Intelligent building video monitoring system

Abstract

The utility model relates to the technical field of video monitoring, in particular to an intelligent building video monitoring system, which comprises a front-end acquisition layer, a network transmission layer, a data processing layer, a multi-dimensional application layer, a data storage layer, a control interaction layer and a display layer, the front-end acquisition layer is connected with the network transmission layer; the network transmission layer is respectively connected with the data processing layer, the multi-dimensional application layer, the data storage layer, the control interaction layer and the display layer; wherein the front-end acquisition layer comprises a binocular passenger flow statistics camera, a panoramic multi-view splicing camera, an AR spherical eagle eye camera, a trinocular behavior analysis camera, a face access control all-in-one machine, an illegal parking snapshot spherical camera and a face recognition snapshot camera. The problems that a traditional building monitoring system is insufficient in full-dimensional coverage, depends on a central server and is complex in network topology structure are solved.

Description

Technical Field

[0001] This utility model relates to the field of video surveillance technology, and in particular to a smart building video surveillance system. Background Technology

[0002] TOD commercial buildings, as integrated hubs combining transportation, commerce, and office space, are characterized by high pedestrian density, multi-business integration, and transportation hub integration, necessitating effective monitoring and management of this high-density population. Existing building monitoring systems face the following technical bottlenecks at the hardware architecture level:

[0003] (1) Insufficient full-dimensional coverage of equipment: The cameras deployed in traditional building monitoring systems are generally monocular bullet cameras or PTZ cameras, with a single field of view of only 60°-90°, which is insufficient in coverage. A large number of devices need to be stacked to cover the public areas of the building (such as the lobby, corridors, and parking lots), resulting in redundant number of lenses but still having blind spots in the field of view.

[0004] (2) Dependence on central server: In traditional monitoring systems, all data is transmitted to the central server, resulting in high bandwidth pressure and high latency. Moreover, if the central server fails, the entire system will be affected.

[0005] (3) Complex network topology: Due to the multi-layered structure of buildings, the wiring between devices is complex. Utility Model Content

[0006] To address the shortcomings of existing technologies, the purpose of this utility model is to provide a smart building video surveillance system that solves the problems of insufficient full-dimensional coverage of traditional building monitoring systems, reliance on central servers, and complex network topologies.

[0007] The solution adopted in this utility model is as follows:

[0008] A smart building video surveillance system includes a front-end acquisition layer, a network transmission layer, a data processing layer, a multi-dimensional application layer, a data storage layer, a control and interaction layer, and a display layer. The front-end acquisition layer is connected to the network transmission layer. The network transmission layer is connected to the data processing layer, the multi-dimensional application layer, the data storage layer, the control and interaction layer, and the display layer, respectively. The front-end acquisition layer includes a binocular passenger flow statistics camera, a panoramic multi-view stitching camera, an AR spherical eagle-eye camera, a tri-view behavior analysis camera, a face recognition access control integrated machine, a spherical camera for illegal parking capture, and a face recognition capture camera.

[0009] Furthermore, the binocular passenger flow statistics camera and the triocular behavior analysis camera form a detail capture module, which are connected to the network transmission layer via Category 6 network cables; the face recognition capture camera, the face access control integrated machine, and the illegal parking capture spherical camera form a dedicated business module, which are connected to the network transmission layer via Category 6 network cables; the panoramic multi-view stitching camera and the AR spherical eagle eye camera form a full-dimensional perception module, which are connected to the network transmission layer via Category 6 network cables.

[0010] Furthermore, the network transmission layer adopts a three-layer architecture of "access layer-aggregation layer-core layer"; wherein, the access layer is configured with device switches, the aggregation layer is configured with fiber optic distribution frames, and the core layer is configured with core switches; the device switches are connected to each device in the front-end acquisition layer through Category 6 network cables; the fiber optic distribution frames are connected to the device switches through optical fibers; and the fiber optic distribution frames are connected to the core switches through optical fibers.

[0011] Furthermore, the device switch is a PoE+ switch, which is deployed in the low-voltage electrical shaft on each floor of the building.

[0012] Furthermore, the network transport layer also includes edge computing nodes embedded in the device switch.

[0013] Furthermore, the data processing layer includes a situation analysis server and a face recognition server; the situation analysis server and the face recognition server are respectively connected to the core switch via Category 6 network cables.

[0014] Furthermore, the data storage layer includes a storage server; the storage server is connected to the core switch via a Category 6 network cable.

[0015] Furthermore, the control interaction layer includes a control keyboard; the control keyboard is connected to the core switch via a Category 6 network cable.

[0016] Furthermore, the multi-dimensional application layer includes a monitoring workstation and a video control platform; the monitoring workstation and the video control platform are respectively connected to the core switch via Category 6 network cables.

[0017] Furthermore, the display layer includes a splicing screen control matrix, a high-definition decoder, and a large display screen; one end of the splicing screen control matrix and the high-definition decoder are respectively connected to the core switch via Category 6 network cables; the other end of the large display screen and the high-definition decoder are respectively connected to the splicing screen control matrix via HDMI cables.

[0018] The beneficial effects of this utility model are as follows:

[0019] (1) This utility model constructs a three-dimensional monitoring network architecture of "panoramic stitching base + eagle eye close-up extension" through the collaborative work of cameras, which effectively improves the coverage area and solves the problem of blind spots that still exist in the stacking redundancy of traditional equipment.

[0020] (2) The network transmission layer of this utility model includes multiple device switches and embedded edge computing nodes (such as FPGA and NPU), which can perform preliminary processing on the structured data collected at the front end, reduce the amount of redundant data transmitted to other servers, significantly reduce bandwidth occupation and system latency, and at the same time enhance system fault tolerance, avoiding the problem of the whole system being paralyzed due to a single point of failure caused by only having a central server.

[0021] (3) The network transmission layer of this utility model adopts a redundant link design, and other devices are connected to the network transmission layer using Category 6 network cables, which simplifies the wiring complexity and ensures the reliability of the data transmission path; the equipment switches are deployed in layers in the weak current well, which facilitates regional maintenance and expansion.

[0022] Advantages of the present invention in additional aspects will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0023] The accompanying drawings, which form part of this specification, are used to provide a further understanding of this utility model. The illustrative embodiments of this utility model and their descriptions are used to explain this utility model and do not constitute an improper limitation of this utility model.

[0024] Figure 1 This is a schematic diagram of a smart building video surveillance system according to an embodiment of this utility model. Detailed Implementation

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0026] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this utility model.

[0027] Example 1

[0028] like Figure 1As shown, a smart building video surveillance system includes a front-end acquisition layer, a network transmission layer, a data processing layer, a multi-dimensional application layer, a data storage layer, a control and interaction layer, and a display layer; the front-end acquisition layer is connected to the network transmission layer; the network transmission layer is connected to the data processing layer, the multi-dimensional application layer, the data storage layer, the control and interaction layer, and the display layer, respectively.

[0029] Specifically, the front-end acquisition layer includes:

[0030] (1) Binocular passenger flow statistics camera: Supports bidirectional passenger flow counting and density detection, providing data support for the intelligent management of smart buildings; it collects new data such as "crowd gathering, onlookers, and lingering" to help businesses in the building further explore effective information. In this embodiment, it is deployed at building entrances and exits, elevator entrances, shopping mall main passages, etc., to collect video stream data and structured data, with the structured data being passenger flow statistics.

[0031] (2) Tri-lens behavior analysis camera: Monitors and analyzes the behavior of personnel intrusion into key areas, and provides real-time warnings for abnormal behaviors such as violent movement, climbing over fences, leaving posts, gathering in groups, fighting, and abnormal running, so as to deal with them in a timely manner. In this embodiment, it is deployed in key locations in the building, such as: isolation areas, escalators, etc., to collect video stream data and structured data. The structured data is abnormal behavior analysis data.

[0032] (3) Face recognition capture camera: This camera captures the faces of people passing by, extracts and analyzes facial features, and sends the face data to the backend personnel database of the data storage layer for face comparison. In this embodiment, it is deployed at building entrances and exits, personnel passages, etc., to collect image data and structured data, where the structured data is a facial feature vector.

[0033] (4) Face recognition access control all-in-one machine: Utilizing the non-contact nature and accuracy of face recognition, different access permissions are granted to users, which can be used for daily attendance tracking of internal personnel. In this embodiment, it is deployed at the main entrances and exits of the building, office floor entrances, equipment room / computer room entrances, etc., to collect image data and structured data, with the structured data representing access control permissions.

[0034] (5) Dome camera for illegal parking detection: This camera monitors vehicles within the designated area in real time. When an illegally parked vehicle is detected, it automatically takes a picture for identification and continuously monitors the vehicle's status within a certain detection cycle. If the vehicle remains illegally parked, it automatically generates a stream of evidence for illegal parking. In this embodiment, it is mainly deployed in the main passageway and fire lane / emergency exit of the underground parking lot to collect image data and structured data. The structured data constitutes the stream of evidence for illegal parking.

[0035] (6) Panoramic Multi-lens Stitching Camera: Equipped with multiple lenses (4-8), it can seamlessly stitch the captured images to form a complete panoramic image, achieving all-round coverage of the monitored area and reducing blind spots; it can also generate heat maps to mark congested areas. In this embodiment, it is deployed in the main atrium area of ​​the building, installed in a tetrahedral topology with a spacing of ≤15m, covering a space of 8-30m in height, for collecting panoramic images and structured data, with the structured data serving as alarm signals.

[0036] (7) AR Spherical Eagle Eye Camera: A 180° or 270° panoramic eagle eye camera can be selected depending on the environment. It also features multiple lenses (using dual fisheye lenses or multi-lens arrays), extending the monitoring range based on wide-angle imaging; simultaneously, it helps managers quickly locate alarm areas, view the situation on-site, and achieve blind-spot-free macro-real-time monitoring of key areas in large scenes. In this embodiment, it is also deployed in the main atrium area of ​​the building to collect panoramic images and structured data. The structured data is the alarm area location information, i.e., AR tags.

[0037] The binocular passenger flow statistics camera and the triocular behavior analysis camera form a detail capture module, which are connected to the network transmission layer via Category 6 network cables. The binocular passenger flow statistics camera and the triocular behavior analysis camera transmit the collected video data and structured data to the network transmission layer, respectively.

[0038] The face recognition capture camera, the face recognition access control machine, and the illegal parking capture spherical camera form a dedicated business module, which are connected to the network transmission layer via Category 6 network cables. The face recognition capture camera, the face recognition access control machine, and the illegal parking capture spherical camera transmit the collected image data and structured data to the network transmission layer respectively.

[0039] A panoramic multi-view stitching camera and an AR spherical eagle-eye camera form a full-dimensional perception module, which are connected to the network transmission layer via Category 6 network cables. The panoramic multi-view stitching camera and the AR spherical eagle-eye camera transmit the panoramic images and structured data they collect to the network transmission layer, respectively.

[0040] It is important to emphasize that the aforementioned binocular passenger flow statistics camera, tri-lens behavior analysis camera, face recognition capture camera, face recognition access control integrated machine, illegal parking capture spherical camera, panoramic multi-lens stitching camera, and AR spherical eagle-eye camera all use existing equipment and have built-in existing algorithms. This means that every function of the camera is a current function, and no improvements have been made to the algorithms for passenger flow statistics, behavior analysis, face recognition, illegal parking detection, image stitching, and location tracking. The improvement in this embodiment lies in the hardware architecture of the video surveillance system, such as the type of camera, connection method, and data transmission path.

[0041] In this embodiment, the binocular passenger flow statistics camera is Dahua DH-SDT4C1423-2F-DP-i, the panoramic multi-view stitching camera is Dahua DH-IPC-PFW5849-A180-E2, the AR spherical eagle-eye camera is Dahua DH-IPC-HDBW5842R-Z, the tri-view behavior analysis camera is Dahua DH-IPC-HFW5849H-S3, the face recognition access control all-in-one machine is Dahua DH-IPC-HDBW5842E-Z, the illegal parking capture spherical camera is Dahua DH-PSDW82449M-A270-D845L-HM-AR, and the face recognition capture camera is Dahua DH-IPC-HFW5849H-A.

[0042] By leveraging the collaborative functions of the detail capture module, the omni-dimensional perception module, and the business-specific module, scenario adaptability is significantly improved. Unstructured video, image, and structured data from each module are aggregated in real time, and the previously fragmented and mixed video and image data is transmitted to the network transport layer for processing. This solves the two core problems of traditional building monitoring systems: fragmented coverage and chaotic front-end data. In this embodiment, the effective coverage area of ​​multi-level building areas is significantly increased, resolving the problem of blind spots still existing despite a large number of stacked lenses in traditional equipment.

[0043] Specifically, the network transport layer:

[0044] The architecture adopts a three-layer structure: "Access Layer - Aggregation Layer - Core Layer".

[0045] The access layer is configured with device switches. In this embodiment, a PoE+ switch is used, which is directly connected to each device in the front-end acquisition layer via Category 6 network cables to provide device power (IEEE 802.3bt) and data access. The aggregation layer is configured with fiber optic distribution frames, which aggregate access layer data and interconnect with the core layer through redundant links. The core layer is configured with core switches. In this embodiment, a high backplane bandwidth core switch is used, which connects the data storage layer, data processing layer, and multi-dimensional application layer via Category 6 network cables, supporting multi-link load balancing.

[0046] Specifically, the binocular passenger flow statistics camera, the tri-lens behavior analysis camera, the face recognition capture camera, the face recognition access control integrated machine, the illegal parking capture spherical camera, the panoramic multi-lens splicing camera, and the AR spherical eagle eye camera respectively transmit the collected unstructured data (video stream data, image data) and structured data to the equipment switch; the equipment switch is deployed in the weak current shaft on each floor of the building.

[0047] The network transport layer also includes edge computing nodes, which are embedded in the device switches.

[0048] In this embodiment, the integrated edge computing node includes hardware such as FPGA and NPU. Both receive data transmitted from the front-end acquisition layer through the backplane bus of the PoE+ switch and are physically connected to the data processing board of the PoE+ switch through the PCIe 3.0x8 interface. The data after preliminary processing (such as passenger flow statistics compression and behavior alarm tag extraction) is transmitted to the core switch through the fiber optic distribution frame.

[0049] The PoE+ switches at the network transport and access layers embed FPGA / NPU hardware nodes to perform local preprocessing of structured data, reducing redundant video stream uploads by more than 80%, and lowering the bandwidth consumption of the central server and system latency. A three-layer redundant link design enables regional maintenance, with fiber optic links supporting 10 Gigabit data transmission. Redundant links ensure uninterrupted service in the event of a single point of failure, thus improving system reliability.

[0050] Specifically, the data storage layer includes storage servers. These servers are connected to the core switch via Category 6 network cables and are used to uniformly manage the data transmitted by the core switch. In this embodiment, the storage servers adopt a centralized storage architecture, support H.265 encoding, and support a bitrate of 4M. This enables the classified management of unstructured and structured data (such as alarm signals and facial features), meets the ISO 27001 security standard, and reduces the risk of data loss.

[0051] Specifically, the data processing layer includes a situation analysis server and a face recognition server; both are connected to the core switch via Category 6 network cables. The situation analysis server receives structured data (heatmaps, abnormal behavior tags) transmitted from the core switch, retrieves data stored in the storage server, and performs intelligent analysis. The situation analysis server provides users with intuitive situation maps and enables device linkage. The face recognition server retrieves facial feature data from the storage server, identifies abnormal individuals by comparing them with the database, and synchronizes the results to the situation analysis server.

[0052] The intelligent analysis functions of each server in the data processing layer are implemented using existing technologies.

[0053] The data processing layer is decoupled from the storage server, and the situation analysis and face recognition servers process structured data in parallel, avoiding overload of a single server and solving the problem of excessive dependence on a central server.

[0054] The control interaction layer includes a control keyboard; the control keyboard is connected to the core switch via a Category 6 network cable, and the core switch centrally manages and controls the front-end acquisition devices in real time, supports horizontal / vertical control of cameras, lens zoom and alarm response command issuance, and realizes cross-module collaborative response.

[0055] The multi-dimensional application layer includes a monitoring workstation and a video control platform; both are connected to the core switch via Category 6 network cables. The monitoring workstation serves as the interface between the user and the system, displaying real-time monitoring data and alarm information. The video control platform controls data output to the display layer, coordinating multiple video outputs to help administrators make quick decisions.

[0056] The display layer includes a video wall control matrix, a high-definition decoder, and a large display screen. One end of the video wall control matrix and the high-definition decoder are connected to the core switch via a Category 6 Ethernet cable. The other end of the large display screen and the high-definition decoder are connected to the video wall control matrix via HDMI cables. After the high-definition decoder parses the data, it transmits it to the large display screen via the HDMI cable to display panoramic images, heat maps, and alarm location information in real time. In this embodiment, the large display screen can be configured as a monitoring wall composed of nine 55-inch LCD video walls to display monitoring images in real time.

[0057] In this embodiment, modular collaboration and edge computing reduce the pressure on the central bandwidth, redundant links ensure the reliability of data transmission during peak periods, and all devices achieve full-scenario linkage management through a layered architecture.

[0058] In this embodiment, taking peak passenger flow management as an example, the collaborative workflow and data transmission path of the smart building video surveillance system are as follows:

[0059] S1: Front-end data acquisition and transmission:

[0060] The binocular passenger flow statistics camera monitors the flow of people in building entrances, elevators and other areas in real time. When the passenger flow exceeds the limit, the structured early warning signal (such as passenger flow density and aggregation data) is transmitted to the access layer PoE+ switch through Category 6 network cable.

[0061] A panoramic multi-view stitching camera provides panoramic coverage of the building's main atrium, generates heat maps to mark congested areas, and uploads the panoramic video stream and alarm signals to the access layer PoE+ switch via Category 6 network cables.

[0062] The three-lens behavior analysis camera detects abnormal behavior (such as falls or gatherings) in key locations such as escalators and isolation zones, and transmits structured alarm data (such as behavior tags) to the access layer PoE+ switch via Category 6 network cable.

[0063] The facial recognition camera captures facial features of people at entrances and exits, extracts facial feature vectors, and then uploads the data to the access layer PoE+ switch via Category 6 network cable. The data is then further transmitted to the facial recognition server for real-time comparison.

[0064] The facial recognition access control machine verifies entry and exit permissions at the main entrance of the building and uploads the access permission data to the access layer PoE+ switch via a Category 6 network cable, which then synchronizes it to the storage server to record attendance information.

[0065] The spherical camera for capturing illegal parking monitors illegally parked vehicles in the main passage of the underground parking lot and transmits the evidence stream of illegal parking (including license plate and continuous illegal parking status) to the access layer PoE+ switch via Category 6 network cable.

[0066] S2: Edge computing and network transmission:

[0067] The access layer PoE+ switches (deployed in the weak current wells of each layer) embed FPGA / NPU edge computing nodes to perform preliminary processing on structured data (such as passenger flow statistics, behavior alarms, and facial features) to reduce redundant video stream uploads. The processed data is then transmitted to the core layer core switch via optical fiber through the aggregation layer fiber distribution frame.

[0068] S3: Centralized Data Processing and Linked Response:

[0069] The core switch distributes unstructured and structured data to various servers: passenger flow warning signals, heat maps, and behavior alarm data are sent to the situation analysis server to generate real-time situation maps and link with other devices; facial feature data is sent to the facial recognition server for identity comparison, and abnormal personnel information is synchronized to the situation analysis server; illegal parking evidence streams and access control permission data are stored in the storage server.

[0070] The situation analysis server calls the PTZ control interface of the AR spherical eagle-eye camera through the video control platform, locates congested areas based on the heat map, automatically adjusts the camera's viewing angle for panoramic tracking, and overlays AR tags onto the image.

[0071] S4: Real-time monitoring and interactive control:

[0072] The monitoring workstation obtains processed data (such as situation maps and alarm information) from the core switch and outputs panoramic images, AR tags, and illegal parking capture images to the splicing screen of the display layer through the video control platform (which is parsed by a high-definition decoder and transmitted via HDMI cable).

[0073] The control keyboard sends commands through the core switch to manually adjust the camera's viewing angle or trigger dedicated business modules (such as close-up tracking of PTZ cameras for illegal parking detection).

[0074] Although the specific embodiments of the present utility model have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present utility model. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solution of the present utility model are still within the scope of protection of the present utility model.

Claims

1. A smart building video surveillance system, characterized in that, It includes a front-end acquisition layer, a network transmission layer, a data processing layer, a multi-dimensional application layer, a data storage layer, a control and interaction layer, and a display layer; the front-end acquisition layer is connected to the network transmission layer; the network transmission layer is connected to the data processing layer, the multi-dimensional application layer, the data storage layer, the control and interaction layer, and the display layer respectively; wherein, the front-end acquisition layer includes a binocular passenger flow statistics camera, a panoramic multi-view stitching camera, an AR spherical eagle-eye camera, a tri-view behavior analysis camera, a face recognition access control integrated machine, a spherical camera for illegal parking capture, and a face recognition capture camera.

2. The intelligent building video surveillance system as described in claim 1, characterized in that, The binocular passenger flow statistics camera and the triocular behavior analysis camera form a detail capture module, which are connected to the network transmission layer via Category 6 network cables; the face recognition capture camera, the face access control integrated machine, and the illegal parking capture spherical camera form a dedicated business module, which are connected to the network transmission layer via Category 6 network cables; the panoramic multi-view stitching camera and the AR spherical eagle eye camera form a full-dimensional perception module, which are connected to the network transmission layer via Category 6 network cables.

3. The intelligent building video surveillance system as described in claim 1, characterized in that, The network transmission layer adopts a three-layer architecture of "access layer - aggregation layer - core layer"; wherein, the access layer is configured with device switches, the aggregation layer is configured with fiber optic distribution frames, and the core layer is configured with core switches; the device switches are connected to each device in the front-end acquisition layer through Category 6 network cables; the fiber optic distribution frames are connected to the device switches through optical fibers; and the fiber optic distribution frames are connected to the core switches through optical fibers.

4. The intelligent building video surveillance system as described in claim 3, characterized in that, The equipment switches are PoE+ switches and are deployed in the low-voltage electrical shafts on each floor of the building.

5. The intelligent building video surveillance system as described in claim 3, characterized in that, The network transport layer also includes edge computing nodes, which are embedded in the device switch.

6. The intelligent building video surveillance system as described in claim 3, characterized in that, The data processing layer includes a situation analysis server and a face recognition server; the situation analysis server and the face recognition server are respectively connected to the core switch via Category 6 network cables.

7. The intelligent building video surveillance system as described in claim 3, characterized in that, The data storage layer includes a storage server; the storage server is connected to the core switch via a Category 6 network cable.

8. The intelligent building video surveillance system as described in claim 3, characterized in that, The control interaction layer includes a control keyboard; the control keyboard is connected to the core switch via a Category 6 network cable.

9. A smart building video surveillance system as described in claim 3, characterized in that, The multi-dimensional application layer includes a monitoring workstation and a video control platform; the monitoring workstation and the video control platform are respectively connected to the core switch via Category 6 network cables.

10. A smart building video surveillance system as described in claim 3, characterized in that, The display layer includes a splicing screen control matrix, a high-definition decoder, and a large display screen; one end of the splicing screen control matrix and the high-definition decoder are connected to the core switch via Category 6 Ethernet cables; the other end of the large display screen and the high-definition decoder are connected to the splicing screen control matrix via HDMI cables.

Images