Remote sensing automatic identification system based on satellite positioning

Abstract

The invention provides a remote sensing automatic identification system based on satellite positioning, and relates to the technical field of remote sensing identification. The remote sensing automatic identification system based on satellite positioning comprises a database and storage module, a satellite positioning module, a biological inspiration algorithm identification, communication and transmission module, a user interface and display module, a multi-modal data fusion, data processing and analysis module, a remote sensing data acquisition module and a data acquisition module. Through edge calculation and real-time processing, the calculation capability is deployed on edge equipment for data acquisition, the real-time performance and efficiency of the system can be remarkably improved, multi-modal data fusion is combined with different types of data, optics, radar and infrared to provide more comprehensive information, the recognition precision is improved, and the recognition efficiency is improved. Quantum calculation can accelerate processing and analysis of remote sensing data, the real-time performance and efficiency of the system are improved, and the accuracy and efficiency of remote sensing image recognition can be remarkably improved by simulating a biological vision system and a cognitive mechanism through biological heuristic algorithm recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing identification, and specifically relates to a remote sensing automatic identification system based on satellite positioning. Background Technique

[0002] The satellite positioning system is a technology that uses satellites to accurately locate something. It has developed from the initial low positioning accuracy, inability to perform real-time positioning, and difficulty in providing timely navigation services to the current high-precision GPS global positioning system, enabling the simultaneous observation of 4 satellites at any time and any point on the earth to achieve functions such as navigation, positioning, and time service. Remote sensing automatic identification refers to using computer technology and algorithms to analyze and process remote sensing images to automatically identify and classify ground objects or targets in the images. The remote sensing automatic identification system based on satellite positioning can provide high-precision geographical information and target recognition capabilities.

[0003] In the process of using the existing remote sensing automatic identification system based on satellite positioning, there are still some problems. Currently, the amount of remote sensing data is large, and data transmission delay will affect real-time processing and analysis. Traditional calculation methods are difficult to process efficiently. Moreover, in a complex environment, the recognition ability may be limited, and single-source data may not provide comprehensive information, resulting in low recognition accuracy. Therefore, those skilled in the art have provided a remote sensing automatic identification system based on satellite positioning to solve the problems raised in the above background technique. Summary of the Invention

[0004] (1) Technical Problems to be Solved

[0005] In view of the deficiencies of the prior art, the present invention provides a remote sensing automatic identification system based on satellite positioning, which solves the problems that the current amount of remote sensing data is large, data transmission delay will affect real-time processing and analysis, traditional calculation methods are difficult to process efficiently, and in a complex environment, the recognition ability may be limited, and single-source data may not provide comprehensive information, resulting in low recognition accuracy.

[0006] (2) Technical Solutions

[0007] To achieve the above objectives, the present invention is realized through the following technical solutions: A remote sensing automatic identification system based on satellite positioning, including:

[0008] A database and storage module, a satellite positioning module, a bio-inspired algorithm for identification, a communication and transmission module, a user interface and display module, multi-modal data fusion, a data processing and analysis module, a remote sensing data acquisition module, and a data collection module;

[0009] The data collection module further includes edge computing and real-time processing. Deploying computing power on edge devices for data collection can reduce data transmission delay and improve real-time processing capabilities;

[0010] The data processing and analysis module also includes quantum computing and remote sensing data processing. Quantum computing has significant advantages in processing large-scale data and complex computing tasks. Through quantum algorithms, the processing and analysis of remote sensing data can be accelerated, improving the real-time performance and efficiency of the system;

[0011] The multi-modal data fusion combines different types of data, such as optical, radar, and infrared, for fusion analysis, which can provide more comprehensive information and improve the recognition accuracy;

[0012] The bio-inspired algorithm recognition also includes visual cognitive models, deep learning, and bio-inspiration. The visual cognitive model, the HMAX model, a multi-layer cognitive model, simulates the structure and function of the human visual cortex. By introducing associative and memory mechanisms, the HMAX model can achieve higher-precision object recognition and classification. It simulates the attention mechanism in the human visual system and selects the initial feature regions by generating a saliency map, thereby improving the recognition efficiency.

[0013] The database and storage module is responsible for storing and managing a large amount of remote sensing data and recognition results, supporting the rapid retrieval and analysis of data;

[0014] The satellite positioning module provides accurate geolocation information and achieves high-precision target positioning through the satellite navigation system;

[0015] The communication and transmission module is responsible for data transmission and communication, ensuring that remote sensing data and recognition results can be efficiently and securely transmitted to the user terminal or data center;

[0016] The user interface and display module provides a user-friendly interface for displaying remote sensing images, recognition results, and location information, supporting interactive operations and visual displays;

[0017] Preferably, for the edge computing and real-time processing, deploying the computing power on the edge devices for data collection can reduce the data transmission delay and improve the real-time processing ability. The detailed steps are as follows:

[0018] A1. Data collection, sensor data collection, real-time collection of remote sensing data on the Earth's surface through optical remote sensing sensors, infrared remote sensing sensors, and radar remote sensing sensors;

[0019] A2. Edge computing preprocessing, data preprocessing is performed on the edge devices, including denoising, calibration, and preliminary feature extraction. The algorithm formula is as follows:

[0020] D l =f pre (D(t))

[0021] where f preIndicates a preprocessing function;

[0022] A3. Real-time processing and analysis, using edge computing devices for real-time processing and analysis, object detection, classification, and recognition. The algorithm formula of its analysis structure R(t) is as follows:

[0023] R(t) = f ana (D l (t))

[0024] Where f ana Indicates an analysis function;

[0025] A4. Data transmission, transmitting the processed data and analysis results to the data processing and analysis module.

[0026] Preferably, the edge computing and real-time processing algorithm includes a convolutional neural network. The convolutional neural network performs a convolutional operation on the input data through a convolutional operation to extract a feature map. Its feature map F conv The algorithm formula is as follows:

[0027] F conv = Comv(D l (t), W conv )

[0028] Where W conv Indicates a convolutional kernel. Edge computing and real-time processing deploy computing power on edge devices for data acquisition, which can significantly improve the real-time performance and efficiency of the system. By performing data preprocessing and real-time processing on edge devices, data transmission latency is reduced, and real-time response capabilities are improved. The convolutional neural network can perform real-time object detection and classification on edge devices. Preprocessing and preliminary analysis are carried out on edge devices, reducing the burden on the central server and improving the overall efficiency of the system. Through distributed computing and real-time processing, the robustness and reliability of the system are enhanced, adapting to complex and changing application scenarios.

[0029] Preferably, quantum computing has significant advantages in processing large-scale data and complex computing tasks. Through quantum algorithms, the processing and analysis of remote sensing data can be accelerated, and the detailed steps to improve the real-time performance and efficiency of the system are as follows:

[0030] B1. Data preprocessing, quantum data preprocessing uses quantum algorithms to preprocess remote sensing data, including denoising, calibration, and feature extraction. The preprocessed data D q (t) formula is as follows:

[0031] D q (t) = Q pre (D(t))

[0032] Where Q preRepresents a quantum preprocessing algorithm;

[0033] B2. Quantum computing processing. The quantum algorithm processing uses quantum algorithms to perform complex computing tasks, such as large-scale data analysis and multi-source data fusion. The processing result R q (t) has the following formula:

[0034] R q (t) = Q ana (D q (t))

[0035] where Q ana represents a quantum analysis algorithm;

[0036] B3. Data transmission and storage. Quantum data transmission transfers the processed data and analysis results to the central server or user terminal.

[0037] Preferably, the quantum computing and remote sensing data processing algorithm includes a quantum support vector machine, which is used to classify and identify targets in remote sensing data. The classification result C q has the following algorithm formula:

[0038] C q = QSVM(D q (t), W q )

[0039] where, W q represents the quantum weight. Quantum computing can accelerate the processing and analysis of remote sensing data, improve the real-time performance and efficiency of the system. Quantum computing can significantly improve the data processing speed. Quantum algorithms such as quantum support vector machines can improve the accuracy of classification and identification and adapt to complex and changing application scenarios. Quantum computing can handle large-scale data and complex computing tasks that are difficult for traditional computers to solve and support more complex analysis and identification tasks.

[0040] Preferably, the multi-modal data fusion combines different types of data, such as optical, radar, and infrared, for fusion analysis, which can provide more comprehensive information and improve the recognition accuracy. The detailed steps are as follows:

[0041] C1. Data preprocessing. The optical data preprocessing performs denoising, calibration, and geometric calibration on optical remote sensing data, performs denoising, coherence processing, and geometric calibration on radar data, and performs denoising, temperature calibration, and geometric calibration on infrared data;

[0042] C2. Feature extraction. The optical feature extraction extracts spectral features and texture features from the preprocessed optical data, extracts backscattering coefficient and coherence coefficient features from the preprocessed radar data, and extracts temperature features and thermal radiation features from the preprocessed infrared data;

[0043] C3. Data fusion: Multimodal data fusion combines optical, radar and infrared feature data for fusion analysis to improve recognition accuracy;

[0044] C4. Result output, target recognition and classification use the fused feature data for target recognition and classification. Multimodal data fusion combines different types of data. Optical, radar and infrared can provide more comprehensive information and improve recognition accuracy. Multimodal data fusion can combine the advantages of different sensors to provide more comprehensive surface information. Optical data provides spectral information, radar data provides terrain information, and infrared data provides temperature information. Through fusion analysis, it can improve recognition accuracy and robustness, adapt to complex and changeable application scenarios, support the fusion of multiple sensor data, and adapt to different task requirements and environmental conditions.

[0045] Preferably, the visual cognitive model, deep learning and biological inspiration, visual cognitive model, HMAX model multi-layer cognitive model, simulates the structure and function of the human visual cortex, and by introducing association and memory mechanisms, the HMAX model can achieve higher-precision object recognition and classification, simulates the attention mechanism in the human visual system, and selects the initial feature area by generating a saliency map, thereby improving the recognition efficiency. The detailed steps are as follows:

[0046] S1. Multi-layer cognitive structure. The HMAX model simulates the structure and function of the human visual cortex and realizes object recognition and classification through multi-layer processing. The HMAX model preprocesses the multi-layer cognitive input data to obtain the initial feature map F0(t). Through multi-layer convolution kernel pooling operations, features are gradually extracted. The l-th layer feature map F l (t) = ConvPool(F l-1 (t), W l ), where W l Represents the convolution kernel and pooling parameters of the lth layer, and finally obtains the high-level feature map F L (t), for object recognition and classification;

[0047] S2. Attention mechanism, saliency map generation, by generating a saliency map, selecting the initial feature area, improving the recognition efficiency, the attention mechanism generates a saliency map S(t) = f att (D(t)), select the initial feature area, and select the key area for feature extraction according to the saliency map to improve the recognition efficiency;

[0048] S3. Memory mechanism. Associative memory utilizes the memory mechanism to enhance feature extraction and recognition capabilities. The memory mechanism combines the current features and historical memory features to enhance feature extraction and recognition capabilities through the memory mechanism. The bio-inspired algorithm recognition can significantly improve the accuracy and efficiency of remote sensing image recognition by simulating the biological vision system and cognitive mechanism. By simulating the structure and function of the human visual cortex, the bio-inspired algorithm can achieve higher-precision object recognition and classification. The HMAX model can effectively recognize targets in complex backgrounds through a multi-layer cognitive structure and associative memory mechanism. The bio-inspired algorithm generates a saliency map through the attention mechanism to select the initial feature region, thereby improving the recognition efficiency. By the memory mechanism, the feature extraction and recognition capabilities are enhanced to adapt to complex and variable application scenarios.

[0049] (III) Beneficial effects

[0050] The present invention provides a remote sensing automatic recognition system based on satellite positioning, which has the following beneficial effects:

[0051] 1. In the present invention, edge computing and real-time processing deploy the computing power on the edge devices for data acquisition, which can significantly improve the real-time performance and efficiency of the system. By performing data preprocessing and real-time processing on the edge devices, the data transmission delay is reduced, and the real-time response ability is improved. The convolutional neural network can perform real-time object detection and classification on the edge devices. The preprocessing and preliminary analysis are carried out on the edge devices, which reduces the burden on the central server and improves the overall efficiency of the system. Through distributed computing and real-time processing, the robustness and reliability of the system are enhanced to adapt to complex and variable application scenarios.

[0052] 2. In the present invention, multi-modal data fusion combines different types of data. Optical, radar, and infrared can provide more comprehensive information to improve the recognition accuracy. Multi-modal data fusion can combine the advantages of different sensors to provide more comprehensive surface information. Optical data provides spectral information, radar data provides terrain information, and infrared data provides temperature information. Through fusion analysis, the recognition accuracy and robustness can be improved to adapt to complex and variable application scenarios, support the fusion of multiple sensor data, and adapt to different task requirements and environmental conditions.

[0053] 3. In the present invention, quantum computing can accelerate the processing and analysis of remote sensing data, improve the real-time performance and efficiency of the system. Quantum computing can significantly improve the data processing speed. Quantum algorithms such as quantum support vector machines can improve the accuracy of classification and recognition, adapt to complex and variable application scenarios. Quantum computing can process large-scale data and complex computing tasks that are difficult for traditional computers to solve, and support more complex analysis and recognition tasks.

[0054] 4. In the present invention, the bio-inspired algorithm for recognition can significantly improve the accuracy and efficiency of remote sensing image recognition by simulating the biological vision system and cognitive mechanism. By simulating the structure and function of the human visual cortex, the bio-inspired algorithm can achieve higher-precision object recognition and classification. The HMAX model can effectively recognize targets in complex backgrounds through a multi-layer cognitive structure and an associative memory mechanism. The bio-inspired algorithm generates a saliency map through an attention mechanism and selects an initial feature region, thereby improving the recognition efficiency. Through a memory mechanism, the feature extraction and recognition capabilities are enhanced to adapt to complex and changing application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic diagram of the system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0057] Embodiment 1:

[0058] As Figure 1 shown, the embodiment of the present invention provides a remote sensing automatic recognition system based on satellite positioning, including:

[0059] a database and storage module, a satellite positioning module, a bio-inspired algorithm for recognition, a communication and transmission module, a user interface and display module, multi-modal data fusion, a data processing and analysis module, a remote sensing data acquisition module, and a data collection module;

[0060] The data collection module further includes edge computing and real-time processing. Deploying computing capabilities on edge devices for data collection can reduce data transmission latency and improve real-time processing capabilities;

[0061] The data processing and analysis module further includes quantum computing and remote sensing data processing. Quantum computing has significant advantages in processing large-scale data and complex computing tasks. Through quantum algorithms, the processing and analysis of remote sensing data can be accelerated, improving the real-time performance and efficiency of the system;

[0062] Multi-modal data fusion combines different types of data, such as optical, radar, and infrared, for fusion analysis, which can provide more comprehensive information and improve recognition accuracy;

[0063] Biologically inspired algorithm recognition also includes visual cognitive models, deep learning, and biological inspiration. Visual cognitive models, such as the HMAX model and multi-layer cognitive models, simulate the structure and function of the human visual cortex. By introducing association and memory mechanisms, the HMAX model can achieve higher-precision object recognition and classification. It simulates the attention mechanism in the human visual system and selects initial feature regions by generating saliency maps, thereby improving recognition efficiency.

[0064] The database and storage module is responsible for storing and managing a large amount of remote sensing data and recognition results, and supports rapid retrieval and analysis of data.

[0065] The satellite positioning module provides accurate geolocation information and achieves high-precision target positioning through satellite navigation systems.

[0066] The communication and transmission module is responsible for data transmission and communication, ensuring that remote sensing data and recognition results can be efficiently and securely transmitted to user terminals or data centers.

[0067] The user interface and display module provides a user-friendly interface for displaying remote sensing images, recognition results, and location information, and supports interactive operations and visual displays.

[0068] Edge computing and real-time processing, deploying computing power on edge devices for data collection, can reduce data transmission latency and improve real-time processing capabilities. The detailed steps are as follows:

[0069] A1. Data collection, sensor data collection, real-time collection of remote sensing data on the Earth's surface through optical remote sensing sensors, infrared remote sensing sensors, and radar remote sensing sensors.

[0070] A2. Edge computing preprocessing, performing data preprocessing on edge devices, including denoising, calibration, and preliminary feature extraction. The algorithm formula is as follows:

[0071] D l =f pre (D(t))

[0072] Where f pre represents the preprocessing function;

[0073] A3. Real-time processing and analysis, using edge computing devices for real-time processing and analysis, such as target detection, classification, and recognition. The analysis structure R(t) algorithm formula is as follows:

[0074] R(t)=f ana (D l (t))

[0075] Where f ana represents the analysis function;

[0076] A4. Data transmission, transmitting the processed data and analysis results to the data processing and analysis module.

[0077] Preferably, the edge computing and real-time processing algorithm includes a convolutional neural network. The convolutional neural network performs a convolutional operation on the input data through a convolutional operation to extract a feature map, and its feature map F conv The algorithm formula is as follows:

[0078] F conv = Comv(D l (t), W conv )

[0079] Among them, W conv represents the convolutional kernel. Edge computing and real-time processing deploy computing power on edge devices for data acquisition, which can significantly improve the real-time performance and efficiency of the system. By performing data preprocessing and real-time processing on edge devices, data transmission latency is reduced, and real-time response capabilities are improved. The convolutional neural network can perform real-time object detection and classification on edge devices. Preprocessing and preliminary analysis are carried out on edge devices, reducing the burden on the central server and improving the overall efficiency of the system. Through distributed computing and real-time processing, the robustness and reliability of the system are enhanced, adapting to complex and changing application scenarios.

[0080] Quantum computing has significant advantages in processing large-scale data and complex computing tasks. Through quantum algorithms, the processing and analysis of remote sensing data can be accelerated, and the detailed steps to improve the real-time performance and efficiency of the system are as follows:

[0081] B1. Data preprocessing, quantum data preprocessing uses quantum algorithms to preprocess remote sensing data, including denoising, calibration, and feature extraction. The preprocessed data D q (t) formula is as follows:

[0082] D q (t)= Q pre (D(t))

[0083] Among them, Q pre represents the quantum preprocessing algorithm;

[0084] B2. Quantum computing processing, quantum algorithm processing uses quantum algorithms to perform complex computing tasks, such as large-scale data analysis and multi-source data fusion. The processing result R q (t) formula is as follows:

[0085] R q (t)= Q ana (D q (t))

[0086] Among them, Q ana represents the quantum analysis algorithm;

[0087] B3. Data transmission and storage. Quantum data transmission transfers the processed data and analysis results to the central server or user terminals.

[0088] Quantum computing and remote sensing data processing algorithms include quantum support vector machines. Quantum support vector machines are used for classifying and identifying targets in remote sensing data, and its classification result C q has the following algorithm formula:

[0089] C q = QSVM(D q (t), W q )

[0090] where W q represents the quantum weight. Quantum computing can accelerate the processing and analysis of remote sensing data, improve the real-time performance and efficiency of the system. Quantum computing can significantly increase the data processing speed. Quantum algorithms such as quantum support vector machines can improve the accuracy of classification and identification, and adapt to complex and variable application scenarios. Quantum computing can handle large-scale data and complex computing tasks that are difficult for traditional computers to solve, and support more complex analysis and identification tasks.

[0091] Multimodal data fusion combines different types of data, such as optical, radar, and infrared, for fusion analysis, which can provide more comprehensive information. The detailed steps to improve the identification accuracy are as follows:

[0092] C1. Data preprocessing. Optical data preprocessing performs denoising, calibration, and geometric calibration on optical remote sensing data, denoising, coherence processing, and geometric calibration on radar data, and denoising, temperature calibration, and geometric calibration on infrared data;

[0093] C2. Feature extraction. Optical feature extraction extracts spectral features and texture features from the preprocessed optical data, backscattering coefficient and coherence coefficient features from the preprocessed radar data, and temperature features and thermal radiation features from the preprocessed infrared data;

[0094] C3. Data fusion. Multimodal data fusion combines optical, radar, and infrared feature data for fusion analysis to improve the identification accuracy;

[0095] C4. Result Output, Target Recognition and Classification Utilize the fused feature data for target recognition and classification. Multimodal data fusion combines different types of data. Optics, radar, and infrared can provide more comprehensive information, improving the recognition accuracy. Multimodal data fusion can integrate the advantages of different sensors, providing more comprehensive surface information. Optical data provides spectral information, radar data provides terrain information, and infrared data provides temperature information. Through fusion analysis, the recognition accuracy and robustness can be improved to adapt to complex and changing application scenarios, support the fusion of multiple sensor data, and adapt to different task requirements and environmental conditions.

[0096] Visual Cognition Model, Deep Learning, and Bio-Inspiration Visual Cognition Model, HMAX Model (a multi-layer cognitive model) simulates the structure and function of the human visual cortex. By introducing association and memory mechanisms, the HMAX model can achieve higher-precision object recognition and classification. The detailed steps of simulating the attention mechanism in the human visual system to improve the recognition efficiency by generating a saliency map to select the initial feature region are as follows:

[0097] S1. Multi-Layer Cognitive Structure The HMAX model simulates the structure and function of the human visual cortex and realizes object recognition and classification through multi-layer processing. The HMAX model preprocesses the input data of multi-layer cognition to obtain the initial feature map F0(t). Through multi-layer convolution kernel pooling operations, features are gradually extracted. The feature map Fl(t) of the l-th layer = ConvPool(Fl−1(t), Wl), where Wl represents the convolution kernel and pooling parameters of the l-th layer. Finally, the high-level feature map Fh(t) is obtained for object recognition and classification; l (t) = ConvPool(F l-1 (t), W l ), where, W l represents the convolution kernel and pooling parameters of the l-th layer, and finally the high-level feature map F L (t) is obtained for object recognition and classification;

[0098] S2. Attention Mechanism, Saliency Map Generation Generate a saliency map to select the initial feature region and improve the recognition efficiency. The attention mechanism generates a saliency map S(t) = f att (D(t)), select the initial feature region, and according to the saliency map, select the key region for feature extraction to improve the recognition efficiency;

[0099] S3. Memory mechanism. Associative memory utilizes the memory mechanism to enhance feature extraction and recognition capabilities. The memory mechanism combines the current features and historical memory features, and through the memory mechanism, enhances feature extraction and recognition capabilities. The bio-inspired algorithm recognition can significantly improve the accuracy and efficiency of remote sensing image recognition by simulating the biological visual system and cognitive mechanism. By simulating the structure and function of the human visual cortex, the bio-inspired algorithm can achieve higher-precision object recognition and classification. The HMAX model can effectively recognize targets in complex backgrounds through a multi-layer cognitive structure and associative memory mechanism. The bio-inspired algorithm generates a saliency map through the attention mechanism and selects the initial feature region, thereby improving the recognition efficiency. Through the memory mechanism, it enhances feature extraction and recognition capabilities and adapts to complex and variable application scenarios.

[0100] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote sensing automatic identification system based on satellite positioning, characterized in that: Including: Database and storage module, satellite positioning module, bio-inspired algorithm recognition, communication and transmission module, user interface and display module, multi-modal data fusion, data processing and analysis module, remote sensing data acquisition module, and data collection module; The data collection module further includes edge computing and real-time processing. By deploying computing power on edge devices for data collection, it can reduce data transmission latency and improve real-time processing capabilities; The data processing and analysis module further includes quantum computing and remote sensing data processing. Quantum computing has significant advantages in processing large-scale data and complex computing tasks. Through quantum algorithms, it can accelerate the processing and analysis of remote sensing data, improving the real-time performance and efficiency of the system; The multi-modal data fusion combines different types of data, such as optical, radar, and infrared, for fusion analysis, which can provide more comprehensive information and improve recognition accuracy; The bio-inspired algorithm recognition also includes visual cognitive models, deep learning, and bio-inspiration. The visual cognitive model, the HMAX model, a multi-layer cognitive model, simulates the structure and function of the human visual cortex. By introducing association and memory mechanisms, the HMAX model can achieve higher-precision object recognition and classification. It simulates the attention mechanism in the human visual system and selects initial feature regions by generating saliency maps, thereby improving recognition efficiency.

2. The remote sensing automatic recognition system based on satellite positioning according to claim 1, characterized in that: The detailed steps for the edge computing and real-time processing, which deploy computing power on edge devices for data collection to reduce data transmission latency and improve real-time processing capabilities, are as follows: A1. Data collection, sensor data collection, real-time collection of remote sensing data on the Earth's surface through optical remote sensing sensors, infrared remote sensing sensors, and radar remote sensing sensors; A2. Edge computing preprocessing, data preprocessing on edge devices, including denoising, calibration, and preliminary feature extraction. The algorithm formulas are as follows: D l = f pre (D(t)) where f pre represents a preprocessing function; A3. Real-time processing and analysis, using edge computing devices for real-time processing and analysis, target detection, classification, and recognition. The analysis structure R(t) algorithm formula is as follows: R(t) = f ana (Dl(t)) where f ana represents an analysis function; A4. Data transmission, transmitting the processed data and analysis results to the data processing and analysis module.

3. The remote sensing automatic identification system based on satellite positioning according to claim 2, characterized in that: The edge computing and real-time processing algorithm includes a convolutional neural network, which performs a convolutional operation on the input data through a convolutional operation to extract a feature map, and its feature map F conv The algorithm formula is as follows: F conv = Comv(D l (t), W conv ) Among them, W conv represents the convolution kernel.

4. A remote sensing automatic identification system based on satellite positioning according to claim 1, characterized in that: The detailed steps for quantum computing, which has significant advantages in processing large-scale data and complex computing tasks and can accelerate the processing and analysis of remote sensing data through quantum algorithms to improve the real-time performance and efficiency of the system, are as follows: B1. Data preprocessing. Quantum data preprocessing uses quantum algorithms to preprocess remote sensing data, including denoising, calibration, and feature extraction. The preprocessed data D q (t) is given by the following formula: D q Q(t) = Q pre (D(t)) Among which Q pre represents a quantum preprocessing algorithm; B2. Quantum computing processing, where quantum algorithm processing utilizes quantum algorithms for complex computational tasks such as large-scale data analysis and multi-source data fusion, and the processing result R q (t) has the following formula: R q r(t) = Q ana (D q (t)) Among which Q ana represents a quantum analysis algorithm; B3. Data transmission and storage, quantum data transmission transmits the processed data and analysis results to the central server or user terminal.

5. The remote sensing automatic recognition system based on satellite positioning according to claim 4, wherein: The quantum computing and remote sensing data processing algorithm includes a quantum support vector machine, which is used to classify and identify targets in remote sensing data, and its classification result C q has the following algorithm formula: C q = QSVM(D q (t), W q ) Among them, W q represents the quantum weight.

6. The remote sensing automatic recognition system based on satellite positioning according to claim 1, characterized in that: The detailed steps for the multi-modal data fusion, which combines different types of data, such as optical, radar, and infrared, for fusion analysis to provide more comprehensive information and improve recognition accuracy, are as follows: C1. Data preprocessing, optical data preprocessing performs denoising, calibration, and geometric calibration on optical remote sensing data, denoising, coherence processing, and geometric calibration on radar data, and denoising, temperature calibration, and geometric calibration on infrared data; C2. Feature extraction. For optical feature extraction, spectral features and texture features are extracted from the preprocessed optical data, backscattering coefficient and coherence coefficient features are extracted from the preprocessed radar data, and temperature features and thermal radiation features are extracted from the preprocessed infrared data. C3. Data fusion. Multimodal data fusion combines optical, radar, and infrared feature data for fusion analysis to improve the recognition accuracy. C4. Result output. Object recognition and classification are performed using the fused feature data.

7. The remote sensing automatic recognition system based on satellite positioning according to claim 1, wherein: The visual cognitive model, deep learning, and biological inspiration. The visual cognitive model, the HMAX model, a multi-layer cognitive model, simulates the structure and function of the human visual cortex. By introducing an association and memory mechanism, the HMAX model can achieve higher-precision object recognition and classification. To simulate the attention mechanism in the human visual system, the initial feature region is selected by generating a saliency map, and the detailed steps to improve the recognition efficiency are as follows: S1. Multilayer cognitive structure. The HMAX model simulates the structure and function of the human visual cortex and realizes object recognition and classification through multilayer processing. The HMAX model preprocesses the input data of multilayer cognition to obtain the initial feature map F0(t). Through the operations of multilayer convolutional kernels and pooling, features are gradually extracted. The feature map Fl(t) = ConvPool(Fl-1(t), Wl), where Wl represents the convolutional kernel and pooling parameters of the l-th layer. Finally, the high-level feature map Fh(t) is obtained for object recognition and classification; l (t) = ConvPool(F l-1 (t), W l ), where W l represents the convolutional kernel and pooling parameters of the l-th layer, and finally the high-level feature map F L (t) is obtained for object recognition and classification; S2. Attention mechanism, saliency map generation. By generating a saliency map, the initial feature region is selected to improve the recognition efficiency. The attention mechanism generates the saliency map S(t) = f att (D(t)), selects the initial feature region, and according to the saliency map, selects the key region for feature extraction to improve the recognition efficiency; S3. Memory mechanism. Associative memory uses the memory mechanism to enhance the feature extraction and recognition capabilities. The memory mechanism combines the current features and historical memory features to enhance the feature extraction and recognition capabilities through the memory mechanism.

Images