Civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data

Abstract

The invention relates to a civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data, and the method comprises the following steps: constructing a risk assessment index system which comprises a first-level index and a second-level index, the first-level index comprises 7 dimensions, and the second-level index comprises 29 risk indexes; determining weights of a first-level index and a second-level index in the risk assessment index system; for each secondary index in the risk assessment index system, constructing a secondary index situation library and performing quantification, and calculating a risk coefficient by using a set value statistical method; and selecting a trapezoidal distribution membership function to calculate a corresponding membership value, and establishing a civil aviation passenger risk comprehensive evaluation model through multi-stage fuzzy evaluation. The civil aviation passenger risk comprehensive assessment model established by the invention has good feasibility and applicability, can be used as an auxiliary method for risk assessment and grading of civil aviation passengers in an actual scene of civil aviation operation, and provides theoretical decision support for civil aviation safety risk management.

Description

Technical Field

[0001] The present invention relates to the technical field of passenger risk assessment, and in particular to a civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data. Background Art

[0002] Civil aviation transport is an important part of the transportation industry. As a distribution center for cargo and passengers transported by civil aviation, airports have always been burdened with high-load security work and are at the forefront of facing risks. After decades of rapid development, my country's civil aviation transport industry has become a popular way of travel. The contradiction between the rapidly rising volume of civil aviation transport and the relatively backward civil aviation security protection capabilities has become increasingly prominent. It can be seen that it is urgent to strengthen the security level of civil aviation airports, which is related to the safety of the lives of the majority of passengers and the security and stability of the country. Civil aviation airports are important places for passengers and crew members to enter and exit aircraft. At present, my country's civil aviation security inspections mainly focus on isolated security inspection strategies of "objects" or "people", and there is no "equal emphasis on people" security inspection model. Given the inadequacy of airport security management capabilities and the fact that front-line security personnel are working under high pressure for a long time, over-reliance on traditional manual security methods can easily lead to security loopholes. Therefore, we use technical means such as big data and facial recognition to integrate data from civil aviation public security, airports, airlines, CAAC and other civil aviation units, conduct pre-flight analysis and judgment of passengers, and conduct research on dynamic risk assessment of the entire process within the airport. This has deepened the application of civil aviation big data scenarios, can effectively prevent typical security risk events, and help all parties at the airport to quickly, promptly and accurately take risk control measures. It also plays an important role in maintaining a safe, efficient and orderly airport operating environment, strengthening airport security prevention and control capabilities, and enhancing passenger confidence and satisfaction. Summary of the invention

[0003] The technical problem to be solved by the present invention is to overcome the deficiencies in the prior art and provide a civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data.

[0004] The present invention is achieved through the following technical solutions:

[0005] A civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data comprises the following steps:

[0006] S1. Construct a risk assessment indicator system, which includes primary indicators and secondary indicators. The primary indicators include 7 dimensions, and the secondary indicators include 29 risk indicators.

[0007] S2. Determine the weights of the primary and secondary indicators in the risk assessment indicator system;

[0008] S3. For each secondary indicator in the risk assessment indicator system, a secondary indicator scenario library is constructed and quantified, and the risk coefficient is calculated using the set value statistics method;

[0009] S4. Select the trapezoidal distribution membership function to calculate its corresponding membership value, and establish a comprehensive assessment model for civil aviation passenger risks through multi-level fuzzy evaluation.

[0010] According to the above technical solution, preferably, before step S1, the risk factors of civil aviation passengers are identified, and systematically analyzed and screened to establish a hazard source database, and a risk assessment indicator system is constructed based on the hazard source database.

[0011] According to the above technical solution, preferably, in step S1, the first-level indicators include social environment, passenger identity, passenger background, abnormal behavior, location, checked baggage, and carry-on baggage; the second-level indicators include sensitive periods, sensitive stopovers, long routes, age, gender, occupation, ethnicity, place of residence, key personnel, public opinion monitoring, abnormal vehicles, historical travelers, abnormal insurance personnel, abnormal consumers, climbing over isolation facilities, terminal buildings close to sensitive areas, frequent reconnaissance, violence in the area, close to sensitive areas in the cabin, importance of airport areas, regional population density, abnormal liquids, abnormal gases, abnormal solids, other checked items, carry-on controlled devices, carry-on flammable and explosive materials, human door alarms, metal door alarms, and trace equipment alarms.

[0012] According to the above technical solution, preferably, in step S2, the group hierarchy analysis method is used to determine the weights of the primary indicators and secondary indicators in the risk assessment indicator system; the group hierarchy analysis method includes establishing a hierarchical structure, constructing an expert judgment matrix, consistency testing, and determining the comprehensive weight of indicators.

[0013] According to the above technical solution, preferably, in step S3, an expert group decision method is adopted to introduce interval numbers to quantify the risk indicator situation, and a risk assessment interval matrix is ​​obtained;

[0014] Among them, for n situations and m experts under the kth secondary risk indicator, the risk assessment matrix is ​​defined as:

[0015]

[0016] In the formula is the interval score given by expert p to situation i under the kth secondary risk indicator.

[0017] According to the above technical solution, preferably, in step S3, the quantitative risk coefficient of situation i under the kth secondary indicator is synthesized by using set value statistics method:

[0018]

[0019] The m×n matrix U is compressed into a one-dimensional row vector U * =[u 1 ,u 2 , ..., u n ].

[0020] According to the above technical solution, preferably, step S4 includes: determining an indicator set and a comment set; using a trapezoidal distribution membership function to construct a membership function corresponding to the level of the indicator; and establishing a comprehensive assessment model for civil aviation passenger risks through multi-level fuzzy evaluation.

[0021] According to the above technical solution, preferably, in step S4, the comment set V = {v1, v2, v3, v4, v5}, representing low risk, lower risk, medium risk, higher risk, and high risk respectively, and the median of the interval is set for each level, V = {10, 30, 50, 70, 90}.

[0022] According to the above technical solution, preferably, in step S4, the membership function of the corresponding level of the indicator is constructed as follows:

[0023]

[0024]

[0025]

[0026]

[0027]

[0028] The beneficial effects of the present invention are:

[0029] The present invention provides a method that takes passengers as the core, dynamically monitors the pre-flight link, terminal entry link (gate link), check-in link, passenger security link, and on-board link of passengers throughout the entire process, and analyzes and extracts multi-dimensional information of passengers from the acquired multi-source heterogeneous intelligence data to conduct dynamic risk assessment. The comprehensive risk assessment model for civil aviation passengers established by the present invention has good feasibility and applicability, and can be used as an auxiliary method for risk assessment and classification of civil aviation passengers in actual civil aviation operation scenarios, providing theoretical decision support for civil aviation safety risk management. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is the comprehensive risk assessment index system for civil aviation passengers in the present invention.

[0031] Figure 2 It is a schematic diagram of the standard value of the average random consistency index RI in the consistency test process of the group hierarchical analysis method of the present invention.

[0032] Figure 3 It is a membership function diagram of the risk index in the present invention relative to each risk level.

[0033] Figure 4 It is a schematic diagram of the calculation process of the single factor evaluation vector in the present invention. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and the best embodiment. Based on the embodiments in the invention, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the invention.

[0035] As shown in the figure, the present invention comprises the following steps:

[0036] S1. Construct a risk assessment indicator system, wherein the risk assessment indicator system includes primary indicators and secondary indicators. The primary indicators include 7 dimensions, and the secondary indicators include 29 risk indicators.

[0037] Before step S1, a comprehensive method of statistical analysis, expert investigation and laws and regulations is used to identify the risk sources of civil aviation passengers in the pre-flight stage and in each stage of entering the terminal, including the pre-flight stage, gate stage, check-in stage, security check stage and on-board stage. By identifying the risk factors of civil aviation passengers and systematically analyzing and screening them, a risk source database is established, and a risk assessment indicator system is constructed based on the risk source database.

[0038] Based on statistical analysis and expert survey methods, risk sources are identified in the pre-flight and gate stages. Using the Aviation Safety Network (ASN), the Global Terrorism Database (GTD) and dangerous event records in search engines as data sources, statistical analysis is conducted on the temporal and spatial distribution and patterns of events, the characteristics of perpetrators, criminal motives, and tools used to commit crimes. In addition, correlation analysis is conducted to include criminal records, occupations, ethnicity, native places, registered permanent residences, destinations, and religious beliefs as risk factors for passengers. Based on surveys of some airports and airport public security units and expert consultations, it is recommended to integrate multi-source data, including information from shopping platforms, search engines, social platforms, China Aviation Information Network, insurance companies, financial institutions, and public security departments, and include passengers' online shopping, searches, public opinions, ticket purchase information, blacklists, credit records, large insurance, historical travel records, criminal records, etc. as risk factors.

[0039] Based on laws and regulations, we identify risk sources in the check-in and security check stages. Based on the specific contents of passenger luggage in the "Public Security Administration Punishment Law of the People's Republic of China", "Civil Aviation Security Protection Regulations of the People's Republic of China", "Civil Aviation Security Inspection Rules", and "List of Restricted Carry-on or Checked Items for Civil Aviation Passengers", we systematically and comprehensively summarize risk factors including but not limited to firearms, explosive or incendiary materials and devices, controlled devices, and self-igniting and flammable dangerous goods.

[0040] Identification of risk sources in all links based on laws, regulations and expert investigation. According to the prohibited behaviors in the Civil Aviation Security Regulations of the People's Republic of China and the survey and consultation with the airport security department, the risk factors of passengers' behaviors in the airport and aircraft are sorted out, including but not limited to fighting, alcoholism, provoking trouble, forcibly boarding aircraft, occupying seats or luggage compartments, falsely reporting dangers, creating chaos, etc. In addition, when assessing airport environmental risks, aircraft, airports, and aviation facilities are the first to be considered. According to the Civil Aviation Security Regulations of the People's Republic of China, Aviation Security Rules for Civil Aviation Transport Airports CCAR-329, Security Work Rules for Public Air Passenger Transport Flights CCAR-332-R1, and Aviation Security Rules for Public Air Transport Enterprises CCAR-343-R1, regional risk factors such as aircraft, airport control areas, key areas, and non-control areas are identified, including but not limited to waiting isolation areas, aircraft activity and maintenance areas, catering work areas and storage areas, aircraft supply storage areas, airport terminals, parking lots, etc.

[0041] After completing the risk source identification of the entire process of passengers' pre-flight stage, entering the terminal (gate stage), check-in stage, security check stage, and on-board stage, the identified risk factors are systematically analyzed and screened to establish a risk source database suitable for civil aviation passenger risk assessment. Afterwards, based on the risk source identification of passengers from the pre-flight stage to the boarding stage, the identified risk factors are systematically analyzed, integrated and evaluated, the impact of risk source characteristics on passenger risks is studied, and typical common characteristic indicators reflecting passenger risk levels are abstracted from risk source intelligence. In this way, a comprehensive risk assessment indicator system for civil aviation passengers is constructed, which includes 7 dimensions and a total of 29 indicators, such as Figure 1 shown.

[0042] S2. Determine the weights of the primary and secondary indicators in the risk assessment indicator system. In this example, the group hierarchical analysis method is preferably used, including establishing a hierarchical structure, constructing an expert judgment matrix, consistency testing, and determining the comprehensive weight of indicators. Specifically:

[0043] S21. Establish a hierarchical structure: This hierarchical structure is the passenger risk assessment indicator system constructed in the previous article.

[0044] S22. Construct an expert judgment matrix: Invite experts in the field of civil aviation safety to compare the relative importance of indicators based on the 1-9 scale method to obtain a pairwise comparison matrix.

[0045]

[0046] Among them, the scale value 1 means "indicator i is equally important as indicator j", the scale value 3 means "indicator i is slightly more important than indicator j", the scale value 5 means "indicator i is obviously more important than indicator j", the scale value 7 means "indicator i is strongly more important than indicator j", the scale value 9 means "indicator i is extremely more important than indicator j", the scale values ​​2, 4, 6, and 8 mean "the importance of indicator i to indicator j is between the above", and the scale values ​​1, 1 / 2, ..., 1 / 9 mean "the opposite of the above situation".

[0047] S23. Consistency check: Due to the subjectivity of expert judgment, the judgment matrix may be inconsistent. Through consistency check (CR value is between 0 and 1, 0<CR<0.1 indicates the accepted consistency level), subjective bias can be identified and corrected to ensure the logical rationality of the judgment. This process is completed by calculating the consistency ratio CR. The steps for calculating the consistency ratio of the judgment matrix are as follows:

[0048] (a) Calculation of consistency index:

[0049] CI=(λ max -n) / (n-1)

[0050] Where n is the order of the judgment matrix, λ max is the maximum eigenvalue of the judgment matrix A.

[0051] (b) Calculate the consistency ratio:

[0052] CR=CI / RI

[0053] Where RI is the average random consistency index, and the standard value is as follows Figure 2 .

[0054] (c) Determine consistency: If the CR of A satisfies 0≤CR<0.1, A is considered consistent. Otherwise, the matrix elements need to be adjusted to improve the consistency.

[0055] S24. Determine the comprehensive weight of indicators: After ensuring the consistency of the expert judgment matrix, the weight vector can be solved by precise calculation (such as characteristic root method, power method) or approximate calculation (such as square root method, sum product method). In this example, the characteristic root method is preferably used to determine the indicator weight, which specifically includes the following steps:

[0056] (a) Use the eigenvalue method to find the maximum eigenvalue of the judgment matrix and its corresponding maximum normalized eigenvector, that is, Aλ = λ maxw i , find the maximum eigenvalue λ max and its forward normalized eigenvector w i =(ω i1 ,ω i2 , ..., ω in ), corresponding to the weights of each sub-indicator contained in the i-th indicator.

[0057] (b) After calculating the weights of each level of indicators that meet the consistency requirements of each expert through the above steps, use the Hadamard product clustering and take the geometric mean to calculate the comprehensive weight, that is:

[0058]

[0059] Where p is the number of experts participating in the evaluation, and after normalization, the comprehensive weight of the indicator W is obtained: i .

[0060] Assuming that there are m indicators under the target layer, and the kth first-level indicator contains n second-level indicators, after the above calculation process, the weight set of indicators at all levels is obtained as follows:

[0061] Target layer weight set: W = (W 1 , W 2 , ..., W m )

[0062] First-level indicator weight set: W k =(W k1 , W k2 , ..., W kn )

[0063] S3. For each secondary indicator in the risk assessment indicator system, a secondary indicator scenario library is constructed and quantified, and the risk coefficient is calculated using the set value statistics method.

[0064] In order to systematically manage and evaluate risk indicators under different situations, it is necessary to collect all possible situation data associated with each secondary risk indicator in the comprehensive risk assessment indicator system for civil aviation passengers and build a secondary indicator situation library. The following methods and approaches are used to collect data to ensure the comprehensiveness, meticulousness, updateability and operability of the situation library. For example: historical data analysis, extracting high-frequency risk situations and causes from the airport's previous operating data, risk event reports and abnormal situation records; expert discussion and review, inviting experts in the fields of aviation safety, airport management, risk management, etc., combining professional knowledge and practical experience to explore possible risk situations and accumulate relevant data; questionnaire surveys, using questionnaire surveys and interviews to collect risk situation data from airport staff, passengers and other relevant parties; new security technology, on the one hand, uses security technology equipment such as video surveillance, gates, biometrics, and security scanners deployed at airports to monitor key links in real time and automatically collect situation data; on the other hand, it uses big data technology to mine potential risk situations from various data sources (such as analyzing social media, customer feedback, flight dynamics and other data to discover passenger behavior patterns and potential hidden dangers).

[0065] The classified and organized situation data will be stored in the situation database system using standardized templates. Each situation will contain the situation ID, situation name, situation description, associated secondary indicators, and risk factor fields. In addition, the situation database will have the ability to be dynamically updated, and will be able to regularly supplement and revise data to ensure that new situations and risk factors are included in a timely manner. Through this systematic approach, not only can various risk situations be effectively integrated and evaluated, providing a reliable basis for scientific decision-making, but it will also help achieve dynamic management and the ability to respond to changes.

[0066] After building the risk index database for civil aviation passengers, it is necessary to quantify the risk index to qualitatively describe the risk level of the situation. In this case, the expert group decision method is preferably used to introduce interval numbers to quantify the risk index situation. Experts assign a fuzzy interval score to each indicator situation to obtain a risk evaluation interval matrix. This method can effectively avoid the subjective bias caused by a single person's preference and experience, and improve the reliability and accuracy of the scoring results.

[0067] Among them, for n situations and m experts under the kth secondary risk indicator, the risk assessment matrix is ​​defined as:

[0068]

[0069] In the formula is the interval score given by expert p to situation i under the kth secondary risk indicator.

[0070] It should be emphasized that each secondary indicator is a qualitative indicator that abstractly represents a certain type of risk attribute. It contains various specific situations, which can be understood as the relationship between class and entity. All possible situations under each secondary indicator are collected through various channels, and a situation database of passenger risk indicators is established. Experts assign interval values ​​to all situations under each secondary indicator to quantify the degree of risk. When a passenger has a certain risk situation, the quantitative results given by each expert represent the risk coefficient of the secondary indicator corresponding to this situation. In this way, the mapping relationship between the passenger indicator and the single factor evaluation vector in the fuzzy comprehensive evaluation is established.

[0071] Afterwards, the set value statistics method is used to synthesize the quantitative risk coefficient of situation i under the kth secondary indicator by m experts:

[0072]

[0073] The m×n matrix U is compressed into a one-dimensional row vector U * =[u 1 ,u 2 , ..., u n ].

[0074] S4. Select the trapezoidal distribution membership function to calculate its corresponding membership value, and establish a comprehensive assessment model for civil aviation passenger risks through multi-level fuzzy evaluation.

[0075] First, we need to determine the indicator set and comment set. According to the hierarchical structure of the civil aviation passenger comprehensive risk assessment indicator system, the indicators at each level are combined into an indicator set as shown below:

[0076] R={E,I,B,A,S,C,H},

[0077] in:

[0078] E={E1,E2,E3}

[0079] I={I1,I2,I3,I4}

[0080] B={B1, ​​B2, B3, B4, B5, B6}

[0081] A={A1,A2,A3,A4,A5}

[0082] S={S1,S2}

[0083] C={C1,C2,C3,C4}

[0084] H={H1,H2,H3,H4,H5}

[0085] In the fuzzy comprehensive evaluation, the comment set has also been improved, that is, fuzziness is introduced, and the single value corresponding to the original comment is changed to a corresponding interval value. Before conducting a comprehensive evaluation, the characteristics of indicators at all levels are comprehensively considered, and combined with the knowledge and experience of experts in the field, the risk level is divided into five levels, namely V = {v1, v2, v3, v4, v5}, representing low risk, relatively low risk, medium risk, relatively high risk, and high risk, respectively. The median of the interval is set for each level, V = {10, 30, 50, 70, 90}.

[0086] After determining the comment set, it is necessary to construct the membership function of the indicator. The trapezoidal membership function has the advantages of being simple, flexible, easy to calculate, and able to well represent fuzzy boundaries. It has been widely used in many engineering practices. Therefore, the trapezoidal distribution membership function is used in this example. Based on the previously determined comment set and combined with the characteristics of the risk indicator and the actual situation, the membership function of the corresponding level of the indicator is constructed as follows:

[0087]

[0088]

[0089]

[0090]

[0091]

[0092] Traditional fuzzy comprehensive evaluation methods often use a single fuzzy logic operation and a unified evaluation framework. The imbalance of fuzzy operators and indicator weights leads to deviations from reality, which limits the accuracy and adaptability of the evaluation results in civil aviation passenger risk assessment. In view of the limitations of traditional fuzzy comprehensive evaluation methods, this paper proposes a new evaluation scheme suitable for civil aviation passenger risk assessment. This scheme has two key innovations: first, different fuzzy operators are used at each level of multi-level fuzzy comprehensive evaluation; second, a strategy for accurately processing evaluation vectors is introduced. In the evaluation matrix at each level, the single factor evaluation vector of the indicator with the largest weight does not participate in the operation when it is mapped from the default membership degree.

[0093] (a) Fuzzy operator:

[0094] In fuzzy evaluation, after the evaluation matrix and weight vector are determined, it is necessary to select a suitable fuzzy operator for synthesis operation. There are four commonly used fuzzy operators:

[0095] M(∧,∨): small or large operator

[0096] The operation rules are: j =max{min(W i , R ij)|1<i<n}(j=1, 2,...,m)

[0097] M(·, V): multiplication to obtain the largest operator

[0098] The operation rules are: j =max{(W i ·R ij )|1<i<n}(j=1, 2,...,m)

[0099] Take the small sum operator

[0100] The operation rules are:

[0101] Weighted Average Operator

[0102] The operation rules are:

[0103] In this example, different fuzzy operators are selected to consider the characteristics and requirements of different levels. The first-level fuzzy comprehensive evaluation uses the weighted average operator, and the second-level fuzzy comprehensive evaluation uses the multiplication operator.

[0104] (b) Single factor evaluation matrix:

[0105] like Figure 4 As shown in the figure, for the risk indicator situations existing in the obtained passenger structured intelligence information, first match them with the situations in the situation library. If the match is successful, the comprehensive risk coefficient corresponding to this situation is substituted into the constructed membership function, the membership of the passenger risk situation is calculated, and these memberships are mapped to the single factor evaluation vector of the secondary indicator. If the match fails, that is, the passenger does not have this secondary indicator situation, the set default membership is mapped to the single factor evaluation vector.

[0106] Assume that a certain indicator set contains t sub-indicators, according to Figure 4 The calculation process obtains the single factor evaluation vector corresponding to each sub-indicator:

[0107] r i =[r i1 , r i2 ,…,r i5 ],i=1,…,t

[0108] Then the single factor evaluation matrix of the indicator set is:

[0109]

[0110] (c) Comprehensive assessment:

[0111] First-level fuzzy comprehensive evaluation:

[0112]

[0113] Among them, if the weight set W k The single factor evaluation vector corresponding to the indicator with the largest weight is mapped by the default membership degree and does not participate in the fuzzy operation.

[0114] Secondary fuzzy comprehensive evaluation:

[0115] R=(B 1 , …, B m ) T

[0116]

[0117] Similarly, if the single factor evaluation vector corresponding to the indicator with the largest weight in the weight set W is mapped by the default membership degree, it does not participate in the fuzzy operation. Represents a fuzzy operator.

[0118] Finally, the passenger risk level is determined according to the principle of maximum membership.

[0119] The present invention provides a method that takes passengers as the core, dynamically monitors the pre-flight link, terminal entry link (gate link), check-in link, passenger security link, and on-board link of passengers throughout the entire process, and analyzes and extracts multi-dimensional information of passengers from the acquired multi-source heterogeneous intelligence data to conduct dynamic risk assessment. The comprehensive risk assessment model for civil aviation passengers established by the present invention has good feasibility and applicability, and can be used as an auxiliary method for risk assessment and classification of civil aviation passengers in actual civil aviation operation scenarios, providing theoretical decision support for civil aviation safety risk management.

[0120] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data, characterized in that: The steps include: S1. Construct a risk assessment indicator system, which includes primary indicators and secondary indicators. The primary indicators include 7 dimensions, and the secondary indicators include 29 risk indicators. S2. Determine the weights of the primary and secondary indicators in the risk assessment indicator system; S3. For each secondary indicator in the risk assessment indicator system, a secondary indicator scenario library is constructed and quantified, and the risk coefficient is calculated using the set value statistics method; S4. Select the trapezoidal distribution membership function to calculate its corresponding membership value, and establish a comprehensive assessment model for civil aviation passenger risks through multi-level fuzzy evaluation.

2. According to claim 1, a civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data is characterized in that: Before step S1, the risk factors of civil aviation passengers are identified, systematically analyzed and screened, a hazard source database is established, and a risk assessment indicator system is constructed based on the hazard source database.

3. According to claim 1, a civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data is characterized in that: In step S1, the first-level indicators include social environment, passenger identity, passenger background, abnormal behavior, location, checked baggage, and carry-on baggage; The secondary indicators include sensitive periods, sensitive stopovers, long routes, age, gender, occupation, ethnicity, place of residence, key personnel, public opinion monitoring, abnormal vehicles, historical travel personnel, abnormal insurance personnel, abnormal consumer personnel, climbing over isolation facilities, terminal buildings close to sensitive areas, frequent reconnaissance, violence in the area, proximity to sensitive areas in the cabin, importance of airport areas, regional population density, abnormal liquids, abnormal gases, abnormal solids, other checked items, carry-on controlled devices, carry-on flammable and explosive materials, human door alarms, metal door alarms, and trace equipment alarms.

4. According to claim 3, a civil aviation passenger risk assessment method based on multi-dimensional and multi-modal data is characterized in that: In step S2, the weights of the primary and secondary indicators in the risk assessment indicator system are determined using the group analytic hierarchy process; The group hierarchical analysis method includes establishing a hierarchical structure, constructing an expert judgment matrix, consistency testing, and determining the comprehensive weight of indicators.

5. The method for assessing civil aviation passenger risk based on multi-dimensional and multi-modal data according to claim 4, characterized in that: In step S3, an expert group decision method is adopted to introduce interval numbers to quantify the risk indicator situation and obtain a risk assessment interval matrix; Among them, for n situations and m experts under the kth secondary risk indicator, the risk assessment matrix is ​​defined as: In the formula is the interval score given by expert p to situation i under the kth secondary risk indicator.

6. The method for assessing civil aviation passenger risk based on multi-dimensional and multi-modal data according to claim 5, characterized in that: In step S3, the set value statistics method is used to synthesize the quantitative risk coefficient of situation i under the kth secondary indicator by m experts: Among them, the 6×8 matrix U is compressed into a one-dimensional row vector U * =[u1,u2,…,u n ].

7. The method for assessing civil aviation passenger risk based on multi-dimensional and multi-modal data according to claim 6, characterized in that: Step S4 includes: determining an indicator set and a comment set; using a trapezoidal distribution membership function to construct a membership function corresponding to the level of the indicator; and establishing a comprehensive assessment model for civil aviation passenger risks through multi-level fuzzy evaluation.

8. The method for assessing civil aviation passenger risk based on multi-dimensional and multi-modal data according to claim 7, characterized in that: In step S4, the comment set V = {v1, v2, v3, v4, v5}, representing low risk, relatively low risk, medium risk, relatively high risk, and high risk respectively, and the median of the interval is set for each level, V = {10, 30, 50, 70, 90}.

9. The method for assessing civil aviation passenger risk based on multi-dimensional and multi-modal data according to claim 8, characterized in that: In step S4, the membership function of the corresponding level of the indicator is constructed as:

Images