An intelligent sensing system of a remote sensing satellite with autonomous cooperative computing capability

Abstract

The application relates to a remote sensing satellite intelligent sensing system with autonomous cooperative computing capability, which comprises an exchange module, a power supply module, an LRM backboard and multiple edge node modules; the edge node modules are used for running star service control and remote control and telemetry application, platform autonomous dynamic management application, remote sensing data preprocessing application, intelligent sensing application and intelligent sensing information integration application, and have storage space integration and centralized management functions of the edge node modules; the exchange module is used for data interaction of the edge node modules; the power supply module is used for power supply of the edge node modules and the exchange module; and the LRM backboard is used for connecting the edge node modules, the exchange module and the power supply module. The application solves the problem that the existing on-board processing single machines of remote sensing satellites lack autonomous cooperation capability in and between single machines, multiple on-board processing single machines of multiple satellites or a single satellite work independently, and on-board processing resources of remote sensing satellites cannot be effectively integrated.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing satellite on-orbit data processing technology, and relates to a remote sensing satellite intelligent sensing system with autonomous collaborative computing capabilities. Background Technology

[0002] With the rapid development of wide-field-of-view and high-resolution imaging technology in remote sensing satellites, the volume of remote sensing image data has expanded dramatically. Meanwhile, key application scenarios such as environmental monitoring, disaster prevention and control, and smart cities have high demands for the diversity and real-time performance of remote sensing image data interpretation tasks. Currently, remote sensing satellite data processing platforms based on fixed data processing tasks have low application flexibility and cannot meet diverse needs; traditional applications based on downlinked remote sensing images for ground processing have high data transmission latency and cannot meet real-time requirements. By deploying a collaborative edge micro-cloud intelligent computing platform on remote sensing satellites to complete diverse remote sensing image interpretation tasks, and utilizing the flexible communication capabilities of Ethernet to achieve collaborative task scheduling and user information push, the diverse and real-time requirements of applications can be effectively addressed.

[0003] The paper "Design and Key Technology Analysis of Edge Computing Architecture for Remote Sensing Satellites [J]. Journal of Information Engineering University" by Wu Xu et al. proposes a "cloud-edge-device" intelligent computing scheme and envisions the space-based intelligent computing architecture for future remote sensing satellites, but fails to provide a data transmission solution for large-throughput remote sensing raw image data. Patent document CN116155895A by Zhang Fei et al. proposes a cloud-edge collaborative computing system and management method for satellite clusters, solving the problem of satellite-ground collaboration between satellites and ground cloud computing platforms, but it does not provide adaptation and optimization for large-scale on-orbit processing of remote sensing images. Patent document CN110647289A by Li Lingjun et al. proposes a satellite remote sensing cloud computing platform and system, designing a system to meet the needs of on-orbit processing of remote sensing images, but it contains a main control server, which may cause the system to malfunction due to a main control server failure. Summary of the Invention

[0004] The technical problem solved by this invention is to overcome the shortcomings of the prior art and propose a remote sensing satellite intelligent sensing system with autonomous collaborative computing capabilities. This solves the problem that the existing on-board processing units of various remote sensing satellites lack autonomous collaborative capabilities within and between individual units, and that the computing nodes within the intelligent processing unit achieve collaboration through manual design. Furthermore, multiple on-board processing units of multiple satellites or multiple on-board processing units of a single satellite work independently, making it impossible to effectively integrate the on-board processing resources of various remote sensing satellites.

[0005] The solution to the technical problem of this invention is: a remote sensing satellite intelligent sensing system with autonomous collaborative computing capabilities, including a switching module, a power supply module, an LRM backplane, and multiple edge node modules;

[0006] The edge node module is used to run onboard satellite control and remote telemetry applications, platform autonomous dynamic management applications, remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications. It also has the function of integrating and centrally managing the storage space of each edge node module; among which:

[0007] The remote sensing data preprocessing application includes: noise reduction, registration, slicing, and distribution of single-scene remote sensing data;

[0008] The intelligent sensing application: receives single-scene remote sensing data slices, performs intelligent target detection, and sends the detection results to the intelligent sensing information integration application through distributed communication;

[0009] The intelligent sensing information integration and application: The detection results are integrated by combining the satellite time and attitude orbit information to generate intelligent sensing information of the latitude, longitude, flight speed and direction of the suspected moving target, which is then sent to other satellites through distributed communication in the remote sensing satellite intelligent sensing system.

[0010] The platform's autonomous dynamic management application collects the health status of each edge node module and performs automated management, expansion, and load balancing of applications within each edge node module.

[0011] The exchange module is used for data interaction between various edge node modules;

[0012] The power supply module is used to power the edge node modules and switching modules;

[0013] The LRM backplane is used to connect various edge node modules, switching modules, and power supply modules.

[0014] Furthermore, the exchange module includes an ultra-high data rate remote sensing data exchange module and other general-purpose data exchange modules;

[0015] The ultra-high data rate remote sensing data exchange module is set as a point-to-point high-speed data exchange module, used to receive remote sensing data generated by the remote sensing camera and distribute it to each edge node module.

[0016] The other general data exchange modules are configured as 10 Gigabit Ethernet exchange modules, used for interconnection and synchronization control between edge node modules.

[0017] Furthermore, the hardware platform of the edge node module includes a computing unit, a storage unit, a point-to-point high-speed interface circuit, a 10 Gigabit Ethernet interface circuit, and a power supply unit; among which:

[0018] The edge node module acquires remote sensing data from the ultra-high data rate remote sensing data exchange module through a point-to-point high-speed interface circuit. The communication rate of each point-to-point high-speed interface is 10Gb / s.

[0019] The 10 Gigabit Ethernet interface circuit enables the edge node module to interact with other general-purpose data exchange modules;

[0020] The computing unit includes a CPU, an AI unit, and an FPGA; wherein: the CPU is connected to a 10 Gigabit Ethernet interface circuit and a storage unit, and is used to deploy the software architecture of the remote sensing satellite intelligent sensing system; the AI ​​unit is connected to the CPU and the storage unit, and is used for co-processing acceleration of intelligent algorithms; the FPGA is connected to a high-speed point-to-point interface circuit and the CPU, and is used to transmit data sent by the ultra-high data rate remote sensing data exchange module to the CPU.

[0021] The power supply unit is used to power the hardware platform of the edge node module.

[0022] Furthermore, the software architecture of the remote sensing satellite intelligent perception system includes an operating system, a container execution layer, a container orchestration and management tool, a distributed communication middleware, a distributed storage management middleware, and a software programming model;

[0023] Operating system: Provides the basic software environment for container execution layer, container orchestration and management tools, distributed communication middleware, distributed storage management middleware, and software programming models;

[0024] Container Execution Layer: Containerizes applications for satellite control and remote telemetry, platform autonomous dynamic management, remote sensing data preprocessing, intelligent sensing, and intelligent sensing information integration, achieving hardware and software decoupling.

[0025] Container orchestration and management tool: Enables automated management, expansion, and load balancing of applications within each edge node module, integrates the computing resources of each edge node module to achieve collaborative processing; has a negotiation and master election mechanism to ensure the decentralized operation of the remote sensing satellite intelligent perception system;

[0026] Distributed communication middleware: Satellite control and remote telemetry applications, platform autonomous dynamic management applications, remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications complete microsecond-level latency interaction of general data through distributed communication middleware.

[0027] Distributed storage management middleware: It integrates and centrally manages the storage space of each edge node module, and builds an application repository through the distributed storage management middleware for storing applications for satellite control and remote telemetry, platform autonomous dynamic management, remote sensing data preprocessing, intelligent sensing, and intelligent sensing information integration, and performs multiple backup redundant storage.

[0028] Software programming model: Enables the scheduling of computing resources for edge node modules, providing computing power support for remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications.

[0029] Furthermore, the container execution layer uses the Docker engine.

[0030] Furthermore, the container orchestration and management tool adopts Kubernetes tools.

[0031] Furthermore, the distributed communication middleware employs a data distribution service;

[0032] The distributed storage management middleware uses the Ceph tool.

[0033] Furthermore, the software programming model is compatible with AI frameworks including PyTorch, TensorFlow, and PaddlePaddle.

[0034] Furthermore, the other general data exchange modules connect the external onboard computer and internal edge node modules of the remote sensing satellite intelligent sensing system to transmit satellite control data streams, remote control and telemetry data streams, satellite time, attitude and orbit information data streams, and ground injection program data streams; they also connect to the internal edge node modules of the remote sensing satellite intelligent sensing system to transmit container orchestration and management data streams, intelligent sensing information data streams, and automatic backup and recovery data streams.

[0035] Furthermore, the process for multiple remote sensing satellite intelligent sensing systems to achieve on-orbit collaborative computing is as follows:

[0036] Coarse detection is performed by infrared remote sensing satellites: During operation, the remote sensing cameras of infrared remote sensing satellites create images, generating single-scene infrared remote sensing data. Remote sensing data preprocessing applications perform noise reduction, registration, slicing, and distribution of the single-scene infrared remote sensing data. Intelligent sensing applications perform intelligent target detection on the slices of single-scene infrared remote sensing data and send the detection results to the intelligent sensing information integration application through distributed communication middleware. The intelligent sensing information integration application integrates the detection results with satellite time, attitude, and orbit information and generates intelligent sensing information on the latitude, longitude, flight speed, and direction of suspected moving targets. This information is then sent to the remote sensing satellite intelligent sensing system on the optical remote sensing satellite through distributed communication middleware.

[0037] Precise detection is performed by an optical remote sensing satellite: Based on the received intelligent sensing information, the intelligent sensing system of the optical remote sensing satellite rotates the angle of the remote sensing camera to form an image, generating single-scene optical remote sensing data. The remote sensing data preprocessing application performs noise reduction, registration, slicing, and distribution of the single-scene optical remote sensing data. The intelligent sensing application performs intelligent target detection on the slices of single-scene optical remote sensing data and sends the detection results to the intelligent sensing information integration application through a distributed communication middleware. The intelligent sensing information integration application integrates the detection results with the satellite's time, attitude, and orbit information and generates a clear image of the target, which is sent to the onboard computer of the optical remote sensing satellite through the distributed communication middleware and then transmitted back to the ground.

[0038] The advantages of this invention compared to the prior art are:

[0039] (1) The system of the present invention integrates the computing and storage resources inside the on-board processing unit and can autonomously and dynamically manage each computing node inside the on-board processing unit, thus solving the problem of each computing node working independently inside the on-board processing unit. In addition, the system of the present invention has a collaborative computing interface, which can provide a cloud computing platform for intelligent sensing applications, thus solving the problem of multiple on-board processing units working independently in single-satellite and multi-satellite scenarios.

[0040] (2) The system of the present invention realizes decentralized autonomous dynamic management, avoids system failure caused by failure of the central control unit, and provides highly integrated intelligent computing power and storage capabilities for remote sensing satellite intelligent sensing missions.

[0041] (3) This invention enables remote sensing satellite intelligent sensing systems to have autonomous collaborative computing capabilities through container orchestration management tools, distributed communication middleware and distributed storage management middleware of edge node modules. Multiple remote sensing satellite intelligent sensing systems can be interconnected autonomously and have balanced computing task loads, providing an autonomous collaborative computing platform for remote sensing satellite intelligent sensing tasks.

[0042] (4) The system of the present invention realizes high-throughput remote sensing raw data transmission based on high-speed point-to-point data transmission, ensuring real-time data processing of remote sensing images on orbit. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the hardware platform of the remote sensing satellite intelligent sensing system according to an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of the edge node module according to an embodiment of the present invention;

[0045] Figure 3 This is a schematic diagram of the data flow in a spaceborne application environment according to an embodiment of the present invention. Detailed Implementation

[0046] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0047] Example 1

[0048] like Figure 1 The diagram shown is a structural schematic of the hardware platform of the remote sensing satellite intelligent sensing system according to an embodiment of the present invention, including a switching module, a power supply module, a standard LRM backplane, and at least three edge node modules.

[0049] The edge node module is used to run onboard satellite control and remote telemetry applications, platform autonomous dynamic management applications, remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications. It also has the function of integrating and centrally managing the storage space of each edge node module; among which:

[0050] The remote sensing data preprocessing application includes: noise reduction, registration, slicing, and distribution of single-scene remote sensing data;

[0051] The intelligent sensing application: receives single-scene remote sensing data slices, performs intelligent target detection, and sends the detection results to the intelligent sensing information integration application through distributed communication;

[0052] The intelligent sensing information integration and application: The detection results are integrated by combining the satellite time and attitude orbit information to generate intelligent sensing information of the latitude, longitude, flight speed and direction of the suspected moving target, which is then sent to other satellites through distributed communication in the remote sensing satellite intelligent sensing system.

[0053] The platform autonomously and dynamically manages applications by collecting the health status of hardware and applications in each edge node module and feeding it back to the container orchestration and management tool integrated in the edge node module for automated management, expansion, and load balancing of applications within each edge node module.

[0054] In this embodiment, four edge node modules are provided.

[0055] The exchange module is used for data interaction between various edge node modules;

[0056] The power supply module is used to power the edge node modules and switching modules;

[0057] The standard LRM backplane is used to connect various edge node modules, switching modules, and power supply modules.

[0058] like Figure 1As shown, in this embodiment, the exchange module includes an ultra-high data rate remote sensing data exchange module and other general data exchange modules. The ultra-high data rate remote sensing data exchange module is used to receive remote sensing data generated by the remote sensing camera and distribute it to each edge node module. The other general data exchange modules are used for interconnection and synchronization control between each edge node module.

[0059] The ultra-high data rate remote sensing data exchange module is configured as a point-to-point high-speed data exchange module, capable of providing high-speed point-to-point data transmission; the other general-purpose data exchange modules are configured as 10 Gigabit Ethernet exchange modules, capable of providing general-purpose data transmission with microsecond-level latency.

[0060] The other general-purpose data exchange modules can be interconnected via optical fiber, network cable or wireless module, providing a hardware foundation for the collaborative operation of multiple intelligent sensing systems within and between satellites.

[0061] The hardware platform of the edge node module includes a computing unit, a storage unit, a point-to-point high-speed interface circuit, a 10 Gigabit Ethernet interface circuit, and a power supply unit; wherein:

[0062] The edge node module acquires remote sensing data from the ultra-high data rate remote sensing data exchange module through a point-to-point high-speed interface circuit. The communication rate of each point-to-point high-speed interface is 10Gb / s.

[0063] The 10 Gigabit Ethernet interface circuit enables the edge node module to interact with other general-purpose data exchange modules;

[0064] like Figure 2 As shown, the computing unit includes a CPU, an AI unit, and an FPGA; wherein: the CPU includes, but is not limited to, the Phytium FT-D2000V octa-core processor, connected to a 10 Gigabit Ethernet interface circuit and a storage unit, for deploying the software architecture of the remote sensing satellite intelligent perception system; the AI ​​unit includes, but is not limited to, intelligent processing units including the Tianshu Zhixin Zhikai 100 GPU and the Denglin Technology Goldwasser GPU, connected to the CPU and a storage unit, for co-processing acceleration of intelligent algorithms; the FPGA includes, but is not limited to, the Fudan Micro V7-690T, connected to a high-speed point-to-point interface circuit and the CPU, for transmitting data sent by the ultra-high data rate remote sensing data exchange module to the CPU.

[0065] The software architecture of the remote sensing satellite intelligent sensing system includes an operating system, a container execution layer, container orchestration and management tools, distributed communication middleware, distributed storage management middleware, and a software programming model.

[0066] (1) Operating system: including but not limited to embedded Linux operating systems including Ubuntu 22.04 and Kylin V10, providing a basic software environment for container execution layer, container orchestration and management tools, distributed communication middleware, distributed storage management middleware and software programming model;

[0067] (2) The container execution layer uses the Docker engine to run applications such as star control and remote telemetry, platform autonomous dynamic management, remote sensing data preprocessing, intelligent sensing, and intelligent sensing information integration in a containerized manner, thereby achieving hardware and software decoupling, which can effectively improve application flexibility and reduce maintenance and update costs.

[0068] (3) The container orchestration and management tool uses Kubernetes to automate the management, expansion, and load balancing of applications within each edge node module, integrating the computing resources of each edge node module to achieve collaborative processing. In addition, the Kubernetes tool has a negotiation-based master election mechanism to ensure the decentralized operation of the remote sensing satellite intelligent sensing system, reduce the risk of system failure due to master node offline, and improve system reliability and stability.

[0069] (4) The distributed communication middleware adopts a data distribution service. Satellite control and remote telemetry applications, platform autonomous dynamic management applications, remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications complete the microsecond-level latency interaction of general data through the distributed communication middleware. In addition, the distributed communication middleware has easy scalability and can realize low-latency interaction of intelligent sensing information between remote sensing satellite intelligent sensing systems of multiple satellites.

[0070] (5) The distributed storage management middleware uses Ceph tools to integrate and centrally manage the storage space of each edge node module. An application repository is built through the distributed storage management middleware to store applications for satellite control and remote telemetry, platform autonomous dynamic management, remote sensing data preprocessing, intelligent sensing, and intelligent sensing information integration, and multiple backups are used for redundant storage.

[0071] (6) The software programming model is compatible with mainstream AI frameworks including PyTorch, TensorFlow, and PaddlePaddle, enabling the scheduling of computing resources for edge node modules and providing computing power support for remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications. Furthermore, it is compatible with intelligent sensing applications using different intelligent algorithm architectures, ensuring that high-quality intelligent sensing applications can be deployed efficiently within the remote sensing satellite intelligent sensing system.

[0072] like Figure 3 The diagram shown is a data flow diagram in the spaceborne application environment of this embodiment, wherein:

[0073] The ultra-high data rate remote sensing data exchange module connects the external remote sensing camera and the internal edge node module of the remote sensing satellite intelligent sensing system for transmitting remote sensing data streams;

[0074] Other general data exchange modules connect the external onboard computer and internal edge node modules of the remote sensing satellite intelligent sensing system, and are used to transmit satellite control data streams, remote control and telemetry data streams, satellite timing, attitude and orbit information data streams, and ground injection program data streams.

[0075] Other general data exchange modules connect to the internal edge node modules of the remote sensing satellite intelligent sensing system, and are used to transmit container orchestration management data streams, intelligent sensing information data streams, and automatic backup and recovery data streams.

[0076] This embodiment takes a joint approach as an example to illustrate in detail the on-orbit collaborative computing function of the remote sensing satellite intelligent sensing system based on the above method, including:

[0077] Step 1: The task begins by powering on the intelligent sensing system of the infrared and optical remote sensing satellites via the power supply module. Each edge node module loads the operating system, container execution layer, container orchestration and management tools, distributed communication middleware, and intelligent target detection algorithm based on software programming model.

[0078] Step 2: Coarse detection by infrared remote sensing satellites with low data rates: During operation, the remote sensing cameras of infrared remote sensing satellites generate single-scene infrared remote sensing data. Remote sensing data preprocessing applications perform noise reduction, registration, slicing, and distribution of the single-scene infrared remote sensing data. In this embodiment, the single-scene data is sliced ​​into four slices. Four intelligent sensing applications perform intelligent target detection on the four slices of single-scene infrared remote sensing data, and send the detection results to the intelligent sensing information integration application through distributed communication middleware. The intelligent sensing information integration application integrates the detection results with the satellite's time, attitude, and orbit information, and generates key intelligent sensing information such as the latitude, longitude, flight speed, and direction of suspected aircraft targets. This information is then sent to the remote sensing satellite intelligent sensing system on the optical remote sensing satellite through distributed communication middleware.

[0079] Step 3: Precise detection by a high-data-rate optical remote sensing satellite: The intelligent sensing system of the optical remote sensing satellite rotates the angle of its remote sensing camera to image based on the received key intelligent sensing information, generating single-scene optical remote sensing data. The remote sensing data preprocessing application performs noise reduction, registration, slicing, and distribution of the single-scene optical remote sensing data. In this embodiment, the single-scene data is sliced ​​into four slices. The four intelligent sensing applications perform intelligent target detection on the four slices of single-scene optical remote sensing data, and send the detection results to the intelligent sensing information integration application through a distributed communication middleware. The intelligent sensing information integration application integrates the detection results with the satellite's time, attitude, and orbit information, and generates a clear image of the target. This image is then sent to the onboard computer of the optical remote sensing satellite through the distributed communication middleware and transmitted back to the ground.

[0080] Step 4: Task completed. Each edge node starts the shutdown application and the power supply module is powered off.

[0081] In summary, this invention integrates the computing and storage resources within a single on-board processing unit, enabling autonomous and dynamic management of each computing node within the unit and solving the problem of independent operation of each node. Furthermore, multiple embodiments of this invention possess collaborative computing interfaces, providing a cloud computing platform for intelligent sensing applications and addressing the issue of multiple on-board processing units operating independently in single-satellite and multi-satellite scenarios.

[0082] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solutions of the present invention by utilizing the methods and techniques disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall fall within the protection scope of the technical solutions of the present invention.

[0083] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capabilities, characterized in that, Includes switching modules, power supply modules, LRM backplane, and multiple edge node modules; The edge node module is used to run onboard satellite control and remote telemetry applications, platform autonomous dynamic management applications, remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications. It also has the function of integrating and centrally managing the storage space of each edge node module; among which: The remote sensing data preprocessing application includes: noise reduction, registration, slicing, and distribution of single-scene remote sensing data; The intelligent sensing application: receives single-scene remote sensing data slices, performs intelligent target detection, and sends the detection results to the intelligent sensing information integration application through distributed communication; The intelligent sensing information integration and application: The detection results are integrated by combining the satellite time and attitude orbit information to generate intelligent sensing information of the latitude, longitude, flight speed and direction of the suspected moving target, which is then sent to other satellites through distributed communication in the remote sensing satellite intelligent sensing system. The platform's autonomous dynamic management application collects the health status of each edge node module and performs automated management, expansion, and load balancing of applications within each edge node module. The exchange module is used for data interaction between various edge node modules; The power supply module is used to power the edge node modules and switching modules; The LRM backplane is used to connect various edge node modules, switching modules, and power supply modules.

2. The remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 1, characterized in that, The exchange module includes an ultra-high data rate remote sensing data exchange module and other general data exchange modules; The ultra-high data rate remote sensing data exchange module is set as a point-to-point high-speed data exchange module, used to receive remote sensing data generated by the remote sensing camera and distribute it to each edge node module. The other general data exchange modules are configured as 10 Gigabit Ethernet exchange modules, used for interconnection and synchronization control between edge node modules.

3. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capabilities according to claim 2, characterized in that, The hardware platform of the edge node module includes a computing unit, a storage unit, a point-to-point high-speed interface circuit, a 10 Gigabit Ethernet interface circuit, and a power supply unit; among which: The edge node module acquires remote sensing data from the ultra-high data rate remote sensing data exchange module through a point-to-point high-speed interface circuit. The communication rate of each point-to-point high-speed interface is 10Gb / s. The 10 Gigabit Ethernet interface circuit enables the edge node module to interact with other general-purpose data exchange modules; The computing unit includes a CPU, an AI unit, and an FPGA; wherein: the CPU is connected to a 10 Gigabit Ethernet interface circuit and a storage unit, and is used to deploy the software architecture of the remote sensing satellite intelligent sensing system; the AI ​​unit is connected to the CPU and the storage unit, and is used for co-processing acceleration of intelligent algorithms; the FPGA is connected to a high-speed point-to-point interface circuit and the CPU, and is used to transmit data sent by the ultra-high data rate remote sensing data exchange module to the CPU. The power supply unit is used to power the hardware platform of the edge node module.

4. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capabilities according to claim 3, characterized in that, The software architecture of the remote sensing satellite intelligent sensing system includes an operating system, a container execution layer, container orchestration and management tools, distributed communication middleware, distributed storage management middleware, and a software programming model. Operating system: Provides the basic software environment for container execution layer, container orchestration and management tools, distributed communication middleware, distributed storage management middleware, and software programming models; Container Execution Layer: Containerizes applications for satellite control and remote telemetry, platform autonomous dynamic management, remote sensing data preprocessing, intelligent sensing, and intelligent sensing information integration, achieving hardware and software decoupling. Container orchestration and management tool: Enables automated management, expansion, and load balancing of applications within each edge node module, integrates the computing resources of each edge node module to achieve collaborative processing; has a negotiation and master election mechanism to ensure the decentralized operation of the remote sensing satellite intelligent perception system; Distributed communication middleware: Satellite control and remote telemetry applications, platform autonomous dynamic management applications, remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications complete microsecond-level latency interaction of general data through distributed communication middleware. Distributed storage management middleware: It integrates and centrally manages the storage space of each edge node module, and builds an application repository through the distributed storage management middleware for storing applications for satellite control and remote telemetry, platform autonomous dynamic management, remote sensing data preprocessing, intelligent sensing, and intelligent sensing information integration, and performs multiple backup redundant storage. Software programming model: Enables the scheduling of computing resources for edge node modules, providing computing power support for remote sensing data preprocessing applications, intelligent sensing applications, and intelligent sensing information integration applications.

5. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 4, characterized in that, The container execution layer uses the Docker engine.

6. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 4, characterized in that, The container orchestration and management tool uses Kubernetes.

7. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 4, characterized in that, The distributed communication middleware employs a data distribution service. The distributed storage management middleware uses the Ceph tool.

8. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 4, characterized in that, The software programming model is compatible with AI frameworks including PyTorch, TensorFlow, and PaddlePaddle.

9. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 4, characterized in that, The other general data exchange modules connect the external onboard computer and internal edge node modules of the remote sensing satellite intelligent sensing system, transmitting satellite control data streams, remote control and telemetry data streams, satellite timing, attitude and orbit information data streams, and ground injection program data streams; they also connect to the internal edge node modules of the remote sensing satellite intelligent sensing system, transmitting container orchestration and management data streams, intelligent sensing information data streams, and automatic backup and recovery data streams.

10. A remote sensing satellite intelligent sensing system with autonomous collaborative computing capability according to claim 9, characterized in that, The process for multiple remote sensing satellite intelligent sensing systems to achieve on-orbit collaborative computing is as follows: Coarse detection is performed by infrared remote sensing satellites: During operation, the remote sensing cameras of infrared remote sensing satellites create images, generating single-scene infrared remote sensing data. Remote sensing data preprocessing applications perform noise reduction, registration, slicing, and distribution of the single-scene infrared remote sensing data. Intelligent sensing applications perform intelligent target detection on the slices of single-scene infrared remote sensing data and send the detection results to the intelligent sensing information integration application through distributed communication middleware. The intelligent sensing information integration application integrates the detection results with satellite time, attitude, and orbit information and generates intelligent sensing information on the latitude, longitude, flight speed, and direction of suspected moving targets. This information is then sent to the remote sensing satellite intelligent sensing system on the optical remote sensing satellite through distributed communication middleware. Precise detection is performed by an optical remote sensing satellite: Based on the received intelligent sensing information, the intelligent sensing system of the optical remote sensing satellite rotates the angle of the remote sensing camera to form an image, generating single-scene optical remote sensing data. The remote sensing data preprocessing application performs noise reduction, registration, slicing, and distribution of the single-scene optical remote sensing data. The intelligent sensing application performs intelligent target detection on the slices of single-scene optical remote sensing data and sends the detection results to the intelligent sensing information integration application through a distributed communication middleware. The intelligent sensing information integration application integrates the detection results with the satellite's time, attitude, and orbit information and generates a clear image of the target, which is sent to the onboard computer of the optical remote sensing satellite through the distributed communication middleware and then transmitted back to the ground.

Images