Information fusion industrial control platform based on protection

Abstract

The invention discloses an information fusion industrial control platform based on protection. The information fusion industrial control platform comprises a localization calculation module, an environment protection module, a software programming module, a camera access module, a video processing module, a video recording management module, a decoding output module and a target identification module. According to the invention, a 24-core processor, namely a national cryptographic algorithm unit, is carried, hardware encryption and high-concurrency processing are realized, and data security is guaranteed; the EMC protection circuit stably operates in a severe environment; 64-channel video concurrence is combined with brightness equalization, geometric correction and motion compensation technologies, so that the picture splicing quality and the dynamic scene precision are improved; carrying out 30-path video stream target detection, fusing equipment state and network intrusion analysis, and realizing multi-dimensional abnormity alarm; various protocol conversion is supported, and heterogeneous equipment communication is simplified; various systems are compatible, and a native application and a deep learning framework are optimized; multi-modal feature extraction and a deep neural network are integrated, and behavior recognition and real-time decision making are supported.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information fusion, and in particular to a protection-based information fusion industrial control platform. BACKGROUND

[0002] With the increasing demand for intelligent border security, multi-source information fusion technology is increasingly important in the field of border monitoring. Traditional border defense systems mostly use general server architecture, which has problems such as dispersed computing resources, weak adaptability to extreme environments, and prominent data security risks.

[0003] In the border defense scenario, the device needs to adapt to complex geographical and climatic conditions such as high-altitude low temperature, desert high temperature, strong wind and sand, and cope with strong electromagnetic interference and field mechanical impact. The existing system core hardware and algorithms rely on imports, which is difficult to meet the nationalization safety standards, and there are performance bottlenecks in video intelligent analysis, multi-target tracking and emergency response, etc.

[0004] The current border monitoring system generally faces problems such as heterogeneous protocols of multi-source sensors such as thermal imaging, visible light camera, radar, unmanned aerial vehicle, multi-view picture distortion, and mobile target tracking frame loss, which reduces the efficiency of border situation awareness; target recognition and early warning mechanism are mostly based on single video data, lacking the ability of collaborative analysis with geographical information, biological recognition, communication signal and other multi-dimensional data; in addition, the border real-time command system is easily affected by network fluctuations caused by complex terrain in cross-regional data transmission, causing loss of key intelligence, and the compatibility of traditional communication protocols for heterogeneous equipment also restricts the system scalability.

[0005] In terms of information security, the border defense system urgently needs encrypted communication and self-controllable algorithms, but most platforms lack national secret-level hardware encryption modules, which poses a risk of data interception. At the same time, the adaptation and optimization of domestic processors and special operating systems are insufficient, resulting in low utilization of computing resources in high-concurrency scenarios such as night monitoring and multi-target identification, which is difficult to support the millisecond-level delay required for border emergency response.

[0006] Therefore, it is urgent to build a nationalized border defense platform integrating intelligent computing, border protection and multi-modal analysis, which breaks through the constraints of existing technology and meets the information fusion needs of high real-time performance, high reliability and strong anti-interference capability in border control through self-controllable hardware acceleration, multi-spectral video fusion optimization, military-grade environmental protection design and multi-dimensional security joint defense mechanism. Radar, optical, and biometric multi-source data collaboration support border line intelligent patrol, cross-border behavior analysis and emergency response, and ensure all-weather combat capability in complex environments. SUMMARY

[0007] The purpose of the present application is to solve the above technical problems by providing a protection-based information fusion industrial control platform.

[0008] A protection-based information fusion industrial control platform, comprising a localization computing module, an environmental protection module, a software programming module, a camera access module, a video processing module, a video recording management module, a decoding output module, and a target recognition module. The localization computing module is equipped with a 24-core Kunpeng processor and internally built SM2, SM3, and SM4 national cryptographic algorithm units. The environmental protection module realizes a wide-temperature-range protection component, a vibration suppression structure, and an EMC protection circuit, and is suitable for harsh environments. The software programming integrated module is internally built with PLC programming and integrates motion control, a communication protocol stack, a real-time database, and a deep neural network. The multi-camera access module is compatible with ONVIF and GB / T28181 access protocols. The video processing module has 64-way video concurrent processing capability and integrates video stitching, target feature extraction engine, and multi-level video cache architecture. The video recording management module provides video playback control and video download functions. The decoding output module is configured with 6-way HDMI high-definition decoding interfaces. The target recognition module detects 30-way video streams, realizes multi-dimensional alarm linkage, and generates abnormal alarm signals.

[0009] Further, a protection-based information fusion industrial control platform, wherein the localization computing module is adapted to SUSE, Ubuntu, CentOS, OpenEuler, UOS, and Kylin software operating systems.

[0010] Further, a protection-based information fusion industrial control platform, wherein the video stitching supports spatiotemporal domain fusion processing and comprises a brightness equalization module, a geometric correction module, and a motion compensation module. The brightness equalization module automatically adjusts the brightness and color consistency of each picture during multi-way video stitching and synthesis, eliminates uneven brightness caused by differences in camera models and lighting conditions, and ensures the overall visual effect of the synthesized picture. The geometric correction module corrects picture deformation caused by camera installation angle, lens distortion, or perspective differences, aligns the edges of multi-way video pictures through coordinate mapping and image transformation technology, and seamlessly stitches the pictures. The motion compensation module reduces the trailing or misplacement of synthesized video caused by object movement or camera shaking in dynamic scenes through inter-frame motion estimation and compensation algorithms.

[0011] Further, a protection-based information fusion industrial control platform, the target feature extraction engine includes human feature recognition, vehicle feature recognition and abnormal behavior detection multi-modal analysis.

[0012] Further, a protection-based information fusion industrial control platform, the multi-dimensional alarm linkage includes a three-level linkage model of device state monitoring, network intrusion detection and video anomaly analysis, supporting SNMP / Modbus TCP alarm protocol conversion.

[0013] Further, a protection-based information fusion industrial control platform, the real-time database adopts a time series data compression algorithm, supports microsecond-level timestamp synchronization, and has a breakpoint resume mechanism.

[0014] Further, a protection-based information fusion industrial control platform, the video processing module simultaneously supports 4-way analog standard definition video and 4-way high-definition coaxial video signal access coding.

[0015] Further, a protection-based information fusion industrial control platform, the communication protocol stack includes Modbus, Profinet, and OPC UA protocol conversion interfaces.

[0016] Further, a protection-based information fusion industrial control platform, the deep neural network adopts a domestic framework optimization model, supporting target feature extraction and behavior pattern analysis.

[0017] Further, a protection-based information fusion industrial control platform, the information fusion industrial control platform supports multi-device networking communication through 8 RS485 interfaces, expanding communication capabilities.

[0018] The beneficial effects of the present application are: through a protection-based information fusion industrial control platform, a 24-core Kunpeng processor and an SM2 / SM3 / SM4 national encryption algorithm unit are carried, hardware encryption and high-concurrency processing are realized, data security is ensured, wide-temperature-range protection, vibration suppression and EMC circuit are used for stable operation in harsh environments, 64-way video concurrency is combined with brightness equalization, geometric correction and motion compensation technology to improve picture stitching quality and dynamic scene precision, 30-way video stream target detection is combined with state and network intrusion analysis to realize multi-dimensional abnormal alarm, Modbus, Profinet and OPC UA protocol conversion are supported to simplify heterogeneous device communication, systems such as Kirin, UOS and OpenEuler are compatible, ARM native applications and deep learning frameworks are optimized, multi-modal feature extraction and deep neural networks are integrated, behavior recognition and real-time decision-making are supported. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 It is a framework diagram of a protection-based information fusion industrial control platform. DETAILED DESCRIPTION

[0020] In order to have a clearer understanding of the technical features, objectives and effects of the present application, the specific embodiments of the present application will now be described with reference to the accompanying drawings.

[0021] As shown in the accompanying drawings Figure 1 Specific embodiment 1: border panoramic monitoring and multi-level alarm This embodiment is applied to the border control scene. Through the multi-camera access module compatible with ONVIF and GB-T28181 protocol, 30 cameras are accessed. The video processing module is based on 64-way concurrent capability to splice video streams in real time, and integrates the following hardware units: Brightness equalization module: eliminate the illumination difference of moonlight and searchlight in night monitoring; Geometric correction module: FPGA correction chip, suitable for wide-angle lens distortion correction in border defense scene; Motion compensation module: dynamic optimization for moving targets such as patrol vehicles and unmanned aerial vehicles; The target recognition module is based on a 30-way video analysis architecture, and the algorithm library is upgraded to support border crossing personnel identification and suspicious vehicle feature extraction, such as modified vehicle chassis features. Combined with the real-time database's time compression, microsecond-level synchronization, and breakpoint resume storage alarm data, and through a three-level linkage model expansion: border defense equipment state monitoring, communication link intrusion detection, and video abnormal behavior analysis; The triggering mechanism uses Modbus and Profinet protocol to output alarm signals; The decoding output module outputs spliced video through 6-way HDMI hardware interface, and the video management module continues the 64-way concurrent playback architecture, supporting encrypted return download; The localization computing module uses a 24-core Kunpeng 920 processor and SM2 / SM3 / SM4 national encryption algorithm to adapt to Kirin and UOS systems for encrypted transmission; The environmental protection module uses EMC protection circuit to meet the operation requirements of highland permafrost and desert high temperature environment.

[0022] Specific embodiment 2: border emergency command and data fusion This embodiment is applied to the border patrol system. The software module is based on PLC programming architecture, and the protocol stack is expanded to support Beidou positioning and military encryption communication protocol. The control module coordinates unmanned patrol vehicles, reconnaissance drones and other devices through deep neural networks. The real-time database hardware is configured with time compression algorithm chip, microsecond-level synchronous clock generator, and breakpoint resume network controller. The video processing module maintains 64-way concurrent processing capability. The target recognition module is upgraded based on original industrial detection algorithm, supporting illegal border crossing recognition, such as personnel climbing action pattern detection and suspicious object left alarm; Environmental protection module design 1+1 redundant power supply architecture; The video management module is used to store 30 channels of planned videos, and the decoding output module outputs high-definition situational images via 6 channels of HDMI, including infrared / visible light fusion display.

[0023] Specific Example 3: Border Security Multi-Protocol Integration Management and Control This embodiment is applied to border security. It expands video input sources through device cascading capabilities, supports 64*4Mbps concurrent forwarding and 30-channel target recognition. The video processing module stitches together multi-channel camera videos, optimizes brightness, geometry, and motion compensation, and outputs them as a single-channel composite video, which is displayed via 6 HDMI interfaces. The target recognition module is combined with a deep neural network and optimized through a domestic framework to analyze the behavior of people and vehicles and trigger alarms; The industrial communication protocol stack implements Modbus and OPC UA protocol conversion and connects to heterogeneous security equipment; The domestically produced computing module is based on the Kunpeng 920 processor, ensuring high-speed data transmission and encrypted storage; The environmental protection module adapts to harsh outdoor conditions through anti-vibration structure and wide temperature range protection, and the fully redundant power supply ensures 24-hour uninterrupted operation; The video management module supports 64-channel playback and remote Web management. It combines time-series database compression technology to reduce storage resource usage and meet the high-security, low-latency border security management and control requirements.

[0024] Specific embodiment 4 A hardware parameter integration of industrial control equipment based on information fusion protection Specification item Description Processor model Kunpeng920 3211R (24-core 2.5GHz) *2 Memory 4*32G 3200M DDR4 RDIMM Hard disk 2*20T SATA hard disk Network Support 2 on-board network cards Interface Support 8 RS485 interfaces, 1 HDMI input interface, 6 HDMI output interfaces, 4 analog standard definition video interfaces, and 4 high-definition coaxial video interfaces System function Support up to 30 high-definition camera video inputs, compatible with Onvif and GB / T 28181 access protocols, and support RTSP address as a video source for input. 4 analog standard definition video and 4 high-definition coaxial video inputs. It has device access expansion capability and can expand video input source types through secondary development. Maximum support 64*4Mbps concurrent forwarding capability. Maximum support 30 high-definition camera plan recording and alarm recording storage. Maximum support 64-channel video playback concurrency. Support video playback control and video download function. Support scheduling video resources through a video information system and parameter configuration. Support multi-channel video synthesis forwarding, which splices multiple videos and outputs as one video for forwarding. Support 6-channel HDMI high-definition video decoding output. Maximum support 30-channel video target identification. After identifying the target, it can alarm according to the configuration and report the alarm information to the video information system and other third-party systems through the network protocol. It has device cascading capability and can realize horizontal expansion of video access, video output, and other expansion capabilities. Support Modbus, MQTT and other protocol standards to access various sensing devices. Power supply Support 2 hot-swappable 900W AC power modules Operating system SUSE, Ubuntu, CentOS, OpenEuler, UOS, Kylin software, etc. The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A protection-based information fusion industrial control platform, characterized in that: It includes domestic computing module, environmental protection module, software programming module, camera access module, video processing module, video management module, decoding output module, and target recognition module; The domestically produced computing module is equipped with a 24-core Kunpeng processor and built-in SM2, SM3, and SM4 national secret algorithm units; The environmental protection module implements wide temperature range protection components, vibration suppression structure, and EMC protection circuit to adapt to harsh environments; The software programming integration module has built-in PLC programming, integrated motion control, communication protocol stack, real-time database, and deep neural network; The multi-camera access module is compatible with ONVIF and GB / T28181 access protocols; The video processing module has the ability to process 64 channels of video concurrently, integrating video splicing, target feature extraction engine and multi-level video cache architecture; The video management module provides video playback control and video download functions; The decoding output module is equipped with 6-channel HDMI high-definition decoding interfaces; The target recognition module detects 30 video streams, realizes multi-dimensional alarm linkage, and generates abnormal alarm signals.

2. The information fusion industrial control platform based on protection according to claim 1 is characterized in that: The domestically produced computing module is compatible with SUSE, Ubuntu, CentOS, OpenEuler, UOS, and Kylin software operating systems.

3. The information fusion industrial control platform based on protection according to claim 1 is characterized in that: The video splicing supports spatiotemporal fusion processing, including a brightness equalization module, a geometric correction module, and a motion compensation module; The brightness balancing module automatically adjusts the brightness and color consistency of each picture when splicing and synthesizing multiple videos, eliminating the uneven brightness and darkness caused by differences in camera models and lighting conditions, and ensuring a unified overall visual effect of the synthesized picture; The geometric correction module corrects the image deformation caused by camera installation angle, lens distortion or perspective difference, and aligns the edges of multiple video images through coordinate mapping and image transformation technology, so as to achieve seamless splicing; The motion compensation module reduces the image drift or dislocation of the synthesized video by using an inter-frame motion estimation and compensation algorithm for the image offset caused by object movement or camera shaking in dynamic scenes.

4. The information fusion industrial control platform based on protection according to claim 1 is characterized in that: The target feature extraction engine includes multimodal analysis of human feature recognition, vehicle feature recognition and abnormal behavior detection.

5. The protection-based information fusion industrial control platform according to claim 1 is characterized in that: The multi-dimensional alarm linkage includes a three-level linkage model of equipment status monitoring, network intrusion detection, and video anomaly analysis, and supports SNMP / Modbus TCP alarm protocol conversion.

6. The protection-based information fusion industrial control platform according to claim 1 is characterized in that: The real-time database adopts a time series data compression algorithm, supports microsecond-level timestamp synchronization, and has a breakpoint resume mechanism.

7. The information fusion industrial control platform based on protection according to claim 1 is characterized in that: The video processing module supports the access coding of 4-channel analog standard definition video and 4-channel high-definition coaxial video signals at the same time.

8. The protection-based information fusion industrial control platform according to claim 1 is characterized in that: The communication protocol stack includes Modbus, Profinet, and OPC UA protocol conversion interfaces.

9. The protection-based information fusion industrial control platform according to claim 1, characterized in that: The deep neural network adopts a domestically produced framework optimization model to support target feature extraction and behavior pattern analysis.

10. The protection-based information fusion industrial control platform according to claim 1, characterized in that: The information fusion industrial control platform supports multi-device networking communication through 8 RS485 interfaces, thereby expanding communication capabilities.

Images