An image recognition alarm method and system based on artificial intelligence

Abstract

This invention relates to the field of artificial intelligence technology, and provides an image recognition alarm method and system based on artificial intelligence. By focusing and acquiring target monitoring images, the image clarity is improved, thereby enhancing the identification of human characteristics in the monitored area, which is beneficial for personnel identification. By recognizing target monitoring images, the existence of preset scenarios can be determined in real time, thereby improving the identification of dangerous scenarios and protecting the lives and property of the public. Furthermore, for the real-time focused images corresponding to preset scenarios, human identification and scene modeling are performed. After determining the human characteristics of the target person, the movement path of the target person is determined based on the predicted movement route generated by scene modeling.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an image recognition alarm method and system based on artificial intelligence. Background Technology

[0002] Image recognition technology is an important field of artificial intelligence. It refers to the technology of recognizing objects in images to identify targets and objects of various patterns. In human image recognition systems, the recognition of complex images often requires information processing at different levels.

[0003] With the rapid development of artificial intelligence, AI devices have been introduced into image recognition for collaborative use. For example, security monitoring devices used in residential communities or shopping malls are typical products that combine artificial intelligence and image recognition technologies.

[0004] However, this only serves a monitoring function. If the target person engages in dangerous behavior in the monitored area and commits an act of violence, the victim can only report the incident by calling for help from bystanders or stopping the act of violence.

[0005] Meanwhile, during this period, the target is likely to escape, making it impossible to guarantee the safety of the crowd. Summary of the Invention

[0006] This invention provides an image recognition alarm method and system based on artificial intelligence, which solves the technical problem that existing image recognition technologies cannot intelligently identify dangerous behaviors and actions and accurately capture target individuals.

[0007] To address the above technical problems, this invention provides an image recognition alarm method based on artificial intelligence, comprising the following steps:

[0008] S1. Focus the image of the monitored area to obtain the target monitoring image;

[0009] S2. Identify each target monitoring image and determine whether a preset scene exists. If so, proceed to the next step.

[0010] S3. Obtain the real-time focused image of the corresponding preset scene, perform person recognition and scene modeling, and perform route planning to obtain the predicted movement route of the target person.

[0011] S4. Execute an alarm and interception against the target person based on the predicted movement route.

[0012] This basic solution improves image clarity by focusing on the target monitoring image, thereby enhancing the identification of human characteristics in the monitored area and facilitating personnel recognition. By identifying the target monitoring image, it determines in real time whether a preset scene exists, improving the identification of dangerous scenarios and protecting the lives and property of the public. Furthermore, it performs human recognition and scene modeling on the real-time focused image of the corresponding preset scene. After determining the characteristics of the target person, it determines the movement path of the target person based on the predicted movement route generated by the scene modeling.

[0013] In a further basic scheme, step S1 includes the following steps:

[0014] S11. Use infrared sensing technology to determine the personnel entering the current area;

[0015] S12. Focus on tracking the area where the person entering is located, and acquire the area image at each moment in real time as the target monitoring image.

[0016] This solution uses infrared sensing to initially lock onto personnel entering the premises, providing a focusing direction for monitoring equipment. This allows for obtaining high-resolution target monitoring images of the personnel entering the premises and enables real-time tracking.

[0017] In a further basic scheme, step S2 specifically includes:

[0018] Multiple preset scenarios are pre-stored to form a comparison database. Each target monitoring image is compared with each preset scenario. If the comparison is successful, proceed to the next step; otherwise, perform regular monitoring video storage.

[0019] Alternatively, image recognition can be used to extract the limb movements of people entering the target monitoring image and determine whether they are dangerous behaviors. If so, it is confirmed that there is a preset scene in the target monitoring image and proceeds to the next step; otherwise, regular monitoring video storage is performed.

[0020] The preset scenarios include specific scenarios in which the target person performs dangerous actions.

[0021] This solution configures a comparison database or a body movement recognition scheme for different usage scenarios, and performs high-precision recognition of different preset scenarios, thereby improving the accuracy of identifying dangerous scenarios.

[0022] In a further basic scheme, step S3 includes the following steps:

[0023] S31. Call the focus lock module to obtain the real-time focused image of the corresponding preset scene;

[0024] S32. Perform image recognition on the real-time focused image to determine the subject of the dangerous behavior as the target person, and use face recognition to obtain the facial feature data of the target person.

[0025] S33. Perform background removal operation on the real-time focused image to obtain the body feature data of the target person, and merge it with the face feature data to obtain portrait feature data;

[0026] S34. Based on the real-time focused image, perform scene simulation modeling, identify obstacles in the regional scene, and then combine the body feature data to plan at least one predicted movement route away from the corresponding area of ​​the preset scene.

[0027] When a preset scene is identified, this solution calls the focus locking module to obtain the corresponding real-time focused image. Using this image as a reference, the background removal operation is improved, and the human figure in the monitored area is extracted in a targeted manner. Then, facial recognition and posture recognition are used to clarify the human image feature data of the target person to assist in the subsequent target person location. At the same time, scene modeling is performed on the area where the preset scene occurs. Through obstacle recognition, the corresponding predicted movement route is planned according to the body characteristics of the target person to determine the movement path of the target person.

[0028] In a further basic solution, step S4 specifically involves: sending the facial feature data and predicted movement route of the target person to the rescue unit to execute on-site alarm, remote alarm and request emergency rescue services; performing big data retrieval based on the facial feature data to obtain the target person's identity information and address information for continuous tracking;

[0029] The emergency rescue services include remotely calling the police station for dispatch and on-site drone tracking.

[0030] When a preset scenario is detected in the monitored area, this solution will send an alarm to the rescue unit remotely based on facial feature data to request emergency rescue services. Simultaneously, it will send an alarm to the security room within the monitored area based on facial feature data and predicted movement routes, and the area security personnel will intercept the target. Additionally, drones will be deployed for tracking, allowing for continuous positioning of the target during their escape. Furthermore, if the target escapes after committing dangerous acts, big data analysis of the facial feature data can be used to pinpoint their identity and residential address.

[0031] The present invention also provides an artificial intelligence-based image recognition alarm system for implementing the above-mentioned artificial intelligence-based image recognition alarm method, including a main unit housing and a main control module installed inside the main unit housing, as well as an image acquisition component, an image processing component, and an alarm module installed outside the main unit housing; the image acquisition module includes a monitoring module, a real-time focusing module, a focus locking module, and an infrared electronic scanner respectively connected to the main control module;

[0032] The infrared electronic scanner is used to identify personnel entering the current area;

[0033] The real-time focusing module is used to focus on the area where the person entering is located;

[0034] The monitoring module is used, with the assistance of the real-time focusing module, to acquire in real time an area image of the area where the person entering is located as a target monitoring image;

[0035] The main control module is used to identify each target monitoring image and determine whether a preset scene exists.

[0036] The focus locking module is used to focus on the area where the preset scene is located, assisting the monitoring module in obtaining real-time focused images;

[0037] The image processing component is used to perform person recognition and scene modeling on the real-time focused image, and to perform route planning to obtain the predicted movement route of the target person.

[0038] The alarm module is used to issue an alarm to the target person based on the predicted movement route.

[0039] In a further basic scheme, the image processing component includes a graphics digitization module and a graphics processing module, which are respectively connected to the main control module;

[0040] The graphic digitization module is used to process the target monitoring image and upload it to the main control module after performing feature data digitization on the target monitoring image.

[0041] The graphics processing module is used to acquire the real-time focused image from the main control module, perform person recognition and scene modeling, and perform route planning to obtain the predicted movement route of the target person.

[0042] In a further basic scheme, the main control module includes an electrically connected PCB control terminal and a big data information processing terminal; the PCB control terminal is used to perform information processing and program execution, and control the operation of the big data information processing terminal, image acquisition component, image processing component and alarm module; the big data information processing terminal is used to connect to the Internet big data platform and provide big data support to the PCB control terminal.

[0043] In a further basic solution, the alarm module includes an audible and visual alarm, a remote communication module, and a tracking drone. The audible and visual alarm is used to perform on-site alarms, the remote communication module is used to perform remote alarms, and the tracking drone is used to perform on-site tracking based on the facial data of the target person.

[0044] In a further basic scheme, the main housing includes an internal cavity for installing the main control module, a number of first mounting through holes penetrating the housing on the top surface, a second mounting through hole penetrating the housing on the bottom surface, and a charging module embedded in one side wall; cable bundles are sleeved on the first and second mounting through holes.

[0045] The wiring harness on the main control module is connected to the image acquisition component and the image processing component through the wiring tube on the first mounting through hole, and to the alarm module through the wiring tube on the second mounting through hole.

[0046] This alarm system uses various modules to implement each step of the alarm method, providing a hardware foundation for the alarm method and facilitating its implementation. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the workflow of an image recognition alarm method based on artificial intelligence provided in an embodiment of the present invention;

[0048] Figure 2 This is a three-dimensional structural diagram of an image recognition alarm system based on artificial intelligence provided in an embodiment of the present invention;

[0049] Figure 3 This is provided by the embodiments of the present invention. Figure 2 Partial structural diagram of the main unit casing;

[0050] Figure 4 This is provided by the embodiments of the present invention. Figure 2 A 3D structural diagram of the monitoring module;

[0051] Figure 5 This is provided by the embodiments of the present invention. Figure 2 A 3D structural diagram of the real-time focusing module;

[0052] Figure 6This is provided by the embodiments of the present invention. Figure 2 A 3D structural diagram of the focusing and locking module;

[0053] Figure 7 This is provided by the embodiments of the present invention. Figure 2 A 3D structural diagram of the graphics processing module. Detailed Implementation

[0054] The embodiments of the present invention are described in detail below with reference to the accompanying drawings. The embodiments are given for illustrative purposes only and should not be construed as limiting the present invention. The accompanying drawings are for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0055] Example 1

[0056] This invention provides an image recognition alarm method based on artificial intelligence, such as... Figure 1 As shown, in this embodiment, steps S1 to S4 are included:

[0057] S1. Focus the image of the monitored area to obtain the target monitoring image, including steps S11~S12:

[0058] S11. Use infrared sensing technology to determine the personnel entering the current area; specifically, use infrared sensing to detect personnel entering the monitored area and identify the number of people in the monitored area.

[0059] S12. Focus on tracking the area where the personnel enter, and acquire the area image at each moment in real time as the target monitoring image.

[0060] This embodiment uses infrared sensing to initially lock onto the personnel entering the area, providing a focusing direction for the monitoring equipment, thereby obtaining a high-resolution target monitoring image of the personnel entering the area and enabling real-time tracking.

[0061] S2. Identify each target monitoring image and determine if a preset scene exists. If so, proceed to the next step, specifically:

[0062] Multiple preset scenarios are pre-stored to form a comparison database. Each target monitoring image is compared with each preset scenario. If the comparison is successful, proceed to the next step; otherwise, perform regular monitoring video storage.

[0063] Alternatively, image recognition can be used to extract the limb movements of people entering the target monitoring image and determine whether they are dangerous behaviors. If so, it is confirmed that there is a preset scene in the target monitoring image and proceeds to the next step; otherwise, regular monitoring video storage is performed.

[0064] The preset scenarios include specific situations in which the target person performs dangerous actions.

[0065] This embodiment configures a comparison database or a body movement recognition scheme for different usage scenarios to perform high-precision recognition of different preset scenarios, thereby improving the accuracy of identifying dangerous scenarios.

[0066] S3. Obtain the real-time focused image of the corresponding preset scene, perform person recognition and scene modeling, and perform route planning to obtain the predicted movement route of the target person, including steps S31~S34:

[0067] S31. Call the focus lock module to obtain the real-time focused image of the corresponding preset scene;

[0068] S32. Perform image recognition on the real-time focused image to identify the subject of the dangerous behavior as the target person, and use face recognition to obtain the facial feature data of the target person.

[0069] S33. Perform background removal on the real-time focused image to obtain the body feature data of the target person, and merge it with the face feature data to obtain the portrait feature data.

[0070] S34. Based on the real-time focused image, perform scene simulation modeling, identify obstacles in the regional scene, and then combine body feature data to plan at least one predicted movement route away from the corresponding area of ​​the preset scene.

[0071] Specifically, the distribution of obstacles in the scene is identified, and movable gaps are determined; then, starting from the current position of the target person, at least one movable route is formed by connecting the movable gaps.

[0072] By using body posture data to determine the target person's center of gravity and head tilt direction, the escape direction can be determined.

[0073] Obtain the route that aligns with the escape direction from the available movement routes as the predicted movement route.

[0074] In this embodiment, when a preset scene is determined to exist, the focus locking module is invoked to obtain the corresponding real-time focused image. At this time, based on this image, the background removal operation is improved, and the human figure in the monitoring area is extracted in a targeted manner. Then, the facial feature data of the target person is clarified through face recognition and posture recognition to cooperate with the subsequent target person positioning. At the same time, scene modeling is performed on the area where the preset scene occurs. Through obstacle recognition, the corresponding predicted movement route is planned according to the body characteristics of the target person to determine the movement path of the target person.

[0075] S4. Based on the predicted movement route, execute alarms and interception of the target person. Specifically, send the target person's facial feature data and predicted movement route to the rescue unit to execute on-site alarms, remote alarms, and request emergency rescue services; perform big data retrieval based on the facial feature data to obtain the target person's identity information and address information for continuous tracking.

[0076] Emergency services include remote police dispatch and on-site drone tracking.

[0077] When a preset scenario is detected in the monitored area, this solution will send an alarm to the rescue unit remotely based on facial feature data to request emergency rescue services. Simultaneously, it will send an alarm to the security room within the monitored area based on facial feature data and predicted movement routes, and the area security personnel will intercept the target. Additionally, drones will be deployed for tracking, allowing for continuous positioning of the target during their escape. Furthermore, if the target escapes after committing dangerous acts, big data analysis of the facial feature data can be used to pinpoint their identity and residential address.

[0078] This invention improves the clarity of target monitoring images by focusing on them, thereby enhancing the identification of human characteristics in the monitored area and facilitating personnel recognition. By identifying target monitoring images, it determines in real time whether a preset scene exists, thereby improving the identification of dangerous scenarios and protecting the lives and property of the public. Furthermore, it performs human recognition and scene modeling on the real-time focused images of the corresponding preset scene. After determining the characteristics of the target person, it determines the movement path of the target person based on the predicted movement route generated by the scene modeling.

[0079] Example 2

[0080] The reference numerals in the accompanying drawings of this embodiment include: main unit housing 1, monitoring module 2, first monitoring unit 21, second monitoring unit 22, and third monitoring unit 23; real-time focusing module 3, focusing locking module 4, first focusing module 41, and second focusing module 42; infrared electronic scanner 5, image digitization module 6, image processing module 7, first image background removal module 71, and second image background removal module 72; PCB control terminal 8, big data information processing terminal 9, alarm module 10, and charging module 11.

[0081] This invention also provides an artificial intelligence-based image recognition alarm system to implement the artificial intelligence-based image recognition alarm method provided in Embodiment 1 above. See [link to embodiment]. Figures 2-7It includes a main housing 1 and a main control module installed inside the main housing 1, as well as an image acquisition component, an image processing component and an alarm module 10 installed outside the main housing 1; the image acquisition module includes a monitoring module 2, a real-time focusing module 3, a focus locking module 4 and an infrared electronic scanner 5, which are respectively connected to the main control module.

[0082] Infrared electronic scanner 5 is used to identify personnel entering the current area;

[0083] The real-time focusing module 3 is used to focus on the area where the person entering is located;

[0084] The monitoring module 2 is used, with the assistance of the real-time focusing module 3, to acquire in real-time area images of the area where the personnel are located as target monitoring images;

[0085] The monitoring module 2 is equipped with a first monitoring unit 21, a second monitoring unit 22, and a third monitoring unit 23 installed side by side. The first monitoring unit 21, the second monitoring unit 22, and the third monitoring unit 23 are all cameras.

[0086] The main control module is used to identify each target monitoring image and determine whether a preset scene exists;

[0087] The focus locking module 4 is used to focus on the area where the preset scene is located, and assists the monitoring module 2 in obtaining real-time focused images;

[0088] The focus locking module 4 includes a first focus module 41 and a second focus module 42.

[0089] The image processing component is used to perform person recognition and scene modeling on real-time focused images, and to perform route planning to obtain the predicted movement route of the target person.

[0090] The alarm module 10 is used to issue an alarm to the target person based on the predicted movement route.

[0091] In a further basic scheme, the image processing component includes a graphics digitization module 6 and a graphics processing module 7, which are respectively connected to the main control module;

[0092] The graphic digitization module 6 is used to digitize the target monitoring image and upload it to the main control module after processing its feature data.

[0093] The graphics processing module 7 is used to acquire real-time focused images from the main control module, perform person recognition and scene modeling, and perform route planning to obtain the predicted movement route of the target person.

[0094] The graphics processing module 7 is equipped with a first image background removal module 71 and a second image background removal module 72.

[0095] In a further basic scheme, the main control module includes an electrically connected PCB control terminal 8 and a big data information processing terminal 9; the PCB control terminal 8 is used to perform information processing and program execution, and to control the operation of the big data information processing terminal 9, the image acquisition component, the image processing component, and the alarm module 10; the big data information processing terminal 9 is used to connect to the Internet big data platform and provide big data support to the PCB control terminal 8.

[0096] PCB control terminal 8 is the core of the system's computation and control. It is the final execution unit for information processing and program execution. PCB control terminal 8 is used for computation and control, processing information, and controlling program execution.

[0097] In a further basic solution, the alarm module 10 includes an audible and visual alarm, a remote communication module, and a tracking drone. The audible and visual alarm is used to perform on-site alarms, the remote communication module is used to perform remote alarms, and the tracking drone is used to perform on-site tracking based on the facial data of the target person.

[0098] In a further basic scheme, the main unit housing 1 includes an internal cavity for installing the main control module, a number of first mounting through holes penetrating the housing on the top surface, a second mounting through hole penetrating the housing on the bottom surface, and a charging module 11 embedded in one side wall; a cable bundle is sleeved on the first mounting through hole and the second mounting through hole.

[0099] The wiring harness on the main control module is connected to the image acquisition component and the image processing component through the wiring tube on the first mounting through hole, and to the alarm module 10 through the wiring tube on the second mounting through hole.

[0100] This alarm system uses various modules to implement each step of the alarm method described above, providing a hardware foundation for the alarm method and facilitating its implementation.

[0101] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. An image recognition alarm method based on artificial intelligence, characterized in that, Including the following steps: S1. Focus the image of the monitored area to obtain the target monitoring image; S2. Identify each target monitoring image and determine whether a preset scene exists. If so, proceed to the next step. S3. Obtain the real-time focused image of the corresponding preset scene, perform person recognition and scene modeling, and perform route planning to obtain the predicted movement route of the target person. S4. Execute alarm and interception on the target person according to the predicted movement route; Step S3 includes the following steps: S31. Call the focus lock module to obtain the real-time focused image of the corresponding preset scene; S32. Perform image recognition on the real-time focused image to determine the subject of the dangerous behavior as the target person, and use face recognition to obtain the facial feature data of the target person. S33. Perform background removal operation on the real-time focused image to obtain the body feature data of the target person, and merge it with the face feature data to obtain portrait feature data; S34. Based on the real-time focused image, perform scene simulation modeling, identify obstacles in the regional scene, and then combine the body feature data to plan at least one predicted movement route away from the corresponding area of ​​the preset scene. Specifically, Identify the distribution of obstacles in the scene and determine the movable gaps among them; then, starting from the current position of the target person, connect the movable gaps to form at least one movable route. By using body posture data to determine the target person's center of gravity and head tilt direction, the escape direction can be determined. Obtain the route that aligns with the escape direction from the available movement routes as the predicted movement route.

2. The image recognition alarm method based on artificial intelligence as described in claim 1, characterized in that, Step S1 includes the following steps: S11. Use infrared sensing technology to determine the personnel entering the current area; S12. Focus on tracking the area where the person entering is located, and acquire the area image at each moment in real time as the target monitoring image.

3. The image recognition alarm method based on artificial intelligence as described in claim 2, characterized in that, Step S2 specifically involves: Multiple preset scenarios are pre-stored to form a comparison database. Each target monitoring image is compared with each preset scenario. If the comparison is successful, proceed to the next step; otherwise, perform regular monitoring video storage. Alternatively, image recognition can be used to extract the limb movements of people entering the target monitoring image and determine whether they are dangerous behaviors. If so, it is confirmed that there is a preset scene in the target monitoring image and proceeds to the next step; otherwise, regular monitoring video storage is performed. The preset scenarios include specific scenarios in which the target person performs dangerous actions.

4. The image recognition alarm method based on artificial intelligence as described in claim 3, characterized in that, Specifically, step S4 involves: sending the facial feature data and predicted movement route of the target person to the rescue unit to execute on-site alarm, remote alarm and request emergency rescue services; and performing big data retrieval based on the facial feature data to obtain the target person's identity information and address information for continuous tracking. The emergency rescue services include remotely calling the police station for dispatch and on-site drone tracking.

5. An artificial intelligence-based image recognition alarm system, used to implement the artificial intelligence-based image recognition alarm method as described in any one of claims 1 to 4, characterized in that: It includes a main unit housing and a main control module installed inside the main unit housing, as well as an image acquisition component, an image processing component, and an alarm module installed outside the main unit housing; the image acquisition module includes a monitoring module, a real-time focusing module, a focus locking module, and an infrared electronic scanner, which are respectively connected to the main control module. The infrared electronic scanner is used to identify personnel entering the current area; The real-time focusing module is used to focus on the area where the person entering is located; The monitoring module is used, with the assistance of the real-time focusing module, to acquire in real time an area image of the area where the person entering is located as a target monitoring image; The main control module is used to identify each target monitoring image and determine whether a preset scene exists. The focus locking module is used to focus on the area where the preset scene is located, assisting the monitoring module in obtaining real-time focused images; The image processing component is used to perform person recognition and scene modeling on the real-time focused image, and to perform route planning to obtain the predicted movement route of the target person. The alarm module is used to issue an alarm to the target person based on the predicted movement route.

6. The image recognition alarm system based on artificial intelligence as described in claim 5, characterized in that: The image processing component includes a graphics digitization module and a graphics computing module, which are respectively connected to the main control module. The graphic digitization module is used to process the target monitoring image and upload it to the main control module after performing feature data digitization on the target monitoring image. The graphics processing module is used to acquire the real-time focused image from the main control module, perform person recognition and scene modeling, and perform route planning to obtain the predicted movement route of the target person.

7. The image recognition alarm system based on artificial intelligence as described in claim 5, characterized in that: The main control module includes an electrically connected PCB control terminal and a big data information processing terminal; the PCB control terminal is used to perform information processing and program execution, and to control the operation of the big data information processing terminal, the image acquisition component, the image processing component, and the alarm module; the big data information processing terminal is used to connect to an Internet big data platform and provide big data support to the PCB control terminal.

8. The image recognition alarm system based on artificial intelligence as described in claim 5, characterized in that: The alarm module includes an audible and visual alarm, a remote communication module, and a tracking drone. The audible and visual alarm is used to trigger an on-site alarm, the remote communication module is used to trigger a remote alarm, and the tracking drone is used to perform on-site tracking based on the facial data of the target person.

9. The image recognition alarm system based on artificial intelligence as described in claim 8, characterized in that: The main housing includes an internal cavity for installing the main control module, a number of first mounting through holes penetrating the housing on the top surface, a second mounting through hole penetrating the housing on the bottom surface, and a charging module embedded in one side wall; cable bundles are fitted onto the first and second mounting through holes. The wiring harness on the main control module is connected to the image acquisition component and the image processing component through the wiring tube on the first mounting through hole, and to the alarm module through the wiring tube on the second mounting through hole.

Images