An integrated system for ground operation and application of high-resolution remote sensing satellites

Abstract

The present invention discloses a ground operation and application integrated system for high-resolution remote sensing satellites. Guided by the support for the operation and application of high-resolution satellites, an integrated space-ground system integrating the design, testing, operation control and application of high-resolution remote sensing satellites is first established to achieve the interconnection and interoperability of high-resolution remote sensing satellite design simulation, comprehensive test verification, on-orbit operation management, and comprehensive application of business data. Then, driven by requirements, an automated task planning and scheduling strategy is constructed to achieve task conflict resolution and optimal planning. Secondly, a training set of typical targets is established, and target intelligent interpretation is realized based on machine learning, effectively improving the timeliness and accuracy of target recognition and confirmation. Finally, data such as meteorology, hydrology, surveying and mapping, and basic maps are introduced to construct target information combined with the three-dimensional environment, and the overall situation awareness of the scene and the target is realized after visualization.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing satellite operation and application, and particularly to an integrated ground operation and application system for high-resolution remote sensing satellites. Background Art

[0002] Generally, the process of remote sensing satellites from design to application includes four stages: simulation design, comprehensive testing, operation control management, and on-orbit application. Currently, these four stages are independent of each other and cannot achieve interconnection and interoperability. On the one hand, the cost is high and the equipment reuse rate is low. On the other hand, there is a lack of knowledge transfer and iteration between different stages, which limits the level of satellite design, development, and application.

[0003] In addition, at present, the link from the proposal of observation requirements to the acquisition of data products for domestic remote sensing satellites is complex, with low efficiency, low intelligence level, and slow response speed, which seriously affects the application efficiency of remote sensing satellites. This is mainly due to the lack of a highly intelligent task conflict resolution mechanism. Currently, the priority of tasks is mainly evaluated manually to form observation and data transmission plans.

[0004] At the same time, during the application process of remote sensing satellites, the intelligence level of data interpretation and intelligence analysis is not high. Most of them are based on the interpretation experience and prior knowledge of intelligence experts, and the timeliness and accuracy of intelligence analysis are severely limited.

[0005] Currently, the form of intelligence information formed is relatively single and the visualization level is not high. Through intelligence analysis, basic information such as target types and target states is provided, but there is a lack of a comprehensive three-dimensional situation awareness of the target and its surrounding environment. In addition, the visualization level of the current intelligence analysis system is not high. Summary of the Invention

[0006] The present invention aims to overcome the above problems in the prior art and provides an integrated ground operation and application system for high-resolution remote sensing satellites. Through the interconnection and interoperability of design, testing, operation, and application, this system solves the problems of poor knowledge transfer and low equipment reuse rate between the current four systems. At the same time, by establishing a mathematical model of external requirements and resources, it automatically resolves resource contention conflicts between various requirements and solves the problems of low task planning efficiency and unreasonable resource allocation. Secondly, new technologies such as artificial intelligence and big data are introduced to solve the problems of low intelligence level, timeliness, and accuracy of current intelligence analysis. Finally, by introducing information such as surveying and mapping, basic base maps, and ground data, it solves the problems of single intelligence form and limited decision-making support.

[0007] The present invention provides an integrated ground operation and application system for high-resolution remote sensing satellites, including:

[0008] Data receiving subsystem: It is used to receive and record satellite payload data, manage receiving resources and monitor data quality at the same time, and transmit satellite payload data to the data processing subsystem;

[0009] Data processing subsystem: It is used to receive the satellite payload data transmitted by the data receiving subsystem, process the data, generate data products and send them to the application service subsystem and users, perform calibration tests and quality assessments of the satellite, receive user requirements and forward them to the integrated management subsystem, and send data products to the application service subsystem;

[0010] Application service subsystem: It is used to receive user requirements transmitted by the integrated management subsystem, receive data products transmitted by the data processing subsystem, classify and process them to generate application products for users, and perform intelligent interpretation analysis, three-dimensional environment analysis, situation visualization and analysis, demand analysis and planning, thematic product production and multi-source data fusion while carrying out standard application product data production for data products;

[0011] Integrated management subsystem: It is used to communicate with the data receiving subsystem, data processing subsystem and application service subsystem, upload and download data, receive user requirements sent by the data processing subsystem and forward them to the application service subsystem, and perform user management, requirement management, task consultation and decision-making, task planning, plan management, system monitoring, integrated display and control, and external coordination.

[0012] In a preferred embodiment, the ground operation and application integrated system for high-resolution remote sensing satellites according to the present invention constructs an automated task planning and scheduling strategy based on the ground operation and application integrated system for high-resolution remote sensing satellites, with task requirements as the traction, to achieve task conflict resolution and optimal planning. Secondly, a training set of typical targets is established, and target intelligent interpretation is realized based on machine learning, effectively improving the timeliness and accuracy of target recognition and confirmation. Finally, data such as meteorology, hydrology, surveying and mapping, and basic maps are introduced to construct target information combined with the three-dimensional environment, and the overall situation perception of the scene and the target is realized after visualization.

[0013] In a preferred embodiment, the data receiving subsystem of the ground operation and application integrated system for high-resolution remote sensing satellites according to the present invention includes:

[0014] Ground receiving station: It is used to receive and record satellite payload data transmitted by the satellite, and transmit satellite payload data to the data transmission module;

[0015] Data transmission module: It is used to receive satellite payload data transmitted by the ground receiving station and transmit it to the data management and monitoring subsystem and the data processing subsystem;

[0016] Data Management and Monitoring Subsystem: It is used to receive the satellite payload data transmitted by the data transmission module and perform receiving resource management and receiving quality monitoring.

[0017] In the ground operation and application integrated system for high-resolution remote sensing satellites described in the present invention, as a preferred mode, the data processing performed by the data processing subsystem on the satellite payload data includes decompression, data preprocessing, cataloging and archiving, quality inspection, radiometric correction, geometric correction, and production of standard images of level 1-4 products.

[0018] In the ground operation and application integrated system for high-resolution remote sensing satellites described in the present invention, as a preferred mode, the automated task planning and scheduling strategy includes the following steps:

[0019] S1. Intelligent Planning and Scheduling of Observation Tasks: The integrated management subsystem selects observation tasks, allocates observation resources, and formulates an observation plan based on the user requirements forwarded by the data processing subsystem, using the task requirement degree, resource competition degree, time window requisition degree, and redundancy amount, and then sends the observation plan to the satellite.

[0020] S2. Intelligent Planning and Scheduling of Data Transmission Tasks: The integrated management subsystem solves the data transmission scheduling sub-problem to determine the observation data transmission sequence, transmission object, and formulates a transmission plan, and then sends the transmission polarization to the data receiving subsystem.

[0021] In the ground operation and application integrated system for high-resolution remote sensing satellites described in the present invention, as a preferred mode, step S1 includes the following steps:

[0022] S11. Preparation: The integrated management subsystem receives the user requirements forwarded by the data processing subsystem.

[0023] S12. Task Selection: Calculate the task requirement degree Need i for observation task l i and sort by urgency.

[0024] S13. Resource Selection: Calculate the resource competition degree Contention ij and the time window contention degree Con IC for observation resource allocation.

[0025] S14. Task Composition: Calculate the redundancy amount for selecting the task composition plan.

[0026] S15. Formulate Observation Plan: Formulate an observation plan and send it to the satellite.

[0027] An integrated ground operation and application system for high-resolution remote sensing satellites according to the present invention, as a preferred mode, realizing target intelligent interpretation based on machine learning includes the following contents: establishing a training set of typical targets for data interpretation and intelligence interpretation, and performing knowledge mining and training learning based on the training set; after obtaining satellite payload data, combining the satellite payload data samples to train knowledge utilization target recognition algorithms and classification recognition algorithms based on cognitive computing to perform target detection, discovery and recognition confirmation.

[0028] An integrated ground operation and application system for high-resolution remote sensing satellites according to the present invention, as a preferred mode, constructing target information combined with a three-dimensional environment includes the following steps:

[0029] SⅠ. Positioning analysis: The integrated management subsystem analyzes and locates the satellite payload data interpreted and judged by the application service subsystem;

[0030] SⅡ. Extracting environmental element information: Extracting environmental element information of the scene where the target is located from the environmental element database, and the environmental element information includes meteorological information, hydrological information, surveying and mapping information, basic base map information and ground data;

[0031] SⅢ. Fusion: Fusing the satellite payload data with the environmental element information to construct a multi-element and multi-dimensional visualization platform integrating multi-scale spatial scenes and variable time series.

[0032] The main goal of intelligent observation task planning is to determine which tasks to observe, what resources to use for observation, and when to observe under the condition of the most observation tasks or the greatest observation benefits. The main goal of data transmission resource scheduling is to optimize and determine which observation data can be transmitted, which ground station to transmit to, and when to start transmission, etc. After the observation and data transmission task planning is completed, the final solution result needs to be evaluated. If the optimization critical value is met, the iteration can be terminated; otherwise, multiple iterations are performed until the conditions are met.

[0033] The intelligent planning and scheduling of observation tasks mainly include task selection, resource selection, time window selection and task synthesis. In the process of task selection, the task requirement degree Need i is defined to describe the urgency of task l i to be arranged.

[0034]

[0035] where w i is the priority of the task, and Opportunities i is the number of remaining observation opportunities for this task, that is, the number of current available time windows. The purpose of the task selection strategy is to give priority to arranging tasks with high value and few remaining observation opportunities.

[0036] During the resource selection process, the resource competition degree and the time window contention degree are used as quantitative indicators. Since resources such as the electric energy, storage capacity, and number of side-swing operations of remote sensing satellites are limited and only part of the tasks can be completed, there is competition for resources among tasks. To express the degree of competition of tasks for satellite resources, the concept of resource competition degree is defined to describe task t i for satellite s i the competition situation of resources:

[0037]

[0038] Suppose the satellite has k types of resources, RequestedCapacit(y,i)r represents the total demand of the candidate tasks of satellite s i for its r-type resources, Capacity(i,r) represents the remaining amount of the r-type resources of satellite s i and Requires(i,j,r) is the demand of task t j for the r-type resources of satellite s i . The resource competition degree reflects the degree of competition of tasks for satellite resources. The larger the Contention ij , the smaller the resource competition degree of task t j for satellite s i ; the smaller the Contention ij , the larger the resource competition degree of task t j for satellite s i . When selecting resources for tasks, satellites with small resource competition degrees should be selected as much as possible, which also meets the requirements of load balancing in multi-satellite scheduling.

[0039] Since a satellite can only adopt one attitude for imaging at the same time, different tasks not only compete for resources such as the electric energy and storage of the satellite, but also compete in the time window. In this project, the time window contention degree is defined to measure the degree of competition of tasks for time window K.

[0040]

[0041] Among them, m represents the number of unassigned tasks, ω j represents the priority (value) of unassigned task t j , ConflictWinSpan ICj represents the duration of the conflict. Using this strategy, tasks will be arranged into the time window with the least conflict with other tasks as much as possible.

[0042] When the satellite images ground targets, data needs to be stored in the on-board memory. After multiple tasks perform synthetic observations, there will be some redundant observation data. The principle of task synthesis is that the redundancy between tasks should be minimized.

[0043] Suppose there are k meta-tasks in the synthesis task, and the start time of the i-th meta-task is ws i , and the end time is we i , and the duration is duration. The redundancy of the synthesis scheme is defined as follows:

[0044]

[0045] The redundant part does not generate benefits and also occupies a certain amount of storage resources. The redundancy not only represents the waste of storage resources by the synthesis observation task, but also represents the waste of other resources such as the continuous power-on time and energy of the satellite. Therefore, a task synthesis scheme with a small redundancy should be selected as much as possible.

[0046] The intelligent planning module for data transmission tasks mainly solves the problem of efficient and reliable transmission of observation data through the data transmission window. If each data transmission window and its possible task observation windows are regarded as a type of planning problem, then the intelligent planning and scheduling problem of data transmission tasks can be composed of multiple sub-problems, that is, to solve the problem of planning with multiple observation windows and a single data transmission window. This problem uses the label updating algorithm to solve the data transmission scheduling sub-problem. The label updating algorithm is a type of solution algorithm that is widely used in practical applications and has the advantages of short solution time, high solution quality, simple and stable algorithm, etc. First, a label array is reserved for each vertex of the same type, and a label array is reserved from the origin to each vertex. The optimal path information from the origin to each vertex is reserved in the label array of that vertex. The vertices in the graph are sequentially selected as candidate vertices, and for each label information reserved on the candidate vertices, the label paths of the vertices connected to this vertex by edges are updated in turn and compared with the label arrays of the corresponding vertices. When a vertex becomes a candidate vertex, the information in its label array is no longer updated. The search stops until it ends. After the algorithm ends, the optimized observation target sequence of the end point can be obtained according to the label set information of the end point.

[0047] Use a bidirectional cognitive computing model. Among them, the forward transformation process is to convert the connotation expressed by the acquired image into a conceptual extension that is convenient for computer storage and calculation and feature recognition; while the reverse transformation process is to learn, abstract or extract the knowledge, information, etc. contained in the acquired image to form a conceptual connotation for people to understand and recognize. Therefore, the forward transformation and the reverse transformation are used in a bidirectional cycle multiple times, that is, a qualitative concept is used to generate quantitative data through the forward transformation algorithm, and then a qualitative concept is formed by the reverse transformation algorithm, and this cycle is carried out multiple times to simulate the human bidirectional cognitive computing process of concepts. The cognitive computing process of dynamic feeding is a process of obtaining corresponding knowledge or concepts from the existing samples or information as the examples or the amount of information of the given cognitive object increases, and studying the dynamics of human concept cognition by comparing the obtained concepts. Data interpretation and intelligence interpretation based on machine learning use forward cognitive computing transformation and different reverse cognitive computing transformations to simulate the dynamics of different people's cognition of things or concepts. The specific algorithm process is as follows: for a given initial concept, first obtain different numbers of data samples (conceptual extensions) through forward cognitive computing transformation, and then use different reverse cognitive computing (including sample random sampling, fixed sampling and repeated sampling) to simulate different cognitive computing capabilities, so as to obtain the digital feature estimation values of the corresponding new concepts.

[0048] The overlay display of basic data mainly solves the problems of single data structure of current intelligence information and limited decision support ability. This system has the ability to access environmental elements. First, it analyzes the interpretation and judgment results of remote sensing satellite data, and after positioning, extracts the environmental element information of the target scene from the environmental element database, mainly including meteorology, hydrology, surveying and mapping, basic base map, ground data and other information. After fusing the target information and the scene information, a multi-element and multi-dimensional visualization platform integrating multi-scale spatial scenes and variable time series is constructed to support application services such as situation awareness, data product customization, information value-added, and solution customization.

[0049] In a preferred embodiment, the ground operation and application integrated system for high-resolution remote sensing satellites according to the present invention has a multi-layer distributed structure, including a data layer, a service layer, and an application layer. The multi-layer distributed application system architecture has the advantages of good reusability, good scalability, manageability and easy maintainability, and can realize information intercommunication, data sharing and application integration of the system. In particular, on the basis of dividing the data layer, the service layer, and the application layer, extracting the process logic from the business logic helps business process automation and business process reengineering, so that the modification of any layer of logic will not affect other layers, thereby minimizing the internal coupling of the system, improving the intensification level of the entire system, and enhancing the system's ability to adapt to changes, providing a basic data structure for the automated operation and intelligent operation of the entire ground application system, and thus effectively saving costs.

[0050] The technical solution of the present invention is as follows: An integrated ground operation and application system for high-resolution remote sensing satellites realizes the interconnection and interoperability of design, testing, operation and application through process optimization and resource integration, realizes the horizontal cross-transfer of knowledge, and improves the reuse rate of equipment while effectively improving the satellite design and application level. Secondly, by integrating internal and external requirements and resources, establishing corresponding decision-making models, and realizing the optimal regulation and allocation of satellite resources, receiving resources, and transmission resources. Then, new technologies such as artificial intelligence and big data are introduced to effectively improve the timeliness and accuracy of intelligence interpretation. Finally, information such as meteorology, hydrology, surveying and mapping, basic maps, and ground data is introduced on the basis of intelligence interpretation and visually displayed to realize the all-element three-dimensional perception of targets and scenes.

[0051] The integrated system of design, testing, operation and application adopts a multi-layer distributed structure design of "data layer / business layer / application layer", which effectively improves the intensification level and adaptability of the system while realizing the interconnection of information and knowledge.

[0052] The task conflict resolution and intelligent planning method identifies task requirements and resource attributes, then establishes corresponding decision-making models according to relevant constraints and conditions, and obtains the results of task planning by solving the models.

[0053] The intelligent intelligence interpretation method first establishes a remote sensing image training set of typical targets, and uses artificial intelligence technologies such as hierarchical multi-task learning and deep learning to realize target detection, discovery, identification and confirmation.

[0054] The intelligent intelligence interpretation method first establishes a remote sensing image training set of typical targets, and uses brain-inspired artificial intelligence technologies such as hierarchical multi-task learning and deep learning to realize target detection, discovery, identification and confirmation.

[0055] The method for diversifying and visualizing intelligence information is as follows: On the basis of intelligence interpretation, information such as meteorology, hydrology, surveying and mapping, basic base maps, and ground data is introduced to effectively enrich the form of intelligence. At the same time, visual display is carried out to realize the all-element three-dimensional perception of targets and scenes, and effectively improve the decision-making support ability of intelligence data.

[0056] The present invention optimizes the structure and process of the existing system, introduces new technologies, and realizes the improvement of the capabilities and efficiency of the ground operation and application system of high-resolution remote sensing satellites; has the capabilities of designing and testing high-resolution remote sensing satellites, and realizes design optimization while improving the reuse rate of equipment through the interconnection and interoperability of design, testing, operation and application; through integrating external requirements and internal resources, realizes the efficient and optimal planning of tasks and resource allocation through the optimal solution of the decision-making model; with the support of technologies such as image simulation, big data, and artificial intelligence, realizes the detection, discovery, identification and confirmation of targets in image data, effectively improving the timeliness and accuracy of intelligence interpretation; on the basis of target recognition and confirmation, introduces information such as meteorology, hydrology, surveying and mapping, basic base maps, and ground data, and can realize the three-dimensional situation awareness of targets and scenes after visual display.

[0057] The beneficial effects of the present invention are as follows:

[0058] (1) The system design proposed by the present invention adopts the idea of multi-layer distributed structure design of "data layer / business layer / application layer". The multi-layer distributed application system architecture has the advantages of good reusability, good scalability, manageability and easy maintainability, and can realize information interconnection, data sharing and application integration of the system. In particular, on the basis of dividing the data layer, business layer and application layer, extracting the process logic from the business logic helps business process automation and business process reengineering, so that the modification of any layer of logic will not affect other layers, thereby minimizing the internal coupling of the system, providing the intensive level of the entire system, improving the system's ability to adapt to changes, providing the basic data structure for the automated operation and intelligent operation of the entire ground application system, thus effectively saving costs, and realizing the interconnection and interoperability of design, testing, operation and application, realizing the effective and smooth transfer of knowledge, improving the equipment reuse efficiency while enhancing the research and development and application levels;

[0059] (2) The task conflict resolution and intelligent planning algorithm disclosed by the present invention can effectively improve the scientificity and planning efficiency of task planning, and further improve the response speed while realizing the optimal allocation of resources;

[0060] (3) The intelligent intelligence interpretation method disclosed by the present invention can realize the intelligent interpretation of intelligence by using methods such as artificial intelligence on the basis of establishing a typical target training set, improving the timeliness and accuracy of intelligence interpretation.

[0061] (4) The intelligence information provided by the present invention can overlay information such as meteorology, hydrology, surveying and mapping, basic base maps, and ground data, and perform visual display, which can further enrich the form of intelligence data and improve the decision-making support ability of intelligence information. Description of the Drawings

[0062] Figure 1It is a structural diagram of an integrated ground operation and application system for high-resolution remote sensing satellites;

[0063] Figure 2 It is a flowchart of task conflict resolution and intelligent planning for an integrated ground operation and application system for high-resolution remote sensing satellites;

[0064] Figure 3 It is a schematic diagram of multi-task redundancy for an integrated ground operation and application system for high-resolution remote sensing satellites;

[0065] Figure 4 It is a schematic diagram of the problem of planning single data transmission windows for multiple observation windows in an integrated ground operation and application system for high-resolution remote sensing satellites;

[0066] Figure 5 It is a flowchart of data interpretation and intelligence based on machine learning for an integrated ground operation and application system for high-resolution remote sensing satellites;

[0067] Figure 6 It is a schematic diagram of a two-way cognitive computing model for an integrated ground operation and application system for high-resolution remote sensing satellites.

[0068] Figure 7 It is a flowchart of a two-way cognitive computing model for an integrated ground operation and application system for high-resolution remote sensing satellites.

[0069] Reference numerals:

[0070] 100, data reception subsystem; 110, ground receiving station; 120, data transmission module; 130, data management and monitoring subsystem; 200, data processing subsystem; 300, application service subsystem; 400, integrated management subsystem. Detailed implementation manners

[0071] Example 1

[0072] As Figure 1 shown, an integrated ground operation and application system for high-resolution remote sensing satellites includes:

[0073] Data reception subsystem 100: used for receiving and recording satellite payload data, simultaneously performing reception resource management and data quality monitoring, and transmitting the satellite payload data to data processing subsystem 200;

[0074] Data processing subsystem 200: used for receiving the satellite payload data transmitted by data reception subsystem 100, processing the data to generate data products, and then sending them to application service subsystem 300 and users, used for satellite calibration testing and quality assessment, used for receiving user requirements and forwarding them to integrated management subsystem 400, and sending the data products to application service subsystem 300;

[0075] Application Service Sub-system 300: It is used to receive the user requirements transmitted by the Integrated Management Sub-system 400, receive the data products transmitted by the Data Processing Sub-system 200, classify and process them to generate user-oriented application products, and conduct intelligent interpretation analysis, three-dimensional environment analysis, situation visualization and analysis, requirement analysis and planning, thematic product production, and multi-source data fusion while carrying out standard application product data production for data products;

[0076] Integrated Management Sub-system 400: It is used to communicate with the Data Reception Sub-system 100, the Data Processing Sub-system 200, and the Application Service Sub-system 300, upload and download data, receive the user requirements sent by the Data Processing Sub-system 200 and forward them to the Application Service Sub-system 300, and conduct user management, requirement management, task consultation and decision-making, task planning, plan management, system monitoring, integrated display and control, and external coordination;

[0077] Based on the ground operation and application integrated system for high-resolution remote sensing satellites, driven by task requirements, an automated task planning and scheduling strategy is constructed to achieve task conflict resolution and optimal planning. Secondly, a training set of typical targets is established, and target intelligent interpretation is realized based on machine learning, effectively improving the timeliness and accuracy of target recognition and confirmation. Finally, data such as meteorology, hydrology, surveying and mapping, and basic maps are introduced to construct target information combined with the three-dimensional environment, and the overall situation perception of the scene and the target is realized after visualization;

[0078] The Data Reception Sub-system 100 includes:

[0079] Ground Receiving Station 110: It is used to receive and record the satellite payload data transmitted by the satellite, and transmit the satellite payload data to the Data Transmission Module 120;

[0080] Data Transmission Module 120: It is used to receive the satellite payload data transmitted by the Ground Receiving Station 110 and transmit it to the Data Management and Monitoring Sub-system 130 and the Data Processing Sub-system 200;

[0081] Data Management and Monitoring Sub-system 130: It is used to receive the satellite payload data transmitted by the Data Transmission Module 120 and conduct receiving resource management and receiving quality monitoring;

[0082] The data processing performed by the Data Processing Sub-system 200 on the satellite payload data includes decompression, data preprocessing, cataloging and archiving, quality inspection, radiation correction, geometric correction, and production of 1-4 level product standard images;

[0083] As Figures 2-3 shown, the automated task planning and scheduling strategy includes the following steps:

[0084] S1. Intelligent planning and scheduling of observation tasks: The integrated management subsystem 400 selects observation tasks, allocates observation resources, and formulates an observation plan based on the user requirements forwarded by the data processing subsystem 200, using the task requirement degree, resource competition degree, time window requisition degree, and redundancy amount, and then sends the observation plan to the satellite.

[0085] S11. Preparation: The integrated management subsystem 400 receives the user requirements forwarded by the data processing subsystem 200.

[0086] S12. Task selection: Calculate the task requirement degree Need i for observation task l i and sort by urgency.

[0087] S13. Resource selection: Calculate the resource competition degree Contention ij and the time window contention degree Con IC for observation resource allocation.

[0088] S14. Task synthesis: Calculate the redundancy amount for the selection of the task synthesis scheme.

[0089] S15. Formulate an observation plan: Formulate an observation plan and send it to the satellite.

[0090] As Figure 4 shown, S2. Intelligent planning and scheduling of data transmission tasks: The integrated management subsystem 400 solves the data transmission scheduling sub-problem to determine the observation data transmission order, transmission object, and formulates a transmission plan, and then sends the transmission polarization to the data receiving subsystem 100.

[0091] As Figure 5 shown, the realization of target intelligent interpretation based on machine learning includes the following: Establish a training set of typical targets for data interpretation and intelligence interpretation, and conduct knowledge mining and training learning based on the training set; after obtaining the satellite payload data, combine the satellite payload data samples to train the knowledge utilization target recognition algorithm and the classification recognition algorithm based on cognitive computing for target detection, discovery, and recognition confirmation.

[0092] The construction of target information combined with the three-dimensional environment includes the following steps:

[0093] SⅠ. Location analysis: The integrated management subsystem 400 analyzes and locates the satellite payload data interpreted and judged by the application service subsystem 300.

[0094] SⅡ. Extract environmental element information: Extract the environmental element information of the target scene from the environmental element database. The environmental element information includes meteorological information, hydrological information, surveying and mapping information, basic base map information, and ground data.

[0095] SⅢ. Fusion: Integrate satellite payload data with environmental element information to build a multi-element and multi-dimensional visualization platform that aggregates multi-scale spatial scenes and variable time series.

[0096] Embodiment 2

[0097] An integrated ground operation and application system for high-resolution remote sensing satellites. This system realizes the interconnection and interoperability of design, testing, operation, and application through process optimization and resource integration, achieving horizontal knowledge transfer, effectively improving the satellite design and application levels while increasing the reuse rate of equipment. Secondly, by integrating internal and external requirements and resources, a corresponding decision-making model is established to achieve the optimal regulation and allocation of satellite resources, receiving resources, and transmission resources. Then, new technologies such as artificial intelligence and big data are introduced to effectively improve the timeliness and accuracy of intelligence interpretation. Finally, information such as meteorology, hydrology, surveying and mapping, basic maps, and ground data is introduced based on intelligence interpretation and visually displayed to achieve all-element three-dimensional perception of targets and scenes.

[0098] I. System Scheme Design

[0099] The system composition is as Figure 1 shown, mainly including four major parts: integrated management, data reception, data processing, and application services. This system can achieve interconnection and interoperability with the overall design simulation, comprehensive testing, and operation control system, maximizing the satellite's usage efficiency while improving the satellite design level, and enabling the effective extension from satellite design and manufacturing to satellite application.

[0100] The integrated management subsystem 400 is the core of the ground system's operation management. Through functions such as user management, requirement management, task consultation and decision-making, task planning, plan management, system monitoring, comprehensive display and control, and external coordination, it realizes the whole-process task management of the high-resolution remote sensing satellite ground system from receiving various user requirements to providing users with various levels of products and services that meet user requirements, ensuring the efficient and business-like operation of the system.

[0101] The data reception subsystem 100 mainly completes the downlink reception of satellite payload data. It mainly consists of modules such as ground receiving stations, data transmission, reception management, and monitoring. In the future, to quickly build a global ground station network at a lower cost, a resource integration model can be adopted, and the construction of the station network can be gradually promoted through a combination of leasing, resource sharing, and self-construction.

[0102] The data processing subsystem 200 mainly completes data decompression, data preprocessing, cataloging and archiving, quality inspection, radiometric calibration, geometric calibration, and the production of standard image products at levels 1-4. At the same time, it completes the calibration test and quality assessment of the satellite. Meanwhile, it sends various products (including data products and application result products at all levels) to users, which is the portal interface for external users of the system. Therefore, it also undertakes the functions of receiving and forwarding user requirements.

[0103] The application service subsystem 300 mainly completes the production of application products for users. For the received data of high-resolution remote sensing satellites, while carrying out the production of standard application product data, it conducts research and verification on new technologies such as intelligent interpretation analysis, three-dimensional environment analysis, situation visualization and analysis, demand analysis and planning, thematic product production, and multi-source data fusion.

[0104] The system design adopts the idea of a multi-layer distributed structure design of "data layer / business layer / application layer". The multi-layer distributed application system architecture has advantages such as good reusability, good scalability, manageability, and easy maintainability, and can achieve information intercommunication, data sharing, and application integration in the system. In particular, on the basis of dividing the data layer, business layer, and application layer, extracting the process logic from the business logic helps with business process automation and business process reengineering, so that any modification of the logic in any layer will not affect other layers, thereby minimizing the internal coupling of the system, providing the intensive level of the entire system, improving the system's ability to adapt to changes, providing the basic data structure for the automated and intelligent operation of the entire ground application system, and thus effectively saving costs.

[0105] II. Task Conflict Resolution and Intelligent Planning

[0106] Task conflict resolution and intelligent planning mainly solve the problem of reasonably planning tasks by rationally allocating observation and data transmission resources after receiving observation requirements, effectively improving the response frequency. Task conflict resolution and intelligent planning are divided into intelligent planning of observation tasks and intelligent planning of data transmission tasks. The main goal of intelligent planning of observation tasks is to determine which tasks to observe, what resources to use for observation, and when to observe under the condition of the most observation tasks or the greatest observation benefits. The main goal of data transmission resource scheduling is to optimize and determine which observation data can be transmitted, to which ground station it is transmitted, and when to start transmission, etc. After the observation and data transmission task planning is completed, the final solution result needs to be evaluated. If it meets the optimization threshold, the iteration can be terminated; otherwise, multiple iterations are carried out until the conditions are met. Its flowchart is as Figure 2 shown.

[0107] (1) Intelligent Planning and Scheduling of Observation Tasks

[0108] The intelligent planning and scheduling of observation tasks mainly include task selection, resource selection, time window selection, and task synthesis.

[0109] During the task selection process, the task requirement degree Need i is defined to describe the urgency of task l i to be scheduled.

[0110]

[0111] Among them, w i is the priority of the task, and Opportunities i is the number of remaining observation opportunities for this task, that is, the number of current available time windows. The purpose of the task selection strategy is to give priority to arranging tasks with high value and few remaining observation opportunities.

[0112] During the resource selection process, the resource competition degree and time window contention degree are used as quantitative indicators to determine that resources such as the electric energy, storage capacity, and number of side-sway times of remote sensing satellites are limited and can only complete some tasks. Therefore, there is competition for resources among tasks. To express the degree of competition of tasks for satellite resources, the concept of resource competition degree is defined to describe the competition situation of task t i for satellite s i resources:

[0113]

[0114] Suppose the satellite has k types of resources. RequestedCapacity(i,r) represents the total demand of the candidate tasks of satellite s i for its r-type resources, Capacity(i,r) represents the remaining amount of the r-type resources of satellite s i , and Requires(i,j,r) is the demand of task t j for the r-type resources of satellite s i . The resource competition degree reflects the degree of competition of tasks for satellite resources. The larger the Contention ij , the smaller the resource competition degree of task t j for satellite s i . The smaller the Contention ij , the larger the resource competition degree of task t j for satellite s i . When selecting resources for tasks, satellites with small resource competition degrees should be selected as much as possible, which also meets the requirements of load balancing in multi-satellite scheduling.

[0115] Since a satellite can only adopt one attitude for imaging at a given moment, different tasks not only compete for resources such as electrical energy and storage on the satellite, but also compete in terms of time windows. In this project, the degree of time window contention is defined to measure the competition degree of tasks for time window K.

[0116]

[0117] Among them, m represents the number of unassigned tasks, and ω j represents the priority (value) of the unassigned task t j , ConflictSpan ICj is the conflict time of the current task, and ConflictWinSpan ICj represents the duration of the conflict. Using this strategy will try to arrange tasks into the time window with the least conflict with other tasks.

[0118] When the satellite images ground targets, it needs to store data on the on-board memory. After multiple tasks perform combined observations, there will be some redundant observation data. The principle of task combination is that the redundancy between tasks should be minimized.

[0119] Suppose the combined task contains k meta-tasks. The start time of the i-th meta-task is ws i , the end time is we i , and the duration is duration. The schematic diagram is as Figure 3 shown, and the redundancy of the combined scheme is defined as:

[0120]

[0121] The redundant part does not generate benefits and also occupies a certain amount of storage resources. The redundancy not only represents the waste of storage resources in the combined observation task, but also represents the waste of other resources such as the continuous power-on time and energy of the satellite. Therefore, a task combination scheme with a small redundancy should be selected as much as possible.

[0122] (2) Intelligent Planning and Scheduling of Data Transmission Tasks

[0123] The intelligent planning and scheduling of data transmission tasks mainly solves the problem of efficient and reliable transmission of observation data through the data transmission window. If each data transmission window and its possible task observation windows are regarded as a type of planning problem, then the intelligent planning and scheduling problem of data transmission tasks can be composed of multiple sub-problems, that is, to solve the problem of single data transmission window planning for multiple observation windows. The schematic diagram is as Figure 4 shown.

[0124] The problem is solved by using the label updating algorithm for the data transmission scheduling sub-problem. The label updating algorithm is a relatively widely used type of solution algorithm in the actual application process, with the advantages of short solution time, high solution quality, simple and stable algorithm, etc. First, a label array is reserved for each vertex of the same type, and a label array is reserved from the origin to each vertex. The optimal path information from the origin to each vertex is reserved in the label array of that vertex. The vertices in the graph are sequentially selected as candidate vertices, and for each label information reserved on the candidate vertices, the label paths of the vertices connected to this vertex by edges are sequentially updated and compared with the label arrays of the corresponding vertices. When a certain vertex becomes a candidate vertex, the information in its label array is no longer updated. The search stops until the end. After the algorithm ends, the optimized observation target sequence of the end point can be obtained according to the label set information of the end point.

[0125] III. Data Interpretation and Intelligence Interpretation Based on Machine Learning

[0126] Data interpretation and intelligence interpretation based on machine learning mainly solve the problems of low intelligence level, low timeliness and accuracy in current intelligence interpretation. The data interpretation and intelligence process based on machine learning is as Figure 5 shown. First, a training set of typical targets is established, and knowledge mining and training learning are carried out based on this training set. After obtaining the observation data, the target detection, discovery and recognition confirmation are realized by using the target recognition algorithm and the classification recognition algorithm based on cognitive computing in combination with the sample training knowledge.

[0127] The schematic diagram of the bidirectional cognitive computing model is as Figure 6As shown in the figure. Among them, the forward transformation process is to convert the connotation expressed by the acquired image into a conceptual extension that is convenient for computer storage and calculation and for feature recognition; while the reverse transformation process is to learn, abstract, or extract the knowledge, information, etc. contained in the acquired image to form a conceptual connotation for people to understand and recognize. Therefore, by performing the two-way cycle of forward transformation and reverse transformation multiple times, that is, generating quantitative data through the forward transformation algorithm for a qualitative concept, and then forming a qualitative concept through the reverse transformation algorithm, and repeating this cycle multiple times to simulate the two-way cognitive calculation process of humans for concepts. The cognitive calculation process of dynamic feeding is a process of obtaining corresponding knowledge or concepts from the existing samples or information as the samples or information volume of the given cognitive object increases, and studying the dynamics of human concept cognition through the comparison of the obtained concepts. The data interpretation and intelligence interpretation based on machine learning use the forward cognitive calculation transformation and different reverse cognitive calculation transformations to simulate the dynamics of different people's cognition of things or concepts. The specific algorithm process is as follows: for a given initial concept, first obtain different numbers of data samples (conceptual extensions) through the forward cognitive calculation transformation, and then use different reverse cognitive calculations (including sample random sampling, fixed sampling, and repeated sampling) to simulate different cognitive calculation abilities, so as to obtain the digital feature estimation values of the corresponding new concepts.

[0128] IV. Superposition and Display of Basic Data

[0129] The superposition display of basic data mainly solves the problems of single data structure of current intelligence information and limited decision-making support ability. This system has the ability to access environmental elements. First, it analyzes the results of remote sensing satellite data interpretation and interpretation, and after positioning, extracts the environmental element information of the target location scene from the environmental element database, mainly including information such as meteorology, hydrology, surveying and mapping, basic base map, and ground data. After fusing the target information and the scene information, it constructs a multi-element and multi-dimensional visualization platform that integrates multi-scale spatial scenes and variable time series to support application services such as situation awareness, data product customization, information value-added, and solution customization.

[0130] The above description is illustrative rather than restrictive for the present invention. Those of ordinary skill in the art understand that any modification, change, or equivalence made without departing from the spirit and scope defined by the claims will fall within the protection scope of the present invention.

Claims

1. An integrated system for ground operation and application of high-resolution remote sensing satellites, characterized in that: Including: Data receiving subsystem (100): It is used to receive and record satellite payload data, manage receiving resources and monitor data quality at the same time, and transmit the satellite payload data to the data processing subsystem (200); Data processing subsystem (200): It is used to receive the satellite payload data transmitted by the data receiving subsystem (100), process the data, generate data products, and then send them to the application service subsystem (300) and users. It is used to conduct calibration tests and quality assessments of satellites, receive user requirements and forward them to the integrated management subsystem (400), and send the data products to the application service subsystem (300); Application service subsystem (300): It is used to receive the user requirements transmitted by the integrated management subsystem (400), receive the data products transmitted by the data processing subsystem (200), classify and process them to generate the application products for users. While carrying out the production of standard application product data for the data products, it conducts intelligent interpretation analysis, three-dimensional environment analysis, situation visualization and analysis, demand analysis and planning, thematic product production, and multi-source data fusion; Integrated management subsystem (400): It is used to communicate with the data receiving subsystem (100), the data processing subsystem (200), and the application service subsystem (300), upload and download data. It is used to receive the user requirements sent by the data processing subsystem (200) and forward them to the application service subsystem (300). It is used to conduct user management, demand management, task consultation and decision-making, task planning, plan management, system monitoring, integrated display and control, and external coordination; Based on the ground operation and application integration system for high-resolution remote sensing satellites, driven by mission requirements, an automated task planning and scheduling strategy is constructed to achieve task conflict resolution and optimal planning. Secondly, a training set of typical targets is established, and target intelligent interpretation is realized based on machine learning, effectively improving the timeliness and accuracy of target recognition and confirmation. Finally, data such as meteorology, hydrology, surveying and mapping, and basic maps are introduced to construct target information combined with the three-dimensional environment, and the overall situation perception of the scene and the target is realized after visualization; The automated task planning and scheduling strategy includes the following steps: S1. Intelligent planning and scheduling of observation tasks: The integrated management subsystem (400) selects observation tasks, allocates observation resources, and formulates an observation plan according to the user requirements forwarded by the data processing subsystem (200), using task requirement degree, resource competition degree, time window requisition degree, and redundancy, and then sends the observation plan to the satellite; Task requirement degree Need i Describe task l i Urgency to be arranged: where w i is the priority of the task, Opportunities i is the number of remaining observation opportunities for the task, that is, the number of current available time windows; Resource Contention xy Describe task t x For satellite s x Resource contention situation: Suppose the satellite has k types of resources, and RequestedCapacity(x,z) is the total demand of candidate mission of satellite s x for type z resources, Capacity(x,z) is the remaining amount of type z resources of satellite s x , and Requires(x,y,z) is the demand of mission t y for type z resources of satellite s x ; Contention Degree of Time Window Con IC Measure the competition degree of a task for time window K; where m is the number of tasks not yet scheduled, ω j is the priority or preference value of the task t not yet scheduled j , and ConflictWinSpan ICj is the duration of the conflict; The redundancy is: Among them, A is the number of meta-tasks included in the synthesis task, and the start time of the a-th meta-task is ws a , and the end time is we a , and the duration is duration a ; S2. Intelligent planning and scheduling of data transmission tasks: The integrated management subsystem (400) solves the data transmission scheduling sub-problem to determine the observation data transmission order, transmission object, and formulates a transmission plan, and then sends the transmission polarization to the data receiving subsystem (100); The target intelligent interpretation based on machine learning includes the following: establishing a training set of typical targets for data interpretation and intelligence interpretation, and conducting knowledge mining and training learning based on the training set; after obtaining the satellite payload data, combining the satellite payload data samples to train knowledge utilization target recognition algorithms and classification recognition algorithms based on cognitive computing for target detection, discovery, and recognition confirmation; The construction of target information combined with the three-dimensional environment includes the following steps: SⅠ. Positioning analysis: The integrated management subsystem (400) analyzes and locates the satellite payload data interpreted and judged by the application service subsystem (300); SⅡ. Extracting environmental element information: Extracting environmental element information of the scene where the target is located from the environmental element database, and the environmental element information includes meteorological information, hydrological information, surveying and mapping information, basic base map information, and ground data; SⅢ. Fusion: Fusing the satellite payload data with the environmental element information to construct a multi-element and multi-dimensional visualization platform integrating multi-scale spatial scenes and variable time series.

2. The integrated ground operation and application system for high-resolution remote sensing satellites according to claim 1, wherein: The data receiving subsystem (100) includes: Ground receiving station (110): Used to receive and record the satellite payload data transmitted by the satellite, and used to transmit the satellite payload data to the data transmission module (120); Data transmission module (120): Used to receive the satellite payload data transmitted by the ground receiving station (110) and transmit it to the data management and monitoring subsystem (130) and the data processing subsystem (200); Data management and monitoring subsystem (130): Used to receive the satellite payload data transmitted by the data transmission module (120) and conduct receiving resource management and receiving quality monitoring.

3. The integrated ground operation and application system for high-resolution remote sensing satellites according to claim 1, wherein: The data processing performed by the data processing subsystem (200) on the satellite payload data includes decompression, data preprocessing, cataloging and archiving, quality inspection, radiometric correction, geometric correction, and production of 1-4 level product standard images.

4. The integrated ground operation and application system for high-resolution remote sensing satellites according to claim 1, wherein: Step S1 includes the following steps: S11. Preparation: The integrated management subsystem (400) receives the user requirements forwarded by the data processing subsystem (200); S12. Task Selection: Calculate the task requirement degree Need i to perform the observation task l i Sort by urgency; S13. Resource Selection: Calculate the resource contention degree Contention xy and the contention degree Con of the time window IC to perform observation resource allocation; S14. Task synthesis: Calculating the redundancy amount to select a task synthesis scheme; S15. Formulating an observation plan: Formulating an observation plan and sending it to the satellite.

Images