Illegal chemical plant detection system based on dangerous chemical transportation track

Abstract

The application discloses a kind of illegal chemical industry facilities detection systems based on dangerous chemical transportation trajectory, compared with existing visualization system, it is applicable to dangerous chemical loading and unloading event detection processing flow;Through integrated space-time visualization method, time and space change are closely combined in a view, to provide more comprehensive analysis tool for identifying illegal chemical industry facilities;Multi-view interactive use supports analyst to quickly browse residence point and provide the timing mode of transport vehicle residence point, the timing mode of transport vehicle residence point can reveal the regularity of transport vehicle transportation, combined with visualization to help analyst to obtain the behavior mode of transport vehicle, and understand dangerous chemical transportation event, for final judgment illegal dangerous chemical facilities decision provides more credible basis, reduces misjudgment rate, saves the manpower and material resources of on-site investigation caused by misjudgment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of illegal facility detection, and particularly relates to an illegal chemical facility detection system based on a dangerous chemical transportation track. BACKGROUND

[0002] Chemicals are widely used in medical treatment, industry and daily life and many other fields. Correct use of chemical materials can benefit society, and illegal operation or transportation of chemicals, especially dangerous chemicals, can cause safety risks, and even endanger the environment and human life. Government departments have invested a lot of effort to find and close illegal facilities that do not have qualifications or are not registered by supervising the production and transportation of chemicals. In order to supervise dangerous chemical transportation vehicles, the government requires GPS trackers to be installed on these transportation vehicles, and based on the tracks of the transportation vehicles, the stay points of these transportation vehicles are analyzed to find illegal facilities.

[0003] In the traditional method, analysts need to check illegal facilities predicted by the stay point detection model in the field, and the analysts check whether the stay point is on the whitelist to determine. But such screening will lead to a large number of false positive results, because the information of the interest points on the whitelist and the map may be out of date. Reference 1 (Z. Zhu, H. Ren, et al. ICFinder: A ubiquitous approach to detecting illegal hazardous chemical facilities with truck trajectories. In Proceedings of ACM SIGSPATIAL, P. 37-40, 2021) develops an ICFinder model to classify loading and unloading events and infer the illegal event probability of suspicious locations, but the model can only provide the classification result of the suspicious location, and cannot provide the reason and basis for the inference, and the stay point is a fixed sensor station, and its position and semantics are known. In addition, the application for an invention with publication number CN 117455344A discloses a logistics track collection and verification method and system, based on vehicle transportation track data provided by multiple parties, the vehicle transportation track is collected and verified, abnormal conditions are found, and the track data is displayed, facilitating data analysis and management by operation personnel. However, the above methods only analyze and calculate the vehicle transportation data through the model or algorithm, lack comprehensive visual analysis methods, cannot directly locate abnormal conditions, and also have false positives in inference.

[0004] Each trajectory is formed by connecting a set of timestamped locations in chronological order, thus possessing both temporal and spatial information. In addition, the movement trajectory of the transport vehicle can also provide clues as to whether the transport vehicle detoured to a suspicious stop, providing a basis for identifying illegal facilities. Reference 2 (N. Ferreira, J. Poco, HTVo, et al. Visual exploration of big spatio-temporal urban data: Astudy of new York city taxi trips. IEEE Transactions on Visualization and Computer Graphics. 19(12): 2149-2158, 2013) designed an interactive interface that allows users to query taxi trips and infer spatial and temporal patterns, but this interactive interface cannot closely integrate trajectory distance and dwell time in a geographical context.

[0005] A scalable and integrated spatiotemporal visualization design can be provided to effectively locate anomalies. For example, invention application CN 116738098A discloses a WebGIS-based urban freight visualization method and system. Based on a WebGIS engine, it constructs a freight visualization model and tightly integrates movement trajectories and time to predict congestion indices. However, this method only provides classification results of freight stops to predict congestion indices; it cannot infer the causes of stops and congestion. Furthermore, since the service target and trajectory data sources are different, it cannot be applied to the detection of illegal facilities. Summary of the Invention

[0006] The purpose of this invention is to provide a detection system for illegal chemical facilities based on the transportation trajectory of hazardous chemicals. Based on the temporal patterns of hazardous chemical transport vehicle behavior, the system visualizes the behavior in three aspects: space, time, and vehicle activity. This visualization allows for interactive viewing of the loading and unloading activities of transport vehicles at potential illegal facilities, analysis of historical temporal patterns of vehicle dwelling behavior, and assessment of suspicious dwelling points to guide analysts in operational and decision-making processes.

[0007] To achieve the above-mentioned objectives, the embodiment provides an illegal chemical facility detection system based on the transportation trajectory of hazardous chemicals, including a data storage module, a map service module, a detection model module, and a visualization module;

[0008] The data storage module is used to store all existing data on the trajectories of hazardous chemical transport vehicles, including geographic points of interest, whitelists, and original trajectory points.

[0009] The map service module is used to provide services including data cleaning, range query, and dwell point detection of stored data;

[0010] The detection model module is used to adjust the clustering results of the docking point detection. The detection model is used to predict the real-time probability of vehicles loading and unloading at docking points based on the clustering results, and to detect suspicious docking points and illegal loading and unloading events based on the probability.

[0011] The visualization module is used to visualize suspicious locations and illegal loading and unloading events, explore suspicious locations for hazardous chemicals, analyze abnormal behavior of hazardous chemical transport vehicles, and identify illegal hazardous chemical facilities.

[0012] In one embodiment, the map service module provides map services using an open-source API or the JD City JUST Data Engine.

[0013] In one embodiment, the detection model module uses the ICFinder model as the detection model to predict the probability of a vehicle loading and unloading goods at a parking point in real time.

[0014] In one embodiment, the visualization module includes a spatial view, a temporal view, and a trajectory activity view;

[0015] The spatial view is used to provide a geographical overview of the stopping points, trajectories, and geographical environment of vehicles transporting hazardous chemicals, helping analysts select suspicious stopping points;

[0016] The time view is used to display the time-series statistics of hazardous chemical transport vehicles staying at selected suspected locations;

[0017] The trajectory activity view is used to visualize the activity trajectory of hazardous chemical transport vehicles, observe the activity patterns of selected hazardous chemical transport vehicles, analyze abnormal behaviors in the activity patterns, and identify illegal hazardous chemical facilities.

[0018] In one embodiment, the spatial view includes multiple functional layer maps, namely a street-style map, a satellite-style map, and a heat map of loading and unloading probabilities. Geographic points of interest are set in the spatial view configuration panel to help understand the semantics of the geographic environment around the points of interest.

[0019] Heatmaps visualize the loading and unloading probabilities of depot locations, with color hue and depth representing loading and unloading probabilities, used to quickly locate suspicious areas; combining heatmaps with clustering results and adjusting the clustering results facilitates precise focusing on suspicious depot locations.

[0020] In one embodiment, the time view includes a time distribution histogram, statistical data of transport vehicles, and a hazardous chemical transport calendar. Figure Three Parts are used to display time information, specifically including:

[0021] The time distribution histogram shows the duration distribution of the frequency of stays of a single transport vehicle at all suspicious stops within a single location area. Records for specific times can be filtered out by scanning the time distribution histogram.

[0022] The statistical data of transport vehicles shows the frequency of hazardous chemical transport vehicles stopping at selected suspicious stops and sorts them. Based on the sorting results, transport vehicles are selected for inspection.

[0023] The hazardous chemicals transport calendar is used to display the transport calendar corresponding to the selected transport vehicle. It shows the dwelling details according to the date. Each date unit is represented by a rectangle. The color visibility of the rectangle's border encodes whether the transport vehicle has made a U-turn. The color depth of the rectangle encodes the total dwelling time of the transport vehicle. The rectangle size encodes the frequency of the transport vehicle's dwelling at suspicious dwelling points within a day. Hovering over a date unit will pop up a prompt showing the transport vehicle's transport activities and highlighting the trajectory of the day in the spatial view and activity view.

[0024] In one embodiment, the trajectory activity view includes: a transportation activity map and a dwell clustering panel;

[0025] The transportation activity map is used to understand and infer the behavior of transport vehicles. It displays the overall events of transport vehicles staying at suspicious stops, including the start time of the stop, the end time of the stop, the number of stops, and the distance between the stop location and the suspicious stop.

[0026] The residency clustering panel is used to display key points in the transportation of hazardous chemicals along the vehicle's trajectory and to summarize the hazardous chemical transportation chain.

[0027] In one embodiment, the dwelling clustering panel uses the DBSCAN method to cluster the dwelling points of a transport vehicle. The clustering results can display whether the dwelling point of the selected transport vehicle is close to a whitelisted hazardous chemical facility and display the name of the hazardous chemical facility.

[0028] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0029] This invention provides an illegal chemical facility detection system based on the transportation trajectory of hazardous chemicals. Compared with existing visualization systems, it is applicable to the loading and unloading detection and processing of hazardous chemicals. Through an integrated spatiotemporal visualization method, it closely combines time and spatial changes in one view, providing a more comprehensive analytical tool for identifying illegal chemical facilities.

[0030] Multi-view interactive features allow analysts to quickly browse vehicle stops and provide time-series patterns of vehicle stops. These time-series patterns can reveal the regularity of vehicle transportation. Combined with visualization, this helps analysts obtain behavioral patterns of vehicles and understand hazardous chemical transportation incidents. This provides a more credible basis for the final judgment of illegal hazardous chemical facilities, reduces the misjudgment rate, and saves manpower and resources for on-site investigations caused by misjudgments. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0032] Figure 1 This is a schematic diagram of the structure of an illegal chemical facility detection system;

[0033] Figure 2 This is a schematic diagram of the visualization module;

[0034] Figure 3 A spatial view for the visualization module;

[0035] Figure 4 A time view for the visualization module;

[0036] Figure 5 This is the active view for the visualization module. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of this invention.

[0038] To help analysts identify illegal chemical facilities in complex detection scenarios, this embodiment provides an illegal chemical facility detection system based on the transportation trajectory of hazardous chemicals. The system visualizes the behavior of hazardous chemical transport vehicles in terms of space, time, and vehicle activity based on the temporal patterns of the vehicles. Through visualization, users can interactively view the loading and unloading activities of transport vehicles of potential illegal facilities, helping analysts to analyze the historical temporal patterns of transport vehicle stay behavior and assess suspicious stay points.

[0039] like Figure 1 As shown, the illegal chemical facility detection system provided in this embodiment includes: a data storage module, a map service module, a detection model module, and a visualization module.

[0040] In this embodiment, the data storage module stores all existing data on the trajectories of hazardous chemical transport vehicles, including geographic points of interest (POIs), original GPS points of vehicle trajectories, and a whitelist of hazardous chemical facilities. The geographic POIs record the location information (longitude and latitude), category, and name of the POIs. The whitelist of hazardous chemical facilities records the location points (longitude and latitude), facility names, and production categories of the facilities. The original GPS points of vehicle trajectories record the vehicle's location information, including timestamps, longitude, and latitude.

[0041] In this embodiment, the map service module provides map services including data cleaning, range querying, and dwell point detection of stored data. Within the map service module, analysts utilize open-source API services or map services provided by JD City JUST Data Engine to clean the raw data stored in the data storage module, ensuring that the cleaned data is of the expected type and range, and ensuring consistency in the data structure before and after cleaning. Next, range querying is performed, specifically allowing queries based on known polygonal regions to obtain dwell points geographically belonging to those regions. Dwell points in the original vehicle trajectory are then detected based on the dwell points within the polygonal regions, with different parameters supporting dwell duration thresholds or dwell distance thresholds.

[0042] In this embodiment, the detection model module is used to adjust the clustering results of the stop point detection. The detection model is used to predict the real-time probability of a vehicle loading or unloading at a stop point based on the clustering results, and to detect suspicious stop points and illegal loading or unloading events based on the probability. Specifically, the ICFinder model is used as the detection model. First, the original GPS points of the vehicle trajectory are spatially clustered for stop points. Analysts analyze and adjust the clustering results to obtain optimized clustering results. Based on the optimized clustering results, the real-time probability of a vehicle loading or unloading at a stop point is predicted, and suspicious stop points and illegal loading or unloading events are detected based on the probability.

[0043] In this embodiment, based on the data storage module, map service module, and detection model module, a visualization module is designed to visualize the spatial, temporal, and vehicle activity behaviors. This visualization allows for interactive viewing of the loading and unloading activities of transport vehicles, helping analysts make decisions and identify illegal facilities. Figure 2 The image shows the interface of the visualization module system, which includes: spatial view, time view, and trajectory activity view.

[0044] In this embodiment, the spatial view includes multiple functional layer diagrams, such as... Figure 3 As shown in a and b, the maps include street-style and satellite-style maps, providing a geographical overview of the stops, tracks, and geographic environment of vehicles transporting hazardous chemicals; the spatial view displays corresponding icons for different point-of-interest categories, which analysts can view in the configured panel.Figure 3 As shown in a3, geographic points of interest (POIs) are set up to help analysts understand the semantics of the geographic environment around the POIs. In addition, the spatial view also provides the original trajectories and original stopping points of hazardous chemical transport vehicles. Analysts can manually check the clustering results of the stopping point detection model in conjunction with the detection model module. If the spatial clustering results are incorrect, analysts can draw polygonal areas on the map and manually cluster the stopping points to remove abnormal results. The normal clustering results are used as model input to obtain the loading and unloading probability of stopping points. Analysts can adjust the spatial clustering of stopping points to predict the probability of vehicles loading and unloading at stopping points in real time, which is used to locate suspicious areas.

[0045] like Figure 3 Figure 'c' shows a heatmap in the spatial view used to visualize the probability of loading and unloading at a depot. The hue and depth of the color represent the probability value of loading and unloading; the redder a location is, the greater the likelihood of a loading or unloading event occurring there. Overlaying the heatmap onto the map allows analysts to perceive the distribution of potential violations by transport vehicles and quickly locate suspicious areas within the heatmap. Considering that loading and unloading events often occur near whitelisted facilities, whitelist icons are also displayed on the heatmap. Analysts can filter loading and unloading locations near whitelisted facilities from the loading and unloading probability heatmap to focus more intently on suspicious depots.

[0046] like Figure 3 As shown in a2, the spatial view uses the original geographic location by default instead of the corrected location matched by the geographic service to display the trajectory. This is because roads within the park cannot be well included in the road network dataset provided by the geographic service, resulting in defects in map matching near the docking point. It is worth noting that a docking point is not just a "point," but can be a small area where the movement range of hazardous chemical transport vehicles is less than the docking distance threshold in docking point detection. Illegal loading / unloading events are precisely hidden in such trajectories. Therefore, these transport trajectories near or within docking points are important references for analysts to judge the behavior of hazardous chemical transport vehicles (e.g., turning details). Within the trajectory, docking points are visualized with different styles of markers according to their type ( Figure 3 As shown in a1), analysts can choose whether to perform transportation trajectory network matching as needed.

[0047] In this embodiment, the time view displays the time-series statistics of events involving transport vehicles staying at selected suspected locations. Analysts can interactively display the historical dwell time of all transport vehicles in the selected area, including dwell duration, dwell frequency, and dwell time distribution, as needed. Figure 4 As shown, this includes a time distribution histogram, statistical data of transport vehicles, and a hazardous chemical transport calendar. Figure ThreeThis view has three sections to display time information;

[0048] When an analyst selects a location area, all suspected locations within that area are selected, and a hazardous chemical transport calendar chart pops up on the left side of the time view, such as... Figure 4 Figure 'a' shows the loading and unloading probability and the number of stopping points for vehicles transporting hazardous chemicals; as shown in Figure 'a'. Figure 4 As shown in b, the time distribution histogram displays the duration distribution of the frequency of stays of a single transport vehicle at all suspicious stops within a single location area. By scanning the time distribution histogram, analysts can decide to filter the time based on specific circumstances. For example, whether a hazardous chemical transport vehicle is loading or unloading chemicals for a short period of time or is parked for a long time, the system will display all stops by default if no filters are used.

[0049] Next, statistical data on transport vehicles is displayed and ranked based on the frequency of their stops at selected suspicious locations. Vehicles are then selected for inspection based on this ranking. Figure 4 As shown in c, the visualization module groups and counts the stops based on the license plates of hazardous chemical transport vehicles. The chart here sorts and displays the stops based on the frequency of hazardous chemical transport vehicles stopping at selected suspicious stops. Specifically, two criteria are provided for experts to sort: the number of stops near the suspicious stops (frequency of stay) and the number of days the stop event occurred at the suspicious stops (number of days of stay).

[0050] The hazardous chemicals transport calendar is used to display the transport calendar corresponding to a selected transport vehicle, such as... Figure 4 As shown in d, the transport calendar displays the daily dwell details of transport vehicles. Considering that the behavior of hazardous chemical transport vehicles involves multiple attributes, a graphic visualization is chosen to represent this data. Each date unit is represented by a rectangle with three visual elements coded. For example, hovering will bring up a tooltip ( Figure 4 As shown in 'e', ​​this plot displays the distribution of dwell time for a specific transport vehicle throughout the day; for example... Figure 4As shown in 'f', the visibility of the orange border encodes whether a transport vehicle makes a U-turn at that location. Based on domain knowledge, when transport vehicles load or unload chemicals at a location, most vehicles will make a U-turn at the chemical facility. If the trajectory corner at a suspicious location is an acute angle, the rectangle will have a border, indicating a U-turn. If a transport vehicle stops more than once a day, the rectangle will have a border as long as one of its stops has a U-turn angle. Secondly, the color depth of the rectangle encodes the total duration of the transport vehicle's stopover; the longer the duration, the darker the rectangle. Thirdly, the rectangle size encodes the frequency of the transport vehicle's stopover at a suspicious stopover point within a day. The motivation for this design is to identify transport vehicles that stop at a location more than once. Each date unit in the calendar is interactive; clicking on the interaction highlights the transport vehicle's trajectory for that day in both the spatial and activity views. By viewing the time patterns of related vehicles, the spatiotemporal information of these suspicious stopover points can be further analyzed.

[0051] In this embodiment, the trajectory activity view is used to visualize the activity trajectories of hazardous chemical transport vehicles, observe the activity patterns of selected hazardous chemical transport vehicles, analyze abnormal behaviors within these patterns, and identify illegal hazardous chemical facilities. For example... Figure 5 As shown, the trajectory activity view includes a transportation activity graph and a dwell clustering panel;

[0052] like Figure 5 As shown in Figure 'a', the transportation activity map closely integrates the Euclidean distance from each stop point to a suspected stop point with the time periods of vehicle movement and stops. This facilitates understanding and inference about vehicle behavior, providing a holistic view of vehicle stops at suspected stops, including the start and end times of stops, the number of stops, and the distance between the stop point and the suspected stop point. This spatiotemporal contextual information enhances analysts' understanding and inference of vehicle transportation behavior. In particular, the close connection between relative distance and absolute time visually demonstrates vehicle movement and stop events, indicating whether a suspected stop point is a starting point, destination, or a point of transit.

[0053] Specifically, the transport activity view displays the spatiotemporal information of the transport vehicle's trajectory in a two-dimensional space. Figure 5(As shown in a1). First, the horizontal axis represents the time of day, with the starting point at 0:00 and the ending point at 24:00. Second, the vertical axis represents the Euclidean distance from the stop point to the suspected stop point. Each stop point has two timestamps: the start time and the end time of the stop. The stop event of a stop point is visualized as a horizontal line segment. The horizontal coordinates of the start and end of the line segment are determined by the start time and the end time of the stop. The diagonal line segment represents the movement of the transport vehicle. The horizontal axis shows the distribution of the absolute number of stops, which helps analysts to have an overall understanding of the stop events of the suspected stop points.

[0054] In this embodiment, the dwelling clustering panel is used to display key points in the hazardous chemical transportation trajectory of transport vehicles, summarizing the hazardous chemical transportation chain. For example, a hazardous chemical transport vehicle may have similar movement and dwelling patterns every day, such as frequent dwelling points. Therefore, the DBSCAN method is used to cluster all transport vehicle dwelling points in geographic locations, and combined with the visualization results from the transportation activity map, the clustering results can show whether the dwelling points of selected transport vehicles are close to hazardous chemical facilities on a whitelist and display the names of the hazardous chemical facilities.

[0055] Specifically, to obtain an appropriate number of dwell points, the minimum sample size was selected as 5 based on the DBSCAN method, and the maximum distance between two samples was 100 meters. For example... Figure 5 As shown in b, relevant locations are sorted according to their frequency of occurrence in the transportation trajectory, and milestone labels, such as origin, destination, or transit points, are provided according to the order in which they appear in the transportation trajectory.

[0056] The transport activity map and the dwell clustering panel share the same color coding. In the transport activity map, horizontal line segments representing dwell events are coded with the cluster color or gray where no clustering occurs, while gray diagonal lines represent events where transport vehicles move. When analysts select a specific line segment in the activity view by hovering or clicking... Figure 5 As shown in a2), in the active view and the space view ( Figure 5 The corresponding trajectory segment (as shown in c) highlights the clustering results, and the calendar cells in the time view also display the corresponding information; when there are overlapping stationary line segments at the hover position, a list of all dates to which the line segments belong will be displayed for analysts to further select and highlight dates; when the analyst hovers the mouse over a stationary event, detailed information about the stationary event will appear (…). Figure 5 (as shown in a3); In addition, analysts can locate cluster locations in the resident clustering panel, and the spatial view will update the viewport to that location. Figure 5 As shown in b1 and c1 in the diagram, the clustering results can show whether the selected transport vehicle's stop is close to a whitelisted hazardous chemical facility and display the name of the hazardous chemical facility.

[0057] The specific embodiments described above illustrate the technical solution and beneficial effects of the present invention in detail. It should be understood that the above description is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A detection system for illegal chemical facilities based on the transportation trajectory of hazardous chemicals, characterized in that, include: Data storage module, map service module, detection model module, and visualization module; The data storage module is used to store all existing data on the trajectories of hazardous chemical transport vehicles, including geographic points of interest, whitelists, and original trajectory points. The map service module is used to provide map services including data cleaning, range query, and dwell point detection of stored data; The detection model module is used to adjust the clustering results of the docking point detection. The detection model is used to predict the probability of vehicles loading and unloading at docking points based on the clustering results, and to detect suspicious docking points and illegal loading and unloading events based on the probability. The visualization module is used to visualize suspicious storage points and illegal loading and unloading events, explore suspicious storage points of hazardous chemicals, analyze abnormal behavior of hazardous chemical transport vehicles, and identify illegal hazardous chemical facilities; wherein, the visualization module includes spatial view, temporal view and trajectory activity view; The spatial view displays the stopping points, trajectories, and geographical overview of hazardous chemical transport vehicles, helping analysts select suspicious stopping points. The spatial view includes multiple functional layers: a street-style map, a satellite-style map, and a heatmap of loading and unloading probabilities. The visibility of geographic points of interest (POIs) can be set in the spatial view configuration panel to help understand the semantics of the surrounding geographic environment. The heatmap visualizes the loading and unloading probabilities of stopping points; the hue and depth of the colors represent the loading and unloading probabilities, used for quickly locating suspicious areas. Combining the heatmap with clustering results and adjusting the clustering results facilitates precise focusing on suspicious stopping points. The time view is used to display the time-series statistics of hazardous chemical transport vehicles' stays at selected suspicious stops. The time view displays time information through three parts: a time distribution histogram, a statistical data graph of transport vehicles, and a hazardous chemical transport calendar graph. Specifically, the time distribution histogram shows the duration distribution of the frequency of a single transport vehicle's stays at all suspicious stops within a single location area; records for specific times are filtered out by refreshing the time distribution histogram. The statistical data graph of transport vehicles displays and sorts the frequency of hazardous chemical transport vehicles' stays at selected suspicious stops, and transport vehicles are selected for inspection based on the sorting results. The hazardous chemical transport calendar graph displays the transport calendar corresponding to the selected transport vehicle, showing stay details according to the date. Each date unit is represented by a rectangle; the color visibility of the rectangle's border encodes whether the transport vehicle has made a U-turn; the color depth of the rectangle encodes the total length of the transport vehicle's stay; and the rectangle's size encodes the frequency of the transport vehicle's stays at suspicious stops within a day. Hovering over a date unit will pop up a prompt showing the transport vehicle's activity behavior and highlighting the trajectory for that day in both the spatial view and the activity view. The trajectory activity view is used to visualize the activity trajectories of hazardous chemical transport vehicles, observe the activity patterns of selected hazardous chemical transport vehicles, analyze abnormal behaviors in the activity patterns, and identify illegal hazardous chemical facilities. The trajectory activity view includes a transport activity map and a dwelling cluster panel. The transport activity map is used to understand and infer the behavior of transport vehicles, displaying the overall dwelling events of transport vehicles at suspected dwelling points, including the start time of the dwelling, end time of the dwelling, number of dwellings, and distance between the dwelling point and the suspected dwelling point. The dwelling cluster panel is used to display key points of hazardous chemical transport in the transport vehicle trajectory and summarize the hazardous chemical transport chain.

2. The illegal chemical facility detection system based on hazardous chemical transportation trajectory according to claim 1, characterized in that, The map service module uses open-source APIs or JD City JUST data engine to provide map services.

3. The illegal chemical facility detection system based on hazardous chemical transportation trajectory according to claim 1, characterized in that, The detection model module uses the ICFinder model as the detection model to predict the probability of loading and unloading events occurring at the dwell point.

4. The illegal chemical facility detection system based on hazardous chemical transportation trajectory according to claim 1, characterized in that, The aforementioned dwelling clustering panel uses the DBSCAN method to cluster the dwelling points of a transport vehicle. The clustering results can display whether the dwelling point of the selected transport vehicle is close to a whitelisted hazardous chemical facility and display the name of the hazardous chemical facility.

Images