CN118413404A - A security alarm system based on artificial intelligence - Google Patents

Abstract

The present invention relates to the field of intelligent security management technology, and discloses an artificial intelligence-based security alarm system, comprising a multi-dimensional acquisition module, an intelligent analysis module and a security alarm module. The artificial intelligence-based security alarm system generates an abnormal data set by calculating the multi-dimensional acquisition module. , clearly judge the target information as normal behavior, and then continue to screen the target information of suspicious behavior in the environmental data set according to the elimination method. The intelligent analysis module calculates and generates a regular data set , through artificial intelligence AI technology to flexibly respond to environmental variables, the intelligent analysis module calculates and generates the risk index and Vulnerability Index , quickly analyze the dangerousness of the target behavior, comprehensively evaluate the system's ability to judge suspicious behavior, the security alarm module judges the security status of the current environment, and implements corresponding security management measures, with high decision-making accuracy and strong anti-attack ability.

Description

A security alarm system based on artificial intelligence

Technical Field

The present invention relates to the technical field of intelligent security management, and in particular to a security alarm system based on artificial intelligence.

Background technique

Security management is a series of management measures to ensure that people, assets and information are in a safe state in a variety of environments such as residences, commercial buildings, and public places. Security management measures usually include access control, video surveillance, intrusion detection systems, and visitor management. The goal of security management is to prevent crime, reduce risks, respond quickly to emergencies, and protect people's safety and property from damage. With the development of technology, artificial intelligence (AI) technology has become an important driving force in the field of security management. The security alarm system based on artificial intelligence uses technologies such as machine learning, deep learning, computer vision, and natural language processing to improve the intelligence level of the security system, enabling it to more accurately identify threats, automate responses, and improve overall efficiency. The security alarm system based on artificial intelligence has the advantages of real-time monitoring, automatic analysis and response, continuous learning to reduce false alarms, resource optimization, and data prediction. The security alarm system based on artificial intelligence can monitor 24/7 without human intervention, reducing errors caused by human fatigue or negligence. The security alarm system based on artificial intelligence can automatically analyze data and respond quickly, shortening the time from threat detection to response. As more data is fed into the system, AI systems can continuously improve their performance through machine learning, reducing the number of false alarms through precise pattern recognition and anomaly detection. AI-based security alarm systems can help security personnel prioritize real threats, allocate resources more efficiently, and analyze historical data to predict and prevent future security incidents.

At present, traditional AI-based security alarm systems are vulnerable to adversarial attacks, such as deceiving AI models through specially designed inputs. The reduction in alertness and intervention capabilities leads to system failure, and there are deviations in the judgment of security status, which affects the accuracy of system decision-making, resulting in false alarms or missed alarms, and weak anti-attack capabilities.

Summary of the invention

In view of the shortcomings of the prior art, the present invention provides a security alarm system based on artificial intelligence, which has the advantages of flexible response to changing environmental factors, high decision-making accuracy and strong anti-attack ability. It solves the problem that traditional security alarm systems based on artificial intelligence are prone to failure after being attacked, and have low decision-making accuracy and weak anti-attack ability.

To achieve the above-mentioned object, the present invention provides the following technical solutions: an artificial intelligence-based security alarm system, comprising a multi-dimensional acquisition module, an intelligent analysis module and a security alarm module;

The multi-dimensional acquisition module includes an environment monitoring unit, a standard data unit and an abnormal data unit. The environment monitoring unit collects an environment data set in real time through a network connection to an environment monitoring device and transmits it to the abnormal data unit through the network. The standard data unit collects a standard data set through a network connection to a database and transmits it to the abnormal data unit through the network. The abnormal data unit calculates and generates an abnormal data set based on the environment data set and the standard data set. , the multi-dimensional acquisition module is connected to the intelligent analysis module via a network;

The intelligent analysis module includes a security analysis unit, an abnormality analysis unit and a comprehensive evaluation unit. The security analysis unit generates a regular data set based on the environmental data set. The abnormal analysis unit calculates and generates a risk index based on the environmental data set and the standard data set. and transmitted to the comprehensive evaluation unit through the network. The comprehensive evaluation unit performs the evaluation based on the environmental data set, the abnormal data set, and risk index , calculate and generate the vulnerability index , the intelligent analysis module is connected to the security alarm module through the network;

The security alarm module is based on the conventional data group , abnormal data group , Danger Index and Vulnerability Index , judge the security status of the current environment and implement corresponding security management measures.

Preferably, the environmental data set includes monitoring data collected in real time by an environmental monitoring device, and the expression of the environmental data set is: , Indicates the time point of monitoring data collection. Indicates the target information contained in the monitoring data. to Corresponding to the target information collected by the environmental monitoring device at each time point.

Preferably, the standard data set includes standard data in a database that defines normal behavior and abnormal behavior, and the expression of the standard data set is: , Represents standard data that defines normal behavior, Represents standard data that defines abnormal behavior, to Corresponding to each standard data that defines normal behavior and abnormal behavior.

Preferably, the abnormal data unit generates an abnormal data set by calculation based on the environmental data set and the standard data set. , the calculation process is as follows:

According to the standard data set, the normal behavior of the target in the environmental data set is filtered out and recorded as ,in, ;

formula, represents an abnormal data set, Represents the target information at all time points in the environment dataset, It indicates the target information belonging to normal behavior in the environmental data set filtered out according to all the standard data that define the normal state in the standard data set. It represents the abnormal data group calculated after excluding all target information belonging to normal behavior in the environmental data set.

Preferably, the security analysis unit calculates and generates a conventional data set based on the environmental data set. , the calculation process is as follows:

The security analysis unit is provided with a convolutional neural network algorithm model, and sequentially analyzes the target information at each time point in the environmental data set through the convolutional neural network algorithm model;

formula, Represents a regular data group, The fixed algorithm model that defines the target information as a normal state in the calculation formula of the regular data group is the convolutional neural network algorithm model. It means that the target information of each time point in the environmental data set is analyzed in turn through the convolutional neural network algorithm model to calculate the regular data group.

Preferably, the abnormality analysis unit calculates and generates a risk index based on the environmental data set and the standard data set. , the calculation process is as follows:

According to the standard data set, the abnormal behavior of the target in the environmental data set is screened out, which is recorded as ,in, ;

formula, Indicates the risk index. It indicates the target information belonging to abnormal behavior in the environmental data set filtered out according to all the standard data defining abnormal behavior in the standard data set. The ratio of target information belonging to abnormal behavior in the environmental data set to standard data defining abnormal behavior in the environmental data set is the danger index.

Preferably, the comprehensive evaluation unit is based on the environmental data set, the abnormal data set and risk index , calculate and generate the vulnerability index , and its calculation formula is as follows: formula, represents the vulnerability index, Represents the target information at all time points in the environment dataset, It indicates the target information belonging to normal behavior in the environmental data set filtered out according to all the standard data defining normal behavior in the standard data set. It indicates the target information belonging to abnormal behavior in the environmental data set filtered out according to all the standard data defining abnormal behavior in the standard data set. It means that after excluding all target information belonging to normal and abnormal behaviors in the environmental data set, target information whose behavior nature cannot be identified is obtained. The ratio of target information whose behavior nature cannot be identified to the target information at all time points in the environmental dataset is the vulnerability index.

Preferably, the abnormal data set When there is data in the security alarm module, the security alarm module determines that the target behavior in the current environment is suspicious, stops the target's suspicious behavior in time and intercepts and asks for detailed information.

Preferably, the conventional data set The data present in the dataset and the normal behavior of the target in the environment dataset If the target behavior is the same, the security alarm module determines that the target behavior in the current environment is normal and allows the target to continue passing. Normal behavior of the target in some data and environment data sets If they are different, the security alarm module will not allow the target to pass and will notify security personnel for manual review.

Preferably, the risk index When the preset value is exceeded, the security alarm module determines that the current environment security is threatened, does not allow all targets to continue to pass, and promptly alarms the security personnel for emergency management. When the preset value is exceeded, the security alarm module determines that the current system is under adversarial attack, does not allow all targets to continue passing, and promptly alarms the security personnel to conduct emergency management.

Compared with the prior art, the present invention provides a security alarm system based on artificial intelligence, which has the following beneficial effects:

1. The present invention is provided with an environment monitoring unit, a standard data unit and an abnormal data unit through a multi-dimensional acquisition module. The environment monitoring unit collects the environment data set in real time. After the standard data unit collects the standard data set, the abnormal data unit calculates and generates the abnormal data set according to the environment data set and the standard data set. , clearly judge the target information as normal behavior, and then continue to screen the target information of suspicious behavior in the environmental data set according to the elimination method. The intelligent analysis module is equipped with a security analysis unit, an abnormal analysis unit and a comprehensive evaluation unit. The security analysis unit calculates and generates a regular data set based on the environmental data set. Through the convolutional neural network algorithm model in artificial intelligence (AI) technology, it can deeply judge the suspicious behavior of targets in environmental data sets and flexibly respond to changing environmental factors.

2. The present invention calculates and generates a risk index based on the environmental data set and the standard data set through an abnormal analysis unit. , clearly judge the target information as abnormal behavior, and then continue to screen the target information of suspicious behavior in the environmental data set according to the elimination method, quickly analyze the danger of the target behavior in the current environment, and facilitate the subsequent implementation of corresponding security management measures based on the value size. The comprehensive evaluation unit and risk index , calculate and generate the vulnerability index The comprehensive evaluation system can judge suspicious behaviors, digitally evaluate the system's anti-attack ability, and timely discover the system's judgment deviation, effectively improving the reliability of security management. The security alarm module can be used to collect regular data. , abnormal data group , Danger Index and Vulnerability Index , judge the security status of the current environment and implement corresponding security management measures. It has high decision-making accuracy and strong anti-attack ability, effectively identifies dangers and abnormal situations, and can more effectively protect personal and property safety at critical moments.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the system flow of the present invention.

Detailed ways

The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

Since the traditional security alarm system based on artificial intelligence is vulnerable to adversarial attacks, such as deceiving the AI model through specially designed inputs, the decrease of alertness and intervention ability leads to the failure of system functions, and there are deviations when judging the security status, which affects the accuracy of system decision-making, generates false alarms or missed alarms, and has weak anti-attack capabilities, therefore, a security alarm system based on artificial intelligence is provided, please refer to Figure 1, the system includes a multi-dimensional acquisition module, an intelligent analysis module and a security alarm module;

The multi-dimensional acquisition module includes an environmental monitoring unit, a standard data unit, and an abnormal data unit. The environmental monitoring unit collects environmental data sets in real time through a network connection to an environmental monitoring device, and transmits the data to the abnormal data unit through the network. The environmental data set includes the monitoring data collected in real time by the environmental monitoring device. The expression of the environmental data set is: , Indicates the time point of monitoring data collection, and marks the time point of monitoring data in real time, ensuring the timeliness and effectiveness of security management, and the subsequent analysis has strong traceability. Indicates the target information contained in the monitoring data. to Corresponding to the target information collected by the environmental monitoring device at each time point, the targets in the monitoring data can be set to faces, vehicles, creatures and objects. In the actual application process, the management personnel may limit the target range according to the actual security needs;

The standard data unit collects the standard data set through the network connection database and transmits it to the abnormal data unit through the network. The standard data set includes the standard data that defines normal behavior and abnormal behavior in the database. The expression of the standard data set is , Represents standard data that defines normal behavior, where a target moves along a regular route. Indicates standard data that defines abnormal behavior, which includes moving along unconventional routes or staying at a certain location for a long time. to Corresponding to each standard data that defines normal and abnormal behaviors, combined with 24-hour uninterrupted monitoring, effectively avoids the system's mistakes of false alarms or missed alarms;

The abnormal data unit calculates and generates the abnormal data set based on the environmental data set and the standard data set. , the calculation process is as follows:

According to the standard data set, the normal behavior of the target in the environmental data set is filtered out and recorded as ,in, , clearly judge the target information as normal behavior, so as to facilitate the subsequent screening of the target information of suspicious behavior in the environmental data set according to the elimination method; formula, represents an abnormal data set, Represents the target information at all time points in the environment dataset, It indicates the target information belonging to normal behavior in the environmental data set filtered out according to all the standard data that define the normal state in the standard data set. It means that after excluding all target information belonging to normal behavior in the environment data set, the abnormal data group is calculated. , quickly screen out the target information of suspicious behaviors in the environmental data set, so as to facilitate the subsequent in-depth judgment of the safety of the target behavior through artificial intelligence (AI) technology;

The multi-dimensional acquisition module is connected to the intelligent analysis module through the network;

The intelligent analysis module includes a security analysis unit, an abnormality analysis unit and a comprehensive evaluation unit. The security analysis unit calculates and generates a regular data set based on the environmental data set. , the calculation process is as follows:

The security analysis unit is provided with a convolutional neural network algorithm model, and sequentially analyzes the target information at each time point in the environmental data set through the convolutional neural network algorithm model;

formula, Represents a regular data group, The fixed algorithm model that defines the target information as a normal state in the calculation formula of the regular data group is the convolutional neural network algorithm model. It means that the target information of each time point in the environmental data set is analyzed in sequence through the convolutional neural network algorithm model, and the conventional data group is calculated. , through the convolutional neural network algorithm model in artificial intelligence AI technology, it can deeply judge the suspicious behavior of the target in the environmental data set and flexibly respond to the changing factors of the environment;

The abnormal analysis unit calculates and generates the danger index based on the environmental data set and the standard data set. , and transmitted to the comprehensive evaluation unit through the network. The calculation process is as follows:

According to the standard data set, the abnormal behavior of the target in the environmental data set is screened out, which is recorded as ,in, ,

Clearly identify the target information that is abnormal, so that it is easier to continue screening the target information of suspicious behavior in the environmental data set by the elimination method; formula, Indicates the risk index. It indicates the target information belonging to abnormal behavior in the environmental data set filtered out according to all the standard data defining abnormal behavior in the standard data set. The ratio of target information belonging to abnormal behavior in the environmental data set to the standard data defining abnormal behavior in the environmental data set is the danger index. , quickly analyze the dangerousness of the target behavior in the current environment, so as to facilitate the subsequent implementation of corresponding security management measures based on the numerical value;

The comprehensive evaluation unit is based on the environmental data set, abnormal data set and risk index , calculate and generate the vulnerability index , and its calculation formula is as follows: formula, represents the vulnerability index, Represents the target information at all time points in the environment dataset, It indicates the target information belonging to normal behavior in the environmental data set filtered out according to all the standard data defining normal behavior in the standard data set. It indicates the target information belonging to abnormal behavior in the environmental data set filtered out according to all the standard data defining abnormal behavior in the standard data set. It means that after excluding all target information belonging to normal and abnormal behaviors in the environmental data set, target information whose behavior nature cannot be identified is obtained. The ratio of the target information that cannot identify the nature of the behavior to the target information at all time points in the environmental data set is the vulnerability index. , the comprehensive evaluation system's ability to judge suspicious behaviors, the digital evaluation system's ability to resist attacks, timely discover the system's judgment deviations, and effectively improve the reliability of security management;

The intelligent analysis module is connected to the security alarm module through the network;

The security alarm module is based on the conventional data group , abnormal data group , Danger Index and Vulnerability Index , judge the security status of the current environment, abnormal data group When there is data in the security alarm module, the security alarm module determines that the target behavior in the current environment is suspicious, stops the target's suspicious behavior in time and intercepts the detailed information. The data present in the environment dataset and the normal behavior of the target If the target behavior is normal, the security alarm module determines that the target behavior is normal in the current environment and allows the target to continue to pass. Normal behavior of the target in some data and environment data sets If they are different, the security alarm module will not allow the target to pass and notify the security personnel to conduct manual review. When the preset value is exceeded, the security alarm module determines that the current environment security is threatened, does not allow all targets to continue to pass, and promptly alarms the security personnel for emergency management. When the preset value is exceeded, the security alarm module determines that the current system is under adversarial attack and does not allow all targets to continue passing. It promptly alarms and notifies security personnel to conduct emergency management. The decision-making accuracy is high and the anti-attack ability is strong. It can effectively identify dangers and abnormal situations, and can more effectively protect personal and property safety at critical moments.

Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims (10)

1. A security alarm system based on artificial intelligence, characterized by: comprising a multi-dimensional acquisition module, an intelligent analysis module and a security alarm module; The multi-dimensional acquisition module includes an environment monitoring unit, a standard data unit and an abnormal data unit. The environment monitoring unit collects an environment data set in real time through a network connection to an environment monitoring device and transmits it to the abnormal data unit through the network. The standard data unit collects a standard data set through a network connection to a database and transmits it to the abnormal data unit through the network. The abnormal data unit calculates and generates an abnormal data set based on the environment data set and the standard data set. , the multi-dimensional acquisition module is connected to the intelligent analysis module via a network; The intelligent analysis module includes a security analysis unit, an abnormality analysis unit and a comprehensive evaluation unit. The security analysis unit generates a regular data set based on the environmental data set. The abnormal analysis unit calculates and generates a risk index based on the environmental data set and the standard data set. and transmitted to the comprehensive evaluation unit through the network. The comprehensive evaluation unit performs the evaluation based on the environmental data set, the abnormal data set, Danger Index , calculate and generate the vulnerability index , the intelligent analysis module is connected to the security alarm module through the network; The security alarm module is based on the conventional data group , abnormal data group , Danger Index and Vulnerability Index , judge the security status of the current environment and implement corresponding security management measures. 2. The artificial intelligence-based security alarm system according to claim 1, characterized in that: the environmental data set includes monitoring data collected in real time by the environmental monitoring device, and the expression of the environmental data set is , Indicates the time point of monitoring data collection. Indicates the target information contained in the monitoring data. to Corresponding to the target information collected by the environmental monitoring device at each time point. 3. The artificial intelligence-based security alarm system according to claim 2, characterized in that: the standard data set includes standard data in the database that defines normal behavior and abnormal behavior, and the expression of the standard data set is , Represents standard data that defines normal behavior, Represents standard data that defines abnormal behavior, to Corresponding to each type of standard data that defines normal behavior and abnormal behavior. 4. The artificial intelligence-based security alarm system according to claim 3, characterized in that: the abnormal data unit calculates and generates the abnormal data set according to the environmental data set and the standard data set. , the calculation process is as follows: According to the standard data set, the normal behavior of the target in the environmental data set is filtered out and recorded as ,in, ; formula, represents an abnormal data set, Represents the target information at all time points in the environment dataset, It indicates the target information belonging to normal behavior in the environmental data set filtered out according to all the standard data that define the normal state in the standard data set. It represents the abnormal data group calculated after excluding all target information belonging to normal behavior in the environmental data set. 5. The artificial intelligence-based security alarm system according to claim 4, characterized in that: the security analysis unit generates a conventional data set by calculating the environmental data set. , the calculation process is as follows: The security analysis unit is provided with a convolutional neural network algorithm model, and sequentially analyzes the target information at each time point in the environmental data set through the convolutional neural network algorithm model; formula, Represents a regular data group, The fixed algorithm model that defines the target information as a normal state in the calculation formula of the regular data group is the convolutional neural network algorithm model. It means that the target information of each time point in the environmental data set is analyzed in turn through the convolutional neural network algorithm model to calculate the regular data group. 6. The artificial intelligence-based security alarm system according to claim 5, characterized in that: the abnormal analysis unit calculates and generates a danger index based on the environmental data set and the standard data set. , the calculation process is as follows: According to the standard data set, the abnormal behavior of the target in the environmental data set is screened out, which is recorded as ,in, ; formula, Indicates the risk index. It indicates the target information belonging to abnormal behavior in the environmental data set filtered out according to all the standard data defining abnormal behavior in the standard data set. The ratio of target information belonging to abnormal behavior in the environmental data set to standard data defining abnormal behavior in the environmental data set is the danger index. 7. The artificial intelligence-based security alarm system according to claim 6 is characterized in that: the comprehensive evaluation unit and risk index , calculate and generate the vulnerability index , and its calculation formula is as follows: formula, represents the vulnerability index, Represents the target information at all time points in the environment dataset, It indicates the target information belonging to normal behavior in the environmental data set filtered out according to all the standard data defining normal behavior in the standard data set. It indicates the target information belonging to abnormal behavior in the environmental data set filtered out according to all the standard data defining abnormal behavior in the standard data set. It means that after excluding all target information belonging to normal and abnormal behaviors in the environmental data set, target information whose behavior nature cannot be identified is obtained. The ratio of target information whose behavior nature cannot be identified to the target information at all time points in the environmental dataset is the vulnerability index. 8. The artificial intelligence-based security alarm system according to claim 4, characterized in that: the abnormal data set When there is data in the security alarm module, the security alarm module determines that the target behavior in the current environment is suspicious, stops the target's suspicious behavior in time and intercepts and asks for detailed information. 9. The artificial intelligence-based security alarm system according to claim 5, characterized in that: the conventional data set The data present in the dataset and the normal behavior of the target in the environment dataset If the target behavior is the same, the security alarm module determines that the target behavior in the current environment is normal and allows the target to continue passing. Normal behavior of the target in some data and environment data sets If they are different, the security alarm module will not allow the target to pass and will notify security personnel for manual review. 10. The artificial intelligence-based security alarm system according to claim 7, characterized in that: the danger index When the preset value is exceeded, the security alarm module determines that the current environment security is threatened, does not allow all targets to continue to pass, and promptly alarms the security personnel for emergency management. When the preset value is exceeded, the security alarm module determines that the current system is under adversarial attack, does not allow all targets to continue passing, and promptly alarms the security personnel to conduct emergency management.