Building safety monitoring method and system based on artificial intelligence

Abstract

The invention belongs to the field of building safety monitoring, and particularly relates to a building safety monitoring method and system based on artificial intelligence. The information acquisition module is used for acquiring image data of internal and external environments of a building; the identification processing module comprises a resident database and an appearance identification model; a data bar in the resident database comprises resident information, resident face capacity and resident address; the appearance recognition model is used for recognizing and marking character features in the image data as appearance features, and querying a data bar of which the face of a resident is similar to the face feature of the appearance features in a resident database as a similar data bar; the data bar with the resident address the same as the resident address of the similar data bar is inquired in the resident database to serve as the similar family member data bar, and the recognition module is used for taking the image data with the same appearance features as the same-flow image data. According to the scheme, the problem that the viewing efficiency of the monitoring picture is low due to the fact that the monitoring picture is analyzed and judged by means of security personnel in the monitoring room is solved.

Description

Technical Field

[0001] This solution belongs to the field of building security monitoring, and specifically involves an artificial intelligence-based building security monitoring method and system. Background Art

[0002] With the development of modern society, the process of urbanization is accelerating, and high-rise buildings have become an important part of the city. Building safety issues have also become increasingly prominent and have become the focus of people's attention. In order to protect the lives and property of residents and park owners, building monitoring systems have emerged and become an indispensable part of modern intelligent buildings. Building monitoring systems use cameras installed in key locations to achieve real-time monitoring of the internal and external environment of the building, so as to detect abnormal situations in time and take corresponding measures. However, these monitoring images are often limited to the display screen in the monitoring room, and they need to rely on the security personnel in the monitoring room to analyze and judge the monitoring images with their naked eyes to identify whether there are abnormalities in the park. This method is very inefficient, and the labor cost of security personnel is extremely high. When the security personnel are tired, distracted, or not on duty, this method cannot accurately identify and judge the abnormal situations in the monitoring. Under the influence of these factors, the monitoring effect is not ideal, and as the number of monitoring equipment increases, the cost and difficulty of manual monitoring are also increasing. Summary of the invention

[0003] The purpose of this solution is to provide a building security monitoring method and system based on artificial intelligence to solve the problem of low efficiency in viewing monitoring images by relying on security personnel in the monitoring room to analyze and judge the monitoring images.

[0004] In order to achieve the above objectives, this solution provides a building security monitoring system based on artificial intelligence, including: Information acquisition module: used to collect image data of the environment inside and outside the building; Identification processing module: including a household database and an appearance recognition model, wherein the data strips in the household database include household information, household faces and household addresses; the appearance recognition model is used to identify and mark the human features in the image data as appearance features, query the household database for data strips with similar facial features to the appearance features as similar data strips, query the household database for data strips with the same household addresses as the similar data strips as similar family members data strips, and the identification module is used to treat image data with the same appearance features as the same-stream image data; Communication transmission module: used to send the same-stream image data to the communication address in the household information of similar family data strips; Resident terminal: used to receive and play the same-stream image data transmitted by the communication transmission module, used to enter the appearance feature correction information during the playback process, and send the appearance feature correction information to the communication transmission module after the playback ends; The communication transmission module receives the appearance feature correction information and transmits it to the recognition processing module. When the recognition processing module receives the appearance feature correction information, it processes the appearance feature correction into appearance feature correction data and trains the appearance recognition model.

[0005] And, an artificial intelligence-based building security monitoring method using an artificial intelligence-based building security monitoring system.

[0006] The principle and technical effect of this solution are as follows: First, this solution can quickly distinguish between residents in the building and outsiders by using the appearance recognition model to perform feature recognition and analysis on the collected image data. The appearance features collected by the appearance recognition model are compared with the faces of the residents in the resident database to determine the resident information corresponding to the appearance features, and the image information corresponding to the resident information is sent to the resident's family members, so that the family members can pay attention to the safety of the residents, thereby reducing the workload of security personnel in analyzing and judging the monitoring screen and improving the viewing efficiency of monitoring (image data). Secondly, the family members of the residents are already aware of the dynamics and safety of the residents. Sending the image information corresponding to the residents to the family members of the residents can enable the family members of the residents to know and pay attention to the situation of the residents in a timely manner. After the family members of the residents view the image information, the dangerous situation of the residents can be discovered in time, and the needs of the family members of the residents to obtain the status information of the residents in time can also be met in time, which improves the living experience of the residents and their families. Furthermore, after viewing the image information, the residents' family members can manually check the marked appearance features, enter the appearance information to correct the appearance features that do not match the residents, and the recognition processing module will process the appearance feature correction information into appearance feature correction data after receiving it and use it to train the appearance recognition model. This method allows the model to continuously learn and adapt to new situations and changes, and continuously improve the accuracy and efficiency of recognition. With the passage of time and the accumulation of more feedback from residents, the model's ability to recognize appearance in complex situations such as different lighting, angles, and expressions will become stronger and stronger, ensuring that the system is always in a good operating and monitoring state.

[0007] In summary: this solution solves the problem of low efficiency in viewing surveillance images, which is caused by relying on security personnel in the surveillance room to analyze and judge the surveillance images.

[0008] Furthermore, when the recognition processing module is used to query family data strips by appearance features, if no family data strips are found, the security monitoring terminal is used as the receiving and judging party of the image information, and the communication transmission module sends the corresponding same-stream image information to the security monitoring terminal.

[0009] Furthermore, when the recognition processing module searches for matching faces of residents in the household database through appearance features, if no matching faces of residents that meet the appearance features are found, an interception instruction is generated based on the appearance features and image data, and the interception instruction is passed to the communication transmission module. After receiving the interception instruction, the communication transmission module sends the interception instruction and its corresponding appearance features and image data to the security monitoring terminal. If the security monitoring terminal receives the interception instruction and checks the appearance features and image data, and then enters and sends appearance feature correction information to the communication transmission module, the communication transmission module passes the appearance feature correction information to the recognition processing module. The recognition processing module processes the appearance feature correction information and uses it to train the appearance recognition model.

[0010] When the recognition processing module cannot find a matching resident face in the resident database based on the appearance features, it means that there is a suspicious person who is likely to be an outsider and unidentified. At this time, the interception instruction is generated and transmitted to the security monitoring terminal, so that the security personnel can receive timely notifications and take corresponding interception actions, take the initiative to deal with potential security threats, and avoid suspicious persons from moving around in the building at will and causing security problems, such as preventing entry and sabotage. If the security personnel find that there may be recognition errors after checking the appearance features and image data and enter the appearance feature correction information, the information will be passed back for training the appearance recognition model. This allows the model to continuously learn from the feedback of actual manual judgment and make up for its own possible recognition defects, such as misjudgment caused by poor lighting, special angles and other factors. With continuous correction training, the model's subsequent recognition accuracy of appearance features will gradually improve, better serving the building security monitoring work.

[0011] Furthermore, the communication transmission module is used to send the same-stream image data to the communication address in the resident information of the similar family data strip, and is also used to receive the security authentication sent back by the resident terminal corresponding to the communication address; after the resident terminal receives and views the same-stream image data, if the security authentication is not entered before closing the same-stream image data, then after closing the same-stream image data, the security authentication is automatically generated according to the same-stream data and sent to the communication transmission module; if the security authentication is entered before closing the same-stream image data, then after closing the same-stream image data, the security authentication will be generated and sent to the communication transmission module.

[0012] Furthermore, when the communication transmission module sends the same-stream image data to the resident terminal, if the received security authentication result contains a dangerous security authentication result, the same-stream image data will be sent to the resident terminal again, and a security verification prompt will be sent to the resident terminal at the same time. After receiving the security verification prompt, the resident terminal will play the security verification prompt and the same-stream image data; if the resident terminal checks the same-stream image data and the security verification prompt, and the security authentication sent to the communication transmission module contains a dangerous security authentication result, the communication transmission module will send the same-stream image data as security warning data to the security monitoring terminal.

[0013] Resident terminals are required to send security authentication after viewing the same-stream image data, prompting residents to carefully examine the relevant situations in the image and actively participate in the confirmation of building safety. Whether it is automatically generated or manually entered, residents pay more attention to the activities of themselves and their families in the building, enhancing their sense of responsibility and participation as a member of building safety. When the communication transmission module receives a dangerous security authentication result, it sends the same-stream image data again with a reminder of security verification, providing residents with an opportunity to check again and carefully identify the situation, avoiding residents from giving inaccurate feedback due to momentary negligence or misoperation, helping to ensure the accuracy of residents' judgment of safety situations, and further ensuring the rigor of safety information transmission and confirmation. After receiving the security warning data, the security monitoring terminal can accurately deploy security forces according to the specific situation, focus limited human and material resources on areas with safety risks and personnel involved, avoid blindly patrolling and checking the entire building, and improve the efficiency and pertinence of security work in responding to security threats.

[0014] Furthermore, the recognition processing module is used to search for matching resident faces in the resident database through appearance features, and is also used to recognize the environment in the image data, recognize environmental facilities according to the object contours in the image data, determine the moving speed of the resident according to the positional relationship of the environmental facilities in time combined with the shooting interval of the image data, and classify the resident whose moving speed exceeds the preset running speed as a running resident; the facial features of the running resident are detected by facial key points to identify the emotional features of the running resident, and the emotional features include happiness, sadness, anger, surprise, disgust, fear, and neutrality; when the emotional features of the running resident are sadness, anger or fear When a running resident is found, it is determined whether the running resident is a minor by judging the resident information of the running resident. If the running residents are all adults, the communication transmission module sends a danger warning to the security monitoring terminal; if there are minors among the running residents, it is determined whether the resident addresses of the running residents are the same. If the resident addresses are the same, the communication transmission module sends a safety reminder to the communication address of the resident address; if the resident addresses are different, it is determined whether there are adults among the running residents. If there are no adults, a child fighting reminder is sent to the communication address of the running resident address; if the resident addresses are different and there are adults among the running residents, a danger warning is sent to the security monitoring terminal.

[0015] By combining appearance features and facial key point detection to identify the emotional characteristics of running residents, and making subsequent judgments based on different emotions, it is no longer limited to focusing on the appearance of behavior, but deeply explores the safety risks that may be contained behind the behavior. When the emotional characteristics are sadness, anger or fear, it means that the running behavior may not be normal, and it is likely to indicate that a dangerous situation is happening or is about to happen, so that potential safety hazards can be discovered more accurately, and corresponding prompts or warnings can be issued in advance, which enhances the accuracy and timeliness of warnings for abnormal situations in the building. According to whether the running resident is a minor and the resident's address, different prompts are sent to the security monitoring terminal or the communication address of the relevant resident address, realizing the pertinence of risk prompts. For adults running under abnormal emotions, sending a danger warning prompt to the security monitoring terminal can enable security personnel to quickly intervene in the investigation and disposal to prevent possible dangerous events from further developing; and for the fights or potential dangerous situations of minors in different scenarios, appropriate prompts are sent to the corresponding residents, so that family members can pay attention and deal with them in time to avoid accidents causing harm to minors, and overall improve the effectiveness of dealing with different risk situations.

[0016] Furthermore, when the recognition module recognizes that the emotional characteristics of the running resident are sadness, anger or fear, it also recognizes the running animal through the contour characteristics, and determines whether a rope is connected between the running animal and the running resident. If no rope is connected, the communication transmission module issues a danger warning to the security monitoring terminal; if a rope is connected, it waits for a preset interval time to determine whether the rope length is shortened. If not shortened, the communication transmission module issues a danger warning to the security monitoring terminal.

[0017] By judging whether a rope is connected between the running animal and the running resident and the change in the length of the rope, scenes with different degrees of danger can be accurately identified. If the rope is not connected, the animal may be out of control and easily cause harm to other residents, public facilities, etc.; even if the rope is connected but the length of the rope is not shortened, it may also mean that there are some abnormal situations, such as the animal pulling uncontrollably. Through such detailed judgment conditions, situations with real safety hazards can be accurately located, improving the accuracy of potential danger judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a schematic diagram of the structure of the artificial intelligence-based building security monitoring system in Example 1 of this solution.

[0019] Figure 2 This is a related judgment flow chart when the recognition module in the second embodiment of this scheme recognizes a running resident.

[0020] Figure 3 This is a related judgment flow chart when a running animal is identified in the third embodiment of this scheme. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the concept and technical effects of the present invention in combination with the embodiments, so as to fully understand the purpose, features and effects of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention: Embodiment 1 like Figure 1 As shown, a building security monitoring system based on artificial intelligence includes: Information acquisition module: used to collect image data of the environment inside and outside the building; Identification and processing module: includes a household database and an appearance recognition model. The data strips in the household database include household information, household faces and household addresses. The appearance recognition model is used to identify and mark the human features in the image data as appearance features, query the household database for data strips with similar facial features to the appearance features as similar data strips, query the household database for data strips with the same household addresses as the similar data strips as similar family data strips. The identification module is used to treat image data with the same appearance features as the same-stream image data.

[0022] The communication transmission module is used to send the same-stream image data to the communication address in the household information of similar family data strips.

[0023] The resident terminal is used to receive and play the same-stream image data transmitted by the communication transmission module, to input appearance feature correction information during the playback process, and to send the appearance feature correction information to the communication transmission module after the playback is completed.

[0024] Among them, the communication transmission module receives the appearance feature correction information and transmits it to the recognition processing module. When the recognition processing module receives the appearance feature correction information, it processes the appearance feature correction into appearance feature correction data and trains the appearance recognition model.

[0025] This solution also provides an artificial intelligence-based building security monitoring method, and the artificial intelligence-based building security monitoring method uses an artificial intelligence-based building security monitoring system.

[0026] When the recognition processing module is used to query the family data strip by appearance features, if no family data strip is found, the security monitoring terminal is used as the receiving and judging party of the image information, and the communication transmission module sends the corresponding same-stream image information to the security monitoring terminal.

[0027] The recognition processing module is used to query and match the face of the resident in the resident database through appearance features. If the face of the resident that meets the appearance features is not found, an interception instruction is generated according to the appearance features and image data, and the interception instruction is passed to the communication transmission module. After receiving the interception instruction, the communication transmission module sends the interception instruction and its corresponding appearance features and image data to the security monitoring terminal. If the security monitoring terminal receives the interception instruction and checks the appearance features and image data, and then enters and sends the appearance feature correction information to the communication transmission module, the communication transmission module passes the appearance feature correction information to the recognition processing module. The recognition processing module processes the appearance feature correction information and uses it to train the appearance recognition model.

[0028] The communication transmission module is used to send the same-stream image data to the communication address in the resident information of the similar family data strip, and is also used to receive the security authentication sent back by the resident terminal corresponding to the communication address; after the resident terminal receives and views the same-stream image data, if the security authentication is not entered before closing the same-stream image data, the security authentication will be automatically generated according to the same-stream data after closing the same-stream image data and sent to the communication transmission module; if the security authentication is entered before closing the same-stream image data, the security authentication will be entered and sent to the communication transmission module after closing the same-stream image data.

[0029] When the communication transmission module sends the same-stream image data to the resident terminal, if the received security authentication result contains a dangerous security authentication result, the same-stream image data will be sent to the resident terminal again, and a security verification prompt will be sent to the resident terminal at the same time. After receiving the security verification prompt, the resident terminal will play the security verification prompt and the same-stream image data; if the resident terminal checks the same-stream image data and the security verification prompt, and the security authentication sent to the communication transmission module contains a dangerous security authentication result, the communication transmission module will send the same-stream image data as security warning data to the security monitoring terminal.

[0030] In specific use, the apartment is used as the collection environment of the artificial intelligence-based building security monitoring system, the property management personnel are used as the users and managers of the security monitoring terminals, Mr. Li is used as the resident of the apartment, and the courier is used as the non-apartment resident.

[0031] The property management staff installed a large number of high-definition cameras in the community, which are distributed in key locations such as the entrance of the apartment building, corridors, elevators, parking lots, and community gardens. As part of the information collection module, these cameras continuously collect image data of the environment inside and outside the building. For example, one morning, the camera at the entrance captured an image of the owner, Mr. Li, leaving for work.

[0032] The appearance recognition model in the recognition processing module starts to work, analyzes Mr. Li's image, and marks Mr. Li's character features as appearance features. Then, a query is made in the resident database. Each data in the resident database contains resident information (such as name, contact information, etc.), resident face and resident address. Through comparison, data strips similar to Mr. Li's facial features are found, and similar family data strips with the same address are confirmed (because Mr. Li's family members are also in the database). Image data containing Mr. Li's appearance features are defined as co-stream image data. After the communication transmission module obtains Mr. Li's co-stream image data, it sends these data to the resident terminal corresponding to the communication address registered in his family information (such as the mobile phone of Mr. Li's family). Mr. Li's family receives and plays the co-stream image data through a special application (resident terminal) on their mobile phone. During the viewing process, if it is found that there is a slight deviation in Mr. Li's appearance recognition in the image (for example, due to the problem of sunlight exposure angle, part of the facial shadow affects the recognition), the family can enter the appearance feature correction information in the terminal. After the viewing is completed, these correction information will be sent back to the recognition processing module through the communication transmission module. After receiving the appearance feature correction information, the recognition processing module processes it into appearance feature correction data, which is further used to train the appearance recognition model to make the model more accurate in subsequent recognition.

[0033] One day, a courier came to deliver a package, but he was not a resident of the community. The information collection module collected the image data of the courier waiting for the recipient at the door of the apartment building. The appearance recognition model could not query the family data in the resident database through his appearance features. At this time, the communication transmission module sent the courier's image information to the security monitoring terminal used by the property management staff. After the property management staff saw the courier's image on the terminal in the monitoring room, they could continue to observe his behavior through monitoring to ensure that he did not behave abnormally during the waiting process. When the recipient came to pick up the package, the property management staff could also confirm whether the entire handover process was safe and orderly.

[0034] Once, the communication transmission module sent the same-stream image data (probably an image of his children playing in the community) to the mobile phone (resident terminal) of the owner Mr. Li. Mr. Li checked it and did not enter the security authentication before closing the image data. The mobile phone automatically generated a security authentication based on the same-stream data and sent it to the communication transmission module.

[0035] Embodiment 2 The difference between this embodiment and the first embodiment is that: Figure 2As shown, the recognition processing module is used to query and match the face of the resident in the resident database through appearance features, and is also used to identify the environment in the image data, identify the environmental facilities according to the object contours in the image data, determine the moving speed of the resident according to the positional relationship of the environmental facilities combined with the shooting interval of the image data, and define the resident whose moving speed exceeds the preset running speed (generally 2m / s, which can be set by the recognition processing module) as a running resident. The facial features of the running resident are detected by facial key points to identify the emotional characteristics of the running resident, and the emotional characteristics include happiness, sadness, anger, surprise, disgust, fear, and neutrality. When the emotional characteristics of the running resident are sadness, anger or fear, it is determined whether the running resident is a minor by judging the resident information of the running resident. If the running residents are all adults, the communication transmission module sends a danger warning to the security monitoring terminal. If there are minors among the running residents, it is determined whether the household addresses of the running residents are the same. If the household addresses are the same, the communication transmission module will send a safety reminder to the communication address of the household address; if the household addresses are different, it is determined whether there are adults among the running residents. If there are no adults, a reminder of children fighting will be sent to the communication address of the running residents' address; if the resident addresses are different and there are adults among the running residents, a danger warning will be sent to the security monitoring terminal.

[0036] In specific use, the apartment is used as the collection environment of the artificial intelligence-based building security monitoring system, the property management personnel are used as the users and managers of the security monitoring terminal, and the male resident A and the child are used as the residents of the apartment.

[0037] In one scenario, the information collection module collected image data in the corridor. When analyzing these image data, the recognition processing module not only searches for matching resident faces in the resident database through appearance features, but also recognizes the environment. According to the outline of the objects in the image, the corridor walls, stair handrails and other environmental facilities are identified. By analyzing the positional relationship of these environmental facilities at different times and combining the shooting interval of the image data, it is found that the movement speed of a male resident A exceeds the preset running speed (2m / s), so he is marked as a running resident. Then, by detecting the facial key points of the running resident's facial features, it is identified that his emotional feature is anger. The system further determines that the resident is an adult through the resident information. In this case, the communication transmission module quickly sends a danger warning to the security monitoring terminal. After receiving the warning, the security personnel immediately went to the corridor to check the situation and found that the male resident A was running because of a fierce quarrel with his neighbor and was emotional. The security personnel mediated in time to avoid further escalation of the conflict and ensure the safety and order in the corridor.

[0038] In another scenario, the monitoring system captured an image of a group of running children. The recognition processing module determined that they were running residents, and found through detection that some of the children had sad emotions (the children were crying), and they were all minors. Further analysis found that the residential addresses of these children were different, and there were no adults among them. At this time, the communication transmission module sent a reminder of children fighting to the communication addresses (parents' mobile phones) of the respective residential addresses of these children. After receiving the reminder, parents can pay attention to the children's situation, or tell their children to be careful not to affect other residents when playing.

[0039] In another scenario, the information collection module collected an image of a group of people running. After analysis, the recognition processing module found that there were both adults and minors among the running residents, and their addresses were different. Through facial key point detection, it was found that the emotional characteristics of some adults and minors were fear. In this case, the communication transmission module issued a danger warning to the security monitoring terminal. The security personnel quickly rushed to the underground parking lot. After questioning, they learned that a vehicle was driving too fast in the parking lot and did not pay attention to pedestrians, causing everyone to run in panic. The security personnel took timely measures to stop the dangerous driving behavior and comforted the residents, preventing possible safety accidents.

[0040] Embodiment 3 The difference between this embodiment and the first and second embodiments is that: Figure 3 As shown, when the recognition module recognizes that the emotional characteristics of the running resident are sadness, anger or fear, it also recognizes the running animal through the contour characteristics, and determines whether a rope is connected between the running animal and the running resident. If no rope is connected, the communication transmission module sends a danger warning to the security monitoring terminal; if a rope is connected, wait for a preset interval time (generally the preset interval time is set to 1 minute, which can be specifically set by the recognition processing module) to determine whether the rope length is shortened. If not shortened, the communication transmission module sends a danger warning to the security monitoring terminal.

[0041] In specific use, the apartment is used as the collection environment of the artificial intelligence-based building security monitoring system, and the property management personnel are used as users and managers of the security monitoring terminal.

[0042] In one scene, the recognition processing module identified a running resident when analyzing the image, and found that his emotional characteristics were fear through facial key point detection. At the same time, the system also identified a running large dog nearby through contour features. Further judgment found that there was no rope connecting the running animal and the running resident. In this case, the communication transmission module quickly issued a danger warning to the security monitoring terminal.

[0043] In another scenario, the information acquisition module collected image data, and the recognition processing module identified a running resident whose emotional characteristics were angry. At the same time, a running pet dog was found next to him. It was confirmed that a rope was connected between the pet dog and the resident. After the system waited for the preset interval time (the preset interval time was set to 1 minute), it analyzed the image again and found that the length of the rope had not shortened. This indicates that there may be some abnormal conditions, such as the pet dog may be pulling the resident hard or showing uncontrolled behavior. Therefore, the communication transmission module issued a danger warning to the security monitoring terminal. The security personnel immediately went to the leisure square and found that the pet dog was stimulated by other pets in the square and became extremely excited. It was trying to rush towards other pets, and the residents were unable to hold it back. The security personnel helped the residents control the pet dog, preventing possible conflicts between pets and potential dangers to surrounding residents.

[0044] In another scenario, a resident was jogging on the track in the community in the morning. He was exercising with his pet dog, and a rope was connected to the two. The recognition and processing module identified the behavior of the resident and the pet dog, and found that the length of the rope changed normally. The pet dog and the resident were in a normal state of exercise, and there were no signs of danger. The system did not issue an alarm, and the security monitoring in the community continued to operate normally, ensuring that the safety of other residents was not affected.

[0045] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. The building security monitoring system based on artificial intelligence is characterized by: include: Information acquisition module: used to collect image data of the environment inside and outside the building; Identification processing module: including a household database and an appearance recognition model, wherein the data strips in the household database include household information, household faces and household addresses; the appearance recognition model is used to identify and mark the human features in the image data as appearance features, query the household database for data strips with similar facial features to the appearance features as similar data strips, query the household database for data strips with the same household addresses as the similar data strips as similar family members data strips, and the identification module is used to treat image data with the same appearance features as the same-stream image data; Communication transmission module: used to send the same-stream image data to the communication address in the household information of similar family data strips; Resident terminal: used to receive and play the same-stream image data transmitted by the communication transmission module, used to enter the appearance feature correction information during the playback process, and send the appearance feature correction information to the communication transmission module after the playback ends; The communication transmission module receives the appearance feature correction information and transmits it to the recognition processing module. When the recognition processing module receives the appearance feature correction information, it processes the appearance feature correction into appearance feature correction data and trains the appearance recognition model.

2. The artificial intelligence-based building security monitoring system according to claim 1, characterized in that: When the identification processing module is used to query family data strips through appearance features, if no family data strips are found, the security monitoring terminal is used as the receiving and judging party of the image information, and the communication transmission module sends the corresponding same-stream image information to the security monitoring terminal.

3. The artificial intelligence-based building security monitoring system according to claim 2, characterized in that: The recognition processing module is used to query and match the face of a resident in the resident database through appearance features. If the face of a resident that meets the appearance features is not found, an interception instruction is generated based on the appearance features and image data, and the interception instruction is passed to the communication transmission module. After receiving the interception instruction, the communication transmission module sends the interception instruction and its corresponding appearance features and image data to the security monitoring terminal. If the security monitoring terminal receives the interception instruction and checks the appearance features and image data, and then enters and sends the appearance feature correction information to the communication transmission module, the communication transmission module passes the appearance feature correction information to the recognition processing module. The recognition processing module processes the appearance feature correction information and uses it to train the appearance recognition model.

4. The artificial intelligence-based building security monitoring system according to claim 3 is characterized in that: The communication transmission module is used to send the same-stream image data to the communication address in the resident information of the similar family data strip, and is also used to receive the security authentication sent back by the resident terminal corresponding to the communication address; after the resident terminal receives and views the same-stream image data, if the security authentication is not entered before closing the same-stream image data, the security authentication will be automatically generated according to the same-stream data after closing the same-stream image data and sent to the communication transmission module; if the security authentication is entered before closing the same-stream image data, the security authentication will be entered and sent to the communication transmission module after closing the same-stream image data.

5. The artificial intelligence-based building security monitoring system according to claim 4 is characterized in that: When the communication transmission module sends the same-stream image data to the resident terminal, if the received security authentication result contains a dangerous security authentication result, the same-stream image data will be sent to the resident terminal again, and a security verification prompt will be sent to the resident terminal at the same time. After receiving the security verification prompt, the resident terminal will play the security verification prompt and the same-stream image data; if the resident terminal checks the same-stream image data and the security verification prompt, and the security authentication sent to the communication transmission module contains a dangerous security authentication result, the communication transmission module will send the same-stream image data as security warning data to the security monitoring terminal.

6. The artificial intelligence-based building security monitoring system according to claim 5, characterized in that: The recognition processing module is used to search for matching resident faces in the resident database by appearance features, and is also used to recognize the environment in the image data, recognize environmental facilities according to the object contours in the image data, determine the moving speed of the resident according to the temporal position relationship of the environmental facilities combined with the shooting interval of the image data, and identify the resident whose moving speed exceeds the preset running speed as a running resident; The facial features of the running resident are detected by detecting key points of the face to identify the emotional features of the running resident, wherein the emotional features include happiness, sadness, anger, surprise, disgust, fear, and neutrality; When the emotional characteristics of the running resident are sadness, anger or fear, it is determined whether the running resident is a minor by judging the resident information of the running resident. If the running residents are all adults, the communication transmission module sends a danger warning to the security monitoring terminal; If there are minors among the running residents, it is determined whether the household addresses of the running residents are the same. If the household addresses are the same, the communication transmission module will send a safety reminder to the communication address of the household address; if the household addresses are different, it is determined whether there are adults among the running residents. If there are no adults, a reminder of children fighting will be sent to the communication address of the running residents' address; if the resident addresses are different and there are adults among the running residents, a danger warning will be sent to the security monitoring terminal.

7. The artificial intelligence-based building security monitoring system according to claim 6, characterized in that: When the recognition module identifies that the emotional characteristics of the running resident are sadness, anger or fear, it also identifies the running animal through the contour features, and determines whether a rope is connected between the running animal and the running resident. If no rope is connected, the communication transmission module sends a danger warning to the security monitoring terminal; if a rope is connected, it waits for a preset interval to determine whether the rope length is shortened. If not shortened, the communication transmission module sends a danger warning to the security monitoring terminal.

8. A building security monitoring method based on artificial intelligence, characterized in that: An artificial intelligence-based building security monitoring system using any one of claims 1-7.

Images