A drug monitoring system

Abstract

The present invention relates to the technical field of drug monitoring, and specifically to a drug monitoring system, including a data acquisition module, a data preprocessing module, a drug analysis module, an abnormal analysis module and a data display module. In the drug analysis module, the data analysis layer uses a clustering method to obtain the input data of each indicator of the indicator layer, and the indicator layer analyzes each indicator according to the input data to obtain the output data of each indicator. The comprehensive evaluation layer is used to comprehensively analyze the output data of each indicator, the weight of each indicator and the preprocessing data to obtain a comprehensive evaluation result. The abnormal analysis module is used to analyze the output data and preprocessing data of each indicator in the drug analysis module to obtain an abnormal analysis report. The data display module visualizes the comprehensive evaluation results and the abnormal analysis report. The present invention solves the problem that drug control units cannot realize alarm monitoring, evaluate drug levels, and view drug information functions in each area in a unified system.

Description

Technical Field

[0001] The present invention relates to the technical field of drug monitoring, and in particular to a drug monitoring system. Background Art

[0002] At present, the global drug problem continues to spread. As the characteristics of drug-related illegal and criminal gangs, such as strong tightness, strong concealment, and wide geographical span of activities, are becoming increasingly obvious, the crackdown and control work has also exposed the drawbacks of difficulty in information sharing and in-depth mining. In order to improve the effectiveness of anti-drug crackdown and control, and strengthen the ability to actively discover and actively crack down, it is necessary for anti-drug units to integrate resources to establish corresponding data system models, automatically analyze suspected hidden drug users, key drug-related places and other objects from massive drug-related data, and focus on their control and crackdown.

[0003] At present, although the business systems of drug control units have a large amount of data, due to the different sources, types, structures and other reasons of this data, it is difficult to precipitate and share information, and it is difficult to dig deep, which leads to incomplete analysis and failure to discover important intelligence in time in actual application. The information query system provided by the drug control unit queries each system separately according to the elements, and it is impossible to unify the functions of alarm monitoring, drug level assessment, and viewing drug information in each area in one system.

[0004] Therefore, the present invention provides a drug monitoring system to solve the above problems. Summary of the invention

[0005] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a drug detection system to solve the problem that the information query systems provided by the above-mentioned drug control units are all queried in each system according to factors, and it is impossible to unify the functions of alarm monitoring, evaluating drug levels, and viewing drug information in each area in one system.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A drug monitoring system includes a data acquisition module, a data preprocessing module, a drug analysis module, an abnormal analysis module and a data display module. The data acquisition module is used to obtain drug-related data in the monitoring area, and the drug-related data at least includes drug-related personnel data, drug data, drug enforcement data, and sewage toxin content data. The data preprocessing module is used to extract data from the drug-related data and perform data format conversion to obtain preprocessed data.

[0008] The poison situation analysis module uses the data analysis layer, indicator layer and comprehensive evaluation layer in sequence to analyze the pre-processed data to obtain a comprehensive evaluation result, wherein the data analysis layer uses a clustering method to obtain the input data of each indicator of the indicator layer, the indicator layer is used to analyze each indicator according to the input data to obtain the output data of each indicator, and the comprehensive evaluation layer is used to comprehensively analyze the output data of each indicator, the weight of each indicator and the pre-processed data to obtain a comprehensive evaluation result, which at least includes the poison situation level and poison situation distribution information of the monitored area.

[0009] The abnormal analysis module uses an abnormal analysis model to analyze the output data of each indicator in the drug situation analysis module that is greater than the corresponding preset threshold and the corresponding pre-processed data to obtain an abnormal analysis report, which at least includes information on drug-related personnel, key locations, and hidden personnel information. The data display module is used to visualize the comprehensive evaluation results and abnormal analysis reports.

[0010] Preferably, the data analysis layer includes extracting input data of each indicator of the indicator layer from the preprocessed data according to preset labels, and storing it in the input list of the corresponding indicator; performing cluster analysis on the data not extracted from the preprocessed data to obtain classification labels; determining the correlation between the classification label and the indicator through expert scoring, and if the correlation is greater than the preset correlation, adding the classification label to the preset label list, and adding the preprocessed data corresponding to the classification label to the input data of the corresponding indicator.

[0011] Preferably, the indicator layer includes location indicators, drug-related personnel indicators, dangerous path indicators, drug indicators and sewage toxin content indicators; wherein, the location indicators include counting the number of repetitions of repeated data in the input data of the statistical location indicators, determining the location anomaly level according to the corresponding relationship between the number of repetitions and the preset interval, and displaying the location distribution information through the data display module; the drug-related personnel indicators include determining the danger level and abnormal record information of drug-related personnel at least according to the illegal records and activity trajectories of drug-related personnel in the pre-processed data; the dangerous path indicators include determining the drug transportation path and transportation method according to the drug-related personnel data and drug enforcement data in the pre-processed data, determining the transportation level according to the relationship between the number of repetitions of the drug transportation path and transportation method and the threshold, and displaying the drug transportation information in the data display module; the drug indicators include determining the drug inflow level according to the relationship between the number of drugs and the type of drugs seized in the pre-processed data and the corresponding preset threshold, and visually displaying the drug data in the data display module; the sewage toxin content indicators determine the toxin content level according to the relationship between the sewage toxin content data in the pre-processed data and the preset toxin content threshold.

[0012] Preferably, in the drug-related personnel index, the relationship between drug-related personnel is analyzed according to the drug-related personnel data in the pre-processed data to establish a relationship network; primary drug-related personnel whose number of relationships exceeds a first set threshold are screened, and the violation records of the primary drug-related personnel are obtained, and the danger level of the primary drug-related personnel is determined according to the number of violation records and the severity of the violations, and the information of the primary drug-related personnel whose danger level exceeds the preset danger level is used as abnormal record information.

[0013] Preferably, the comprehensive evaluation layer includes determining the weight of each indicator by expert scoring; performing weighted average calculation on the evaluation level and weight of each indicator to obtain the toxicity level; and forming a comprehensive evaluation result based on the preprocessed data, the toxicity level, and the output data of each indicator.

[0014] Preferably, in the abnormal analysis model, output data is obtained in which the level of each indicator in the indicator layer is greater than the corresponding preset threshold, and the output data includes information on drug-related persons, location abnormality information, drug inflow information and transportation path information; monitoring images at the location abnormality information and the transportation path are obtained; the monitoring images are identified by using an image recognition model to determine whether there are drug-related persons, and if so, the monitoring images are processed to obtain the head portrait of the interacting person who interacts with the drug-related person in the monitoring image, determine that the interacting person is an initial hidden person, and connect to the identity authentication system to obtain the identity information of the initial hidden person; according to The initial hidden person identity information obtains the travel information of the initial hidden person, and determines whether there is abnormal travel data in the travel information. If there is abnormal travel data, the initial hidden person is determined to be a hidden person; the image recognition model is used to analyze the monitoring image, and the interactive personnel information of the hidden person is obtained, and it is determined whether there is abnormal information in the interactive personnel information of the hidden person. If so, the hidden person is determined to be a secondary hidden person, and a relationship library between the drug-related personnel information, the hidden personnel information and the secondary hidden personnel information and the key positions where the contact frequency exceeds the set frequency are established; an abnormal analysis report is formed according to the relationship library and the key positions.

[0015] Preferably, the process of processing surveillance images by the image recognition model includes: obtaining surveillance images containing drug-related persons according to the itinerary of the drug-related persons, the surveillance images being images of set resolution, and storing the surveillance images in a first database in the order of itinerary; extracting avatar images of the interacting persons in multiple positions in the surveillance images, performing image stitching and image enhancement on the avatar images in multiple positions to obtain frontal avatars of the interacting persons.

[0016] Preferably, the abnormal record information includes at least criminal records, entry and exit records, dangerous driving path records, and records of abnormal entry and exit locations.

[0017] Preferably, the abnormal analysis module also includes: establishing a quantity prediction model using a convolutional neural network based on historical sewage toxin content data and corresponding historical drug addict data, analyzing the sewage toxin content data using the quantity prediction model to obtain a predicted number of drug addicts, and observing and analyzing the personnel in the abnormal analysis report based on the predicted number of drug addicts to determine suspected drug addicts.

[0018] Preferably, the drug situation monitoring system also includes: when the drug situation level exceeds the preset drug situation level, the abnormal analysis report is analyzed using a strategy generation model based on historical abnormal data and response measures to generate a response strategy, and an alarm message, abnormal analysis report and response strategy are sent to the monitoring personnel.

[0019] The beneficial effects of the present invention are:

[0020] 1. In the poison situation analysis module of the present invention, the pre-processed data obtained by the data pre-processing module is analyzed layer by layer through the data analysis layer, the indicator layer and the comprehensive evaluation layer to obtain the comprehensive evaluation result of the poison situation analysis. First, the data analysis layer is used to perform cluster analysis on the pre-processed data to obtain the input data of each indicator of the indicator layer; secondly, in the indicator layer, each indicator analyzes its own input data to determine the current level of each indicator and obtain output data containing the current level; then, the comprehensive evaluation layer is used to perform a comprehensive analysis on the output data of each indicator, the weight of each indicator and the pre-processed data to obtain a comprehensive evaluation result containing the poison situation level and poison situation distribution information of the monitored area. In addition, the abnormal analysis module is used to analyze the output data and corresponding pre-processed data of the indicators in the drug situation analysis module whose levels are greater than the corresponding preset thresholds, and to analyze the abnormal data of the drug situation in the monitored area to obtain an abnormal analysis report including information on drug-related personnel, key locations and hidden personnel; finally, the comprehensive evaluation results and the abnormal analysis report are visualized using the data display module, which can provide detailed and intuitive reports and data, solving the problem that the information query systems provided by anti-drug units all query in each system according to factors, and cannot realize the functions of alarm monitoring, evaluating drug situation levels, and viewing drug situation information in each area in a unified system, saving the time for anti-drug units to obtain data from multiple systems and analyze and process them, thereby improving case handling efficiency.

[0021] 2. In the data analysis layer of the present invention, based on preset labels, input data of each indicator of the indicator layer are extracted from the preprocessed data; at the same time, cluster analysis is performed on the data not extracted from the preprocessed data to obtain the classification labels of the preprocessed data, and the correlation between the classification labels and each indicator is determined by means of multiple expert scores to screen out classification labels with a correlation greater than a preset correlation, and the preprocessed data corresponding to the classification labels are added to the input data of the corresponding indicators, so as to analyze and process each indicator to obtain accurate output data, thereby ensuring that the comprehensive analysis layer and the abnormality analysis module output accurate results.

[0022] 3. In the indicator layer of the drug situation analysis module of the present invention, in the position indicator, the number of repetitions of the position in the input data of the position indicator is counted, and the position abnormality level is determined according to the corresponding relationship between the number of repetitions and the preset interval, so as to focus on monitoring and surprise inspection of these positions; in the drug-related personnel indicator, at least the danger level and abnormal record information of the drug-related personnel are determined according to the illegal records and activity trajectories of the drug-related personnel, so that the subsequent staff can focus on monitoring and inquiring the drug-related personnel and learn more intelligence information; in the dangerous path indicator, the drug transportation path and transportation method are determined according to the drug-related personnel data and the drug enforcement data, and the transportation level is determined according to the corresponding relationship between the number of repetitions of the drug transportation path and the transportation method and the threshold, so as to focus on checking the above-mentioned transportation method at certain positions of the transportation path; in the drug indicator, the drug inflow level is determined according to the relationship between the number of seized drugs, the type of drugs and the corresponding preset threshold; in the sewage toxin content indicator, the toxin content level is determined according to the relationship between the sewage toxin content data and the preset toxin content threshold. The present invention can obtain the comprehensive analysis of the drug situation level and the current drug situation status by analyzing the level and corresponding information of the above-mentioned indicators.

[0023] 4. In the abnormal analysis module of the present invention, firstly, the output data of each indicator in the indicator layer whose level is greater than the corresponding preset threshold value is obtained, and the monitoring images that can be obtained at the abnormal position and on the transportation path are obtained according to the abnormal position information and transportation path information in the output data; then, the monitoring image is identified and analyzed by using the image recognition model to determine whether there are drug-related persons recorded in the monitoring image; if so, the monitoring image is processed to obtain the head portrait of the interacting person who communicates with the drug-related person in the monitoring image, determine that the interacting person is the initial hidden person, and connect to the identity authentication system to obtain the identity information of the initial hidden person, and then obtain the itinerary information of the initial hidden person according to the identity information of the initial hidden person, and determine whether the itinerary information contains the drug-related person. There is abnormal travel data. If there is abnormal travel data, the initial hidden person is determined as a hidden person; then the image recognition model is used again to perform the above analysis process on the hidden person to determine the secondary hidden person; and according to the drug-related person information, the hidden person information and the secondary hidden person information, a relationship library between the drug-related person information, the hidden person information and the secondary hidden person information and the key positions where the contact frequency exceeds the set frequency is established to determine the main hidden persons and key positions, so as to monitor the above-mentioned persons and key positions and find intelligence clues. Compared with the traditional method of manually analyzing monitoring images to determine suspects, the present invention can save the time of staff in analyzing monitoring images and obtaining clues, thereby improving work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic flow chart of the drug monitoring system of the present invention;

[0025] Figure 2 It is a schematic structural diagram of the indicator layer in the drug situation analysis module of the present invention. DETAILED DESCRIPTION

[0026] The following will refer to Figure 1 to Figure 2 The embodiments of the present invention are described in detail. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0027] A drug monitoring system, such as Figure 1 As shown, it includes a data acquisition module, a data preprocessing module, a drug situation analysis module, an abnormal analysis module and a data display module. Among them, the data acquisition module is used to obtain drug-related data in the monitoring area, and the drug-related data at least includes drug-related personnel data, drug data, drug enforcement data, and sewage toxin content data. The data preprocessing module is used to extract data from the drug-related data and perform data format conversion to obtain preprocessed data.

[0028] The toxicity analysis module includes analyzing the pre-processed data in sequence using the data analysis layer, the indicator layer and the comprehensive evaluation layer to obtain a comprehensive evaluation result, wherein the data analysis layer uses a clustering method to obtain the input data of each indicator of the indicator layer, the indicator layer is used to analyze each indicator according to the input data to obtain the output data of each indicator, and the comprehensive evaluation layer is used to conduct a comprehensive analysis of the output data of each indicator, the weight of each indicator and the pre-processed data to obtain a comprehensive evaluation result, which at least includes the toxicity level and toxicity distribution information of the monitored area.

[0029] The abnormal analysis module includes using an abnormal analysis model to analyze the output data and corresponding pre-processed data of each indicator in the drug situation analysis module whose level is greater than the corresponding preset threshold to obtain an abnormal analysis report. The abnormal analysis report at least includes information on drug-related persons, key locations, and hidden personnel information.

[0030] The data display module is used to visualize the comprehensive evaluation results and abnormal analysis reports, and uses conventional visualization tools to display them in the display page of the corresponding function of the toxic monitoring system. For example, key locations, toxic levels, toxic information, etc. are displayed on the map; and some locations can be clicked to view monitoring images, sewage toxin content data, etc.

[0031] Preferably, drug-related personnel refer to personnel who are related to drug contact, including but not limited to drug users, drug dealers, and personnel working in key places. Drug enforcement data refers to data related to drugs seized by anti-drug units during the investigation, including but not limited to the number of people, categories, locations, time, etc. The clustering methods used in the present invention include but are not limited to K-MEANS algorithm, CLARANS algorithm, CHAMELEON algorithm, OPTICS algorithm, CLIQUE algorithm, WAVE-CLUSTER algorithm, etc. When used, different or the same clustering algorithms are used for cluster analysis for different pre-processed data; for example, the K-MEANS algorithm is used, or the CLARANS algorithm is used for drug-related personnel data, and the CLIQUE algorithm is used for sewage toxin content data. In the data preprocessing module, by preprocessing various types of drug-related data from different sources, the format of the drug-related data is converted into the required format of the input data that meets the requirements of the drug situation analysis module, so that the drug situation analysis module can process the data.

[0032] In addition, based on historical drug-related data, experts established preset thresholds corresponding to different levels of each indicator according to existing experience, and calculated the preset thresholds established by these experts using the averaging method; and established a table of the correspondence between the level of each indicator and the preset thresholds, which was stored in the database for easy retrieval.

[0033] In this embodiment, in the drug situation analysis module, the present invention analyzes the preprocessed data obtained by the data preprocessing module layer by layer through the data analysis layer, the indicator layer and the comprehensive evaluation layer to obtain a comprehensive evaluation result of the drug situation analysis. First, the data analysis layer is used to perform cluster analysis on the pre-processed data to obtain the input data of each indicator in the indicator layer; secondly, in the indicator layer, each indicator analyzes its own input data to determine the current level of each indicator and obtain output data containing the current level; then, the comprehensive evaluation layer is used to perform comprehensive analysis on the output data of each indicator, the weight of each indicator and the pre-processed data to obtain a comprehensive evaluation result including the drug situation level and drug distribution information of the monitored area; and, the abnormal analysis module is used to analyze the output data and the corresponding pre-processed data of each indicator in the drug situation analysis module whose level is greater than the corresponding preset threshold, and analyze the abnormal data of the drug situation in the monitored area to obtain an abnormal analysis report including information on drug-related personnel, key locations and hidden personnel; finally, the comprehensive evaluation results and the abnormal analysis report are visualized by the data display module, which can provide detailed and intuitive reports and data, solving the problem that the information query system provided by the anti-drug unit is queried in each system according to the elements, and the alarm monitoring, drug situation level assessment, and drug situation information viewing functions of each area cannot be realized in one system, saving the time of the anti-drug unit to obtain data from multiple systems and analyze and process them, thereby improving work efficiency.

[0034] In one embodiment of the present invention, the data analysis layer includes extracting input data of each indicator of the indicator layer from the preprocessed data according to preset labels, and storing it in the input list of the corresponding indicator; performing cluster analysis on the data not extracted from the preprocessed data to obtain classification labels; determining the correlation between the classification label and the indicator through expert scoring, and if the correlation is greater than the preset correlation, adding the classification label to the preset label list, and adding the preprocessed data corresponding to the classification label to the input data of the corresponding indicator.

[0035] Preferably, the preset labels are classification labels determined based on historical preprocessed data by adopting a clustering algorithm and expert evaluations, so as to obtain accurate input data for each indicator; and after the preprocessed data corresponding to the preset labels are extracted at the data analysis layer, a clustering method is used to perform cluster analysis on the unextracted preprocessed data to obtain new classification labels, and then the correlation between the classification labels and each indicator is determined by taking the average value of the expert scores, and the classification labels with a correlation greater than the preset correlation are added to the preset label list to enrich the content of the preset labels and improve the label types of the input data, so as to improve the accuracy of the analysis results of each indicator.

[0036] In this embodiment, in the data analysis layer, the present invention extracts input data of each indicator of the indicator layer from the preprocessed data based on preset labels; at the same time, cluster analysis is performed on the data not extracted from the preprocessed data to obtain classification labels of the preprocessed data, and the correlation between the classification labels and each indicator is determined by means of multiple expert scores to screen out classification labels with a correlation greater than a preset correlation, and the preprocessed data corresponding to the classification labels are added to the input data of the corresponding indicators, so as to analyze and process each indicator to obtain accurate output data, thereby ensuring that the comprehensive analysis layer and the abnormality analysis module output accurate results.

[0037] In one embodiment of the present invention, Figure 2 As shown, the indicator layer includes location indicators, drug-related personnel indicators, dangerous path indicators, drug indicators and sewage toxin content indicators. Among them, the analysis process of the location indicator includes counting the number of repetitions of repeated data in the input data of the location indicator, determining the location abnormality level according to the corresponding relationship between the number of repetitions and the preset interval, so as to focus on monitoring and surprise inspections of these locations, and displaying the distribution information of the above locations through the data display module so that the personnel of the drug control unit can view it intuitively.

[0038] The analysis process of drug-related personnel indicators includes determining the danger level and abnormal record information of drug-related personnel based on at least the illegal records and activity trajectories of drug-related personnel in the pre-processed data, so that subsequent staff can focus on monitoring and questioning the drug-related personnel and obtain more intelligence information. Among them, illegal records include criminal records and detention records.

[0039] The analysis process of the dangerous path indicator includes determining the drug transportation route and mode based on the drug-related personnel data and drug enforcement data in the pre-processed data, and determining the transportation level based on the relationship between the drug transportation route and the number of repetitions of the transportation mode and the threshold, so as to focus on checking the above-mentioned transportation mode at certain positions on the transportation route; and displaying the drug transportation information through the data display module, and the transportation information includes but is not limited to the transportation route, transportation mode, transportation license plate number, etc.

[0040] The analysis process of drug indicators includes determining the drug inflow level based on the relationship between the quantity and type of drugs seized in the pre-processing data and the corresponding preset thresholds, and visually displaying the drug data in the data display module. The analysis process of sewage toxin content indicators is to determine the toxin content level based on the relationship between the sewage toxin content data in the pre-processing data and the preset toxin content threshold.

[0041] In this embodiment, the corresponding relationship between the analysis data of the above-mentioned indicators and the preset thresholds and levels is a relationship determined by experts based on historical data analysis and stored in the database of the drug monitoring system. The present invention can obtain a comprehensive analysis of the drug level and the current drug status by analyzing the level and corresponding information of the above-mentioned indicators.

[0042] In one embodiment of the present invention, in the drug-related personnel index, the relationship between drug-related personnel is analyzed according to the drug-related personnel data in the preprocessed data, and a relationship network between drug-related personnel is established; the primary drug-related personnel whose number of relationships with drug-related personnel in the relationship network exceeds the first set threshold are screened, and the violation records of the primary drug-related personnel are obtained, and the danger level of the primary drug-related personnel is determined according to the number of violation records and the severity of the violation, and the information of the primary drug-related personnel whose danger level exceeds the preset danger level is used as abnormal record information. The danger level corresponding to a blank violation record is zero, and the danger level includes six levels, among which level 0 is the lowest and level 5 is the highest.

[0043] Preferably, a mind map is used to establish a relationship network among drug-related personnel so that it can be viewed intuitively in the drug monitoring system.

[0044] Furthermore, the abnormal record information includes at least criminal records, entry and exit records, dangerous driving path records, and records of abnormal entry and exit locations.

[0045] In this embodiment, the present invention screens out drug-related persons with complex relationship networks through their relationship networks, and marks their danger levels so that they can be observed, questioned or monitored in a focused manner to obtain more intelligence clues.

[0046] In one embodiment of the present invention, the comprehensive evaluation layer includes determining the weight of each indicator by expert scoring; performing weighted average calculation on the evaluation level and weight of each indicator to obtain the toxicity level; and forming a comprehensive evaluation result based on the preprocessed data, the toxicity level, and the output data of each indicator.

[0047] Preferably, the evaluation level of each indicator includes the location abnormality level, the danger level of drug-related personnel, the transportation level, the drug inflow level and the toxin content level.

[0048] Preferably, the weight of each indicator is determined by scoring by experts and averaging the scores, wherein the experts are experienced investigators in the field or professionals studying drug information. At the same time, averaging can avoid errors caused by personal factors.

[0049] In this embodiment, the output data of the toxicity level and each indicator are formed into a comprehensive report, and the pre-processed data after deduplication is embodied in the comprehensive evaluation result as an attachment, so that the detailed data of the comprehensive evaluation result can be clicked to view when needed.

[0050] In one embodiment of the present invention, in the abnormal analysis model, output data is obtained in which the level of each indicator in the indicator layer is greater than the corresponding preset threshold, and the output data includes information on drug-related persons, location abnormality information, drug inflow information and transportation path information; monitoring images at the location abnormality information and the transportation path are obtained; the monitoring images are identified using an image recognition model to determine whether there are drug-related persons, and if so, the monitoring images are processed to obtain the head portrait of the interacting person who interacts with the drug-related person in the monitoring image, determine that the interacting person is an initial hidden person, and connect to the identity authentication system to obtain the identity of the initial hidden person information; obtain the travel information of the initial hidden person according to the identity information of the initial hidden person, and judge whether there is abnormal travel data in the travel information. If there is abnormal travel data, determine that the initial hidden person is a hidden person; use the image recognition model to analyze the monitoring image, obtain the interactive personnel information of the hidden person, and judge whether there is abnormal information in the interactive personnel information of the hidden person. If so, determine that the person is a secondary hidden person, and establish a relationship library between the information of drug-related persons, hidden personnel information and secondary hidden personnel information, as well as key locations where the contact frequency exceeds the set frequency; form an abnormal analysis report based on the relationship library and key locations.

[0051] Preferably, the travel information includes but is not limited to entry and exit information, high-speed rail travel information, airplane travel information, bus travel information, taxi information, driving vehicle travel information, etc. When judging whether the interactive personnel information of the hidden person has abnormal information, it is mainly judged whether there is entry and exit information, information on entering and leaving key monitoring places, and illegal record information.

[0052] Preferably, the database of the drug monitoring system pre-stores the head portrait information of drug-related persons; or the identity authentication system is connected to obtain the head portrait information of drug users and drug dealers.

[0053] In this embodiment, the present invention first obtains output data in which the level of each indicator in the indicator layer is greater than the corresponding preset threshold value, and obtains the monitoring images that can be obtained at the abnormal position and on the transportation path according to the position abnormality information and the transportation path information in the output data; then, the image recognition model is used to identify and analyze the monitoring image to determine whether there are recorded drug-related persons in the monitoring image. If so, the monitoring image is processed to obtain the head portrait of the interactive person who communicates with the drug-related person in the monitoring image, determine that the interactive person is the initial hidden person, and connect to the identity authentication system to obtain the identity information of the initial hidden person, and then obtain the initial hidden person according to the identity information of the initial hidden person. The itinerary information of the drug-related personnel is used to determine whether there is abnormal itinerary data in the itinerary information. If there is abnormal itinerary data, the initial hidden personnel is determined as a hidden personnel; then the image recognition model is used again to perform the above analysis process on the hidden personnel to determine the secondary hidden personnel; and a relationship library between the drug-related personnel information, the hidden personnel information and the secondary hidden personnel information is established based on the drug-related personnel information, the hidden personnel information and the secondary hidden personnel information, as well as the key positions when the frequency of contact between these personnel exceeds the set frequency, so as to determine the main hidden personnel and key positions, so as to monitor the above personnel and key positions according to different levels and find intelligence clues related to the drug situation. Compared with the traditional method of manually analyzing surveillance images to determine suspects, the method adopted in this embodiment can save the time of staff in analyzing surveillance images and obtaining clues, thereby improving work efficiency.

[0054] In one embodiment of the present invention, the process of processing surveillance images by an image recognition model includes: obtaining surveillance images containing drug-related persons according to the itinerary of the drug-related persons, the surveillance images are images of set resolution, and storing the surveillance images in a first database in the order of itinerary; extracting avatar images of the interacting persons in multiple positions in the surveillance images, and performing image stitching and image enhancement on the avatar images in multiple positions to obtain frontal avatars of the interacting persons.

[0055] Preferably, the image resolution is set to the image with the lowest resolution that the image recognition model can recognize. For example, if the image recognition model requires the minimum image resolution to be 640×480 while ensuring accuracy, then the resolution is set to 640×480 to ensure that the minimum amount of data for monitoring images is obtained, thereby reducing the storage capacity of the database.

[0056] In this embodiment, the image recognition model can be a model built based on an attention mechanism and a convolutional neural network, or a model built based on machine learning. After successful training, it can realize the above-mentioned image recognition, image stitching and image enhancement functions to obtain a frontal portrait of the interacting person, so as to subsequently determine the identity information of the interacting person, and find the travel information, violation record, etc. of the interacting person.

[0057] In one embodiment of the present invention, the abnormal analysis module also includes: establishing a quantity prediction model using a convolutional neural network based on historical sewage toxin content data and corresponding historical drug addict data, analyzing the sewage toxin content data using the quantity prediction model to obtain a predicted number of drug addicts, and observing and analyzing the personnel in the abnormal analysis report based on the predicted number of drug addicts to determine suspected drug addicts.

[0058] Preferably, based on the historical sewage toxin content data of each sampling point in the monitoring area and the drug addict data recorded in the coverage area corresponding to the sampling points, an initial quantity prediction model constructed by a convolutional neural network is trained, and after the training is completed and passed the test, a quantity prediction model is obtained.

[0059] Furthermore, when the constructed quantity prediction model processes the sewage toxin content data, the sewage toxin content data of each sewage sampling point in the monitoring area is first obtained; then, based on the coverage of each sewage sampling point, the predicted number of drug addicts in the coverage area is estimated; and then the predicted number of drug addicts in the entire monitoring area is calculated.

[0060] In this embodiment, the drug addict data predicted by the quantity prediction model can be used as reference data when the staff determines drug addicts from drug-related persons, hidden persons and secondary hidden persons. If the number of drug addicts determined among these persons is too small compared with the predicted number, it indicates that there are still drug addicts hidden among these persons, and it is necessary to further observe these persons or conduct drug testing to accurately identify drug addicts.

[0061] In one embodiment of the present invention, the drug situation monitoring system also includes: when the drug situation level exceeds the preset drug situation level, the abnormal analysis report is analyzed using a strategy generation model established based on historical abnormal data and response measures to generate a response strategy, and an alarm message, abnormal analysis report and response strategy are sent to the monitoring personnel.

[0062] Preferably, the historical abnormal data includes data of drug-related personnel, drug data, drug enforcement data and sewage toxin content data with historical abnormalities. The response measures are conventional response plans corresponding to the historical abnormal data.

[0063] Furthermore, when the strategy generation model is used to process the historical abnormal data, the principal component analysis method is used to extract the key data in the historical abnormal data, and then the decision model based on the random forest is used to calculate the matching degree between the key data and each response plan, and the response plan with the highest matching degree is selected as the response strategy for the abnormal analysis report. When the drug monitoring system of the present invention detects that the drug level exceeds the preset drug level, the strategy generation model will match the best response plan according to the content of the abnormal analysis report for reference by drug control personnel.

[0064] In one embodiment of the present invention, the functions of the drug monitoring system of the present invention also include a task management module: the recorded tasks can be viewed, edited, deleted, searched or paused. Abnormal alarm module: the alarm information can be viewed, deleted, queried, the execution result query, and the corresponding task editing. The execution result query can view the alarm cause, and the corresponding task editing allows the user to edit the corresponding task content according to the alarm information. The system management module includes user management and operation log management. User management includes viewing, adding, resetting passwords, deleting, modifying, and querying user information; operation log management is to view the operation information of the user account.

[0065] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A drug monitoring system, characterized in that: include: A data collection module is used to obtain drug-related data in the monitoring area, wherein the drug-related data at least includes data on drug-related personnel, drug data, drug enforcement data, and sewage toxin content data; A data preprocessing module is used to extract data from drug-related data and convert the data format to obtain preprocessed data; The poison situation analysis module uses the data analysis layer, the indicator layer and the comprehensive evaluation layer to analyze the pre-processed data in sequence to obtain a comprehensive evaluation result, wherein the data analysis layer uses a clustering method to obtain the input data of each indicator of the indicator layer, the indicator layer is used to analyze each indicator according to the input data to obtain the output data of each indicator, and the comprehensive evaluation layer is used to perform a comprehensive analysis on the output data of each indicator, the weight of each indicator and the pre-processed data to obtain a comprehensive evaluation result, and the comprehensive evaluation result at least includes the poison situation level and poison situation distribution information of the monitoring area; The abnormal analysis module uses an abnormal analysis model to analyze the output data of each indicator in the drug situation analysis module that is greater than the corresponding preset threshold and the corresponding pre-processed data to obtain an abnormal analysis report, which at least includes information on drug-related personnel, key locations, and hidden personnel information; Data display module, used to visualize the comprehensive evaluation results and abnormal analysis reports; The data analysis layer includes extracting input data of each indicator of the indicator layer from the preprocessed data according to the preset labels, and storing them in the input list of the corresponding indicators; performing cluster analysis on the data not extracted from the preprocessed data to obtain classification labels; determining the correlation between the classification labels and the indicators through expert scoring, and if the correlation is greater than the preset correlation, adding the classification labels to the preset label list, and adding the preprocessed data corresponding to the classification labels to the input data of the corresponding indicators; The indicator layer includes location indicators, drug-related personnel indicators, dangerous path indicators, drug indicators and sewage toxin content indicators; wherein, the location indicators include counting the number of repetitions of repeated data in the input data of the location indicators, determining the location anomaly level according to the corresponding relationship between the number of repetitions and the preset interval, and displaying the location distribution information through the data display module; the drug-related personnel indicators include determining the danger level and abnormal record information of drug-related personnel at least according to the illegal records and activity trajectories of drug-related personnel in the pre-processed data; the dangerous path indicators include determining the drug transportation path and transportation method according to the drug-related personnel data and drug enforcement data in the pre-processed data, determining the transportation level according to the relationship between the number of repetitions of the drug transportation path and transportation method and the threshold, and displaying the drug transportation information in the data display module; the drug indicators include determining the drug inflow level according to the relationship between the number of drugs and the type of drugs seized in the pre-processed data and the corresponding preset threshold, and visually displaying the drug data in the data display module; the sewage toxin content indicators determine the toxin content level according to the relationship between the sewage toxin content data in the pre-processed data and the preset toxin content threshold; In the drug-related personnel index, the relationship between drug-related personnel is analyzed according to the drug-related personnel data in the preprocessed data, and a relationship network is established; primary drug-related personnel whose number of relationships exceeds a first set threshold are screened, and the violation records of the primary drug-related personnel are obtained, and the danger level of the primary drug-related personnel is determined according to the number of violation records and the severity of the violation, and the information of the primary drug-related personnel whose danger level exceeds the preset danger level is used as abnormal record information; The comprehensive evaluation layer includes determining the weight of each indicator by expert scoring; performing weighted average calculation on the evaluation level and weight of each indicator to obtain the toxicity level; and forming a comprehensive evaluation result based on the preprocessing data, the toxicity level, and the output data of each indicator; In the abnormal analysis model, output data whose level of each indicator in the indicator layer is greater than the corresponding preset threshold value is obtained, and the output data includes information on drug-related personnel, location abnormality information, drug inflow information and transportation path information; monitoring images at the location abnormality information and the transportation path are obtained; the monitoring images are identified by using an image recognition model to determine whether there are drug-related personnel, and if so, the monitoring images are processed to obtain the head portrait of the interacting person who interacts with the drug-related personnel in the monitoring image, determine that the interacting person is an initial hidden person, and connect to the identity authentication system to obtain the identity information of the initial hidden person; according to the initial The hidden person identity information obtains the travel information of the initial hidden person, and determines whether there is abnormal travel data in the travel information. If there is abnormal travel data, the initial hidden person is determined to be a hidden person; the image recognition model is used to analyze the monitoring image, and the interactive personnel information of the hidden person is obtained, and it is determined whether there is abnormal information in the interactive personnel information of the hidden person. If so, the hidden person is determined to be a secondary hidden person, and a relationship library between the drug-related personnel information, the hidden personnel information and the secondary hidden personnel information and the key positions where the contact frequency exceeds the set frequency are established; an abnormal analysis report is formed according to the relationship library and the key positions.

2. The drug monitoring system according to claim 1, characterized in that: The process of processing surveillance images by the image recognition model includes: obtaining surveillance images containing drug-related persons according to the itinerary of the drug-related persons, the surveillance images are images with a set resolution, and storing the surveillance images in a first database in the order of the itinerary; extracting avatar images of the interacting persons in multiple positions in the surveillance images, and performing image stitching and image enhancement on the avatar images in multiple positions to obtain frontal avatars of the interacting persons.

3. The drug monitoring system according to claim 1, characterized in that: The abnormal record information at least includes criminal records, entry and exit records, dangerous driving path records, and records of abnormal entry and exit locations.

4. The drug monitoring system according to claim 1, characterized in that: The abnormal analysis module also includes: using a convolutional neural network to establish a quantity prediction model based on historical sewage toxin content data and corresponding historical drug addict data, using the quantity prediction model to analyze the sewage toxin content data to obtain a predicted number of drug addicts, and observing and analyzing the personnel in the abnormal analysis report based on the predicted number of drug addicts to determine suspected drug addicts.

5. The drug monitoring system according to claim 4, characterized in that: Also includes: When the toxicity level exceeds the preset toxicity level, the abnormal analysis report is analyzed using a strategy generation model based on historical abnormal data and response measures to generate a response strategy, and an alarm message, abnormal analysis report and response strategy are sent to the monitoring personnel.

Images