the radio frequency receiver on the patrol vehicle; Based on the signal strength of the radio frequency positioning tag in the dangerous goods container received by the radio frequency receiver on the inspection vehicle, the distance between the radio frequency positioning tag in the dangerous goods container and the radio frequency receiver on the inspection vehicle is determined; The location information of the dangerous goods container is generated based on the distance between each target-assisted radio frequency positioning tag and the radio frequency receiver on the inspection vehicle and the distance between the radio frequency positioning tag in the dangerous goods container and the radio frequency receiver on the inspection vehicle.

9. A method for safely supervising dangerous goods in a port according to any one of claims 1 to 5, characterized in that: Based on the internal status information, images and location information of dangerous goods containers, dangerous goods safety supervision is carried out, including: Based on the internal status information of dangerous goods containers, internal stability supervision of dangerous goods containers is carried out; Based on the images of dangerous goods containers, the system can monitor the damage and isolation distance of dangerous goods containers. Based on the location information of dangerous goods containers, the stacking location of dangerous goods containers is supervised.

10. A method for safety supervision of dangerous goods in ports according to claim 9, characterized in that: Based on the image of dangerous goods containers, the system conducts container damage monitoring and container isolation distance monitoring, including: For each adjacent dangerous goods container of a dangerous goods container, determining an isolation distance between the dangerous goods container and the adjacent dangerous goods container based on an image of the dangerous goods container, determining an isolation distance threshold based on the dangerous goods type of the dangerous goods container and the dangerous goods type of the adjacent dangerous goods container, and performing container isolation distance supervision based on the isolation distance between the dangerous goods container and the adjacent dangerous goods container and the isolation distance threshold; Based on the image of the dangerous goods container, the surface flatness of the dangerous goods container is determined, and based on the surface flatness of the dangerous goods container, container damage monitoring of the dangerous goods container is performed.

Images