Analysis method and device for on-orbit measurement error of optical sensor

Abstract

The invention provides an optical sensor on-orbit measurement error analysis method and device, and the method comprises the following steps: obtaining and analyzing various error parameters influencing the measurement precision of a sensor, and forming a measurement error feature space; classifying the various error parameters, determining the influence relationship of different error parameters on the measurement accuracy of the sensor, and then constructing an error model; based on the error model, in combination with the in-orbit operation condition of the deep space optical sensor, constructing an interactive coupling form between error parameters, determining influence coefficients of the error parameters under different conditions, and in combination with a fuzzy comprehensive evaluation method, constructing a coupling analysis model between different error parameters; optimizing the error model by using a self-adaptive orthogonal neighborhood, eliminating redundant features in a measurement error feature space, and constructing an error reduction model; and realizing on-orbit measurement of the optical sensor by using the error reduction model. And technical support is provided for efficient identification of system errors and complete estimation of navigation states.

Description

Technical Field

[0001] This invention relates to the field of on-orbit measurement error technology, and specifically to a method and apparatus for analyzing on-orbit measurement errors of optical sensors. Background Technology

[0002] Before a deep space probe is launched from Earth, its onboard sensors undergo precise calibration on the ground. However, factors such as vibrations during launch and the complex space environment can cause various errors to affect the optical sensors during on-orbit measurements, altering the error characteristics of the measurement data, reducing the accuracy of optical measurement data, and ultimately limiting the improvement of the accuracy of the optical autonomous navigation system. Therefore, it is essential to fully understand the error mechanism of optical sensors during on-orbit measurement, refine the modeling of measurement errors, and perform targeted on-orbit calibration and compensation for these errors.

[0003] Optical sensor measurement errors can be categorized into internal and external errors based on their sources. Internal errors primarily arise from factors such as stellar vibrations, uneven solar illumination, and high-energy particle radiation in space, causing translational, tilting, and rotation of the imaging plane, optical distortion, and focal length changes. External errors mainly refer to sensor installation errors and reference errors caused by the on-orbit motion environment of the detector. These errors can be broadly classified into systematic errors, random errors, and other errors. Systematic errors are inherent deviations between measured and true values, including sensor installation errors, measurement imaging deviations, and low-frequency errors. Installation errors and imaging deviations mainly stem from stellar vibrations, optical distortion, and changes in the imaging plane's position. Low-frequency errors are primarily caused by thermal radiation variations due to uneven solar illumination, exhibiting periodic changes similar to the orbital period. Random errors are those that follow certain statistical laws, including thermal noise, shot noise, and photoelectric sampling errors. Other errors mainly include ephemeris errors, parallax, and target relative motion errors. These error types severely restrict the measurement accuracy of optical sensors, thereby significantly reducing the accuracy of optical autonomous navigation.

[0004] In the field of optical sensor error analysis and modeling, researchers both domestically and internationally have conducted in-depth studies. Qin Shiqiao et al. analyzed the impact of celestial body and environmental vibrations on optical sensor imaging errors under complex dynamic environments and provided corresponding error characterization models. Current on-orbit measurement error mechanism analysis of optical sensors mainly focuses on a single key influencing factor, analyzing the sensor measurement error, but the analysis of the generation mechanism of measurement errors is not comprehensive enough. The correlation between measurement errors and system state variables, and the coupling between different error factors, are not sufficiently studied. The comprehensive models for optical sensor measurement errors are complex and diverse, requiring computational methods such as neural networks, resulting in heavy error modeling and processing loads, making them unsuitable for deep space probes with severely limited onboard resources.

[0005] In summary, the existing technology has the following problems: the current analysis of the error mechanism of optical sensor on-orbit measurement is not applicable to deep space probes with severely limited spaceborne resources, resulting in insufficient on-orbit measurement accuracy of optical sensor, which in turn restricts the improvement of the accuracy of optical autonomous navigation system and cannot solve the high-precision requirements of optical measurement in deep space exploration. Summary of the Invention

[0006] The purpose of this invention is to solve the problem of how to improve the accuracy of on-orbit measurement of optical sensors, thereby improving the accuracy of optical autonomous navigation systems and addressing the high-precision requirements of optical measurements in deep space exploration.

[0007] To this end, in one aspect, embodiments of the present invention provide a method for analyzing on-orbit measurement errors of optical sensors, the method comprising the following steps:

[0008] Acquire and analyze various error parameters that affect the measurement accuracy of the sensor to form a measurement error feature space;

[0009] The various error parameters are classified, and the influence of different error parameters on the measurement accuracy of the sensor is determined. Then, an error model is constructed.

[0010] Based on the aforementioned error model and combined with the on-orbit operation of the deep space optical sensor, an interactive coupling form between error parameters is constructed, the influence coefficients of error parameters under different conditions are determined, and a coupling analytical model between different error parameters is constructed by combining the fuzzy comprehensive evaluation method.

[0011] The error model is optimized by adaptive orthogonal neighborhood, and redundant features in the measurement error feature space are eliminated to achieve the fusion and reduction of error features and construct an error reduction model.

[0012] The aforementioned error reduction model is used to achieve on-orbit measurement of the optical sensor.

[0013] On the other hand, embodiments of the present invention provide an analysis apparatus for on-orbit measurement errors of optical sensors, comprising:

[0014] The acquisition unit is used to acquire and analyze various error parameters that affect the measurement accuracy of the sensor, forming a measurement error feature space;

[0015] A classification unit is used to classify the various error parameters, determine the influence relationship of different error parameters on the measurement accuracy of the sensor, and then construct an error model;

[0016] The construction unit is used to construct the interactive coupling form between error parameters based on the error model and the on-orbit operation of the deep space optical sensor, determine the influence coefficient of the error parameters under different conditions, and construct the coupling analytical model between different error parameters by combining the fuzzy comprehensive evaluation method.

[0017] The simplified element is used to optimize the error model using adaptive orthogonal neighborhood, eliminate redundant features in the measurement error feature space, realize the fusion and reduction of error features, and construct an error reduction model;

[0018] The measurement unit is used to perform on-orbit measurement of the optical sensor using the error reduction model.

[0019] The above technical solution has the following beneficial effects:

[0020] This invention addresses the challenges of complex deep space exploration environments and multiple sources of external interference. It systematically analyzes the mechanism of on-orbit measurement errors of optical sensors, obtaining a representation of the correlation and coupling between error influencing factors and system state variables. Balancing high-precision error modeling and efficient on-orbit identification, it employs a feature extraction and dimensionality reduction method for optical sensor measurement errors based on differential manifolds and sparse constraints. This optimizes the feature space of optical measurement errors. Furthermore, combined with a navigation system observability evaluation system, based on qualitative assessment of observability, it constructs a comprehensive error reduction model that integrates multiple source errors, providing technical support for efficient system error identification and complete navigation state estimation. Attached Figure Description

[0021] Figure 1 This is a flowchart of an analysis method for on-orbit measurement error of an optical sensor provided in an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of the structure of an analysis device for on-orbit measurement error of an optical sensor provided in an embodiment of the present invention;

[0023] Figure 3 This is a flowchart of the first embodiment of an analysis method for on-orbit measurement error of an optical sensor provided by an embodiment of the present invention;

[0024] Figure 4 This is a flowchart of a second embodiment of an on-orbit measurement error analysis method for an optical sensor provided by an embodiment of the present invention;

[0025] Figure 5 This is a flowchart of the third embodiment of the method for analyzing on-orbit measurement errors of an optical sensor provided by the present invention; Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] The accuracy of on-orbit measurements of optical sensors is affected by numerous error factors, with complex and interdependent relationships. On the one hand, it is necessary to systematically and comprehensively analyze the generation mechanisms of various internal and external errors throughout the entire on-orbit measurement process of optical sensors, and to analyze the impact of each error factor on the accuracy of optical sensor measurement data. On the other hand, in order to adapt to the constraints of limited on-board resources, it is also necessary to reduce the dimensionality of various errors and error models by considering the correlation and coupling of errors.

[0028] In embodiments of the present invention, such as Figure 1 This paper provides a method for analyzing the on-orbit measurement error of an optical sensor, the method comprising the following steps:

[0029] S101: Acquire and analyze various error parameters that affect the measurement accuracy of the sensor, forming a measurement error characteristic space; for the entire process of deep space probe operation in orbit, combined with optical measurement principles, analyze the hierarchy and progressive relationship of optical measurement error influencing factors, systematically and comprehensively analyze various error sources, error influence characteristics, and error influence magnitude, establish the design space of optical sensor measurement error parameters, and lay the foundation for the construction of a comprehensive error model.

[0030] Specifically, this includes: analyzing the influence weights of various error parameters on the accuracy, update frequency, and amplitude of measurement data; using spectrum analysis techniques to trace the root causes of errors; and establishing quantitative influence functions and transfer matrices between error parameters. Further, it involves: analyzing the impact of different error factors on the output accuracy, output frequency, and output magnitude of measurement data; using spectrum analysis to trace the source of measurement errors; and constructing influence functions and error propagation laws between different types of errors based on a multi-source error hierarchy. Finally, it involves determining whether error factors are controllable based on different mission phases of deep space probes, and classifying errors according to their impact and controllability.

[0031] For example, regarding the impact of internal errors on image point extraction, the imaging position deviation when acquiring measurement data will reduce the measurement accuracy of the optical sensor. In order to analyze the influencing factors of image point extraction, it is necessary to consider the different types of error patterns presented when acquiring different types of measurement data such as point targets (such as star points) and area targets (such as large planets in the transfer phase and geographic landmarks in the landing / descent phase), establish corresponding error impact models, study the error impact characteristics, and analyze the error types they belong to (such as random errors, compensable systematic errors, or uncertain errors).

[0032] For error parameters with well-defined mechanisms, an analytical model is constructed using physical mechanism functions. Statistical quantification features are extracted from complex error parameters. Combined with ground simulation and on-orbit measured data, parameter modeling is completed through time series analysis and empirical mode decomposition to characterize the error parameters. Based on differential manifold theory, a mapping relationship between the error parameter space and the local homeomorphic Euclidean space is established, and a bidirectional mapping is established. The quantification features and dynamic attributes of various error parameters are integrated into the spatial construction to form a measurement error feature space, realizing a one-to-one correspondence between error parameters and the on-orbit motion state of the detector.

[0033] S102: Classify the various error parameters, determine the influence relationship between different error parameters and the sensor measurement accuracy, and then construct an error model, such as... Figure 3 As shown, it includes:

[0034] The various error parameters are classified according to whether they are controllable or uncontrollable, and qualitative or quantitative methods.

[0035] This study analyzes the influence of different error factors on the measurement accuracy of the sensor. Combining the generation mechanism and performance characteristics of various error parameters, it adopts the uncertainty error modeling method to establish parametric, semi-parametric and non-parametric models of various error parameter influencing factors.

[0036] Based on the analysis of the error hierarchy and the transmission laws between errors, and combined with the physical mechanisms of optical measurement, error modeling is achieved through mathematical derivation, physical deduction, and experimental verification. For errors whose physical influence mechanisms are clearly defined through optics, temperature, etc., the error mechanism function can be directly used as the error model. For errors with complex influence relationships that are difficult to directly establish with measurement accuracy, on the one hand, semi-parametric / non-parametric estimation methods can be used, combined with kernel estimation, local polynomial estimation, k-nearest neighbor estimation, etc., to determine the periodic and statistical characteristics of the error influence. On the other hand, simulation data from ground-based experimental verification systems and on-orbit measured data can be combined to construct error influence models through parameter estimation methods such as time series analysis and empirical mode decomposition.

[0037] S103: Based on the aforementioned error model and considering the on-orbit operation of the deep space optical sensor, construct the interactive coupling form between error parameters, determine the influence coefficients of error parameters under different conditions, and construct a coupling analytical model between different error parameters using the fuzzy comprehensive evaluation method, such as... Figure 4 As shown:

[0038] First, considering the on-orbit operation of deep space probes, based on the multi-source error influence mechanism, and combining the analytical functions of various error influences and error propagation relationships, the coupling / correlation relationships among the measurement error parameters of optical sensors are studied. For example, the optical sensor distortion caused by sunlight has a strong coupling with the optical sensor installation error and the measurement reference error, which can be uniformly expressed using multiplicative error form or constant transformation of the installation matrix. The single-event effect of high-energy particles in space, the random errors caused by stellar jitter, and the thermal noise of the optical sensor have strong similarities, and a unified error expression model can be constructed. This unified expression model helps to simplify the error parameter space, thereby improving the efficiency of parameter estimation and system state estimation.

[0039] Secondly, based on the coupling / correlation analysis between measurement error parameters and combined with the on-orbit operation of the deep space probe, the interactive coupling form between errors is constructed, the error influence coefficient under different conditions is determined, and the similarity coefficient calculation method is studied by combining the fuzzy comprehensive evaluation method to construct an analytical expression model of the coupling / correlation between different error influences. For example, when the probe is not exposed to sunlight, the influence of optical sensor measurements decreases, and it is necessary to adjust the coupling form between optical distortion caused by the thermal vacuum environment and installation errors and measurement reference errors; during the landing / descent phase, the influence of celestial body jitter caused by soft landing is greater, and the resulting error form and characteristics will undergo fundamental changes, requiring a reduction in the similarity coefficient with the thermal noise influence of the optical sensor.

[0040] Finally, combining the optical measurement error influence model and the coupling / correlation model between multiple source errors, the design parameters in the error influence model are clarified on the basis of the error parameter space. An analytical relationship is constructed between the optical sensor measurement error and the design parameters such as the dynamic characteristics of rotating parts, on-orbit temperature change, and high-energy particle density change. The range of design parameter variation is given. Through a combination of experimental verification and theoretical derivation, a comprehensive error model of the optical sensor is constructed.

[0041] S104: Optimize the error model using adaptive orthogonal neighborhood, eliminate redundant features in the measurement error feature space, achieve the fusion and reduction of error features, and construct an error reduction model;

[0042] Specifically, it includes:

[0043] Based on the aforementioned error model, and combining the manifold learning strategy with the multi-criteria fusion feature selection method, a threshold for the impact of accuracy is set, and sensitive feature spaces that meet the threshold requirements and play a key role in measurement accuracy are selected. Redundant error features are eliminated simultaneously, thus achieving preliminary simplification of feature dimensions.

[0044] Multiple sensitive feature spaces selected are fused. The core information of the original error parameter space is retained by the feature fusion algorithm to ensure that the fused feature space can fully characterize the key attributes of the original error and establish an error reduction model.

[0045] By combining the observability determination system of the autonomous navigation system, and balancing the feature dimension and onboard resource constraints, if the feature dimension is too high after fusion, resulting in computational overload, it is necessary to further reduce the dimension through manifold learning technology; if the feature dimension is insufficient, resulting in the accuracy not meeting the navigation index, then sensitive features are added within the range of onboard resources to determine the optimal reduction result that balances error representation accuracy and engineering operation efficiency.

[0046] As one implementation method: Figure 5 As shown, firstly, under the influence of the complex on-orbit environment of deep space probes, the measurement errors of optical sensors exhibit nonlinear, strongly coupled, and difficult-to-identify characteristics. Combining a comprehensive error model based on error coupling / correlation analysis, we explore the Euclidean space that is locally homeomorphic to the error parameter space and establish the corresponding local mapping relationship. For the internal and external errors of optical sensor measurements under the influence of complex environments, we construct a complete error feature space that can characterize the parameter space, and fully describe the mapping relationship between the optical sensor measurement error parameters and the on-orbit motion state of the deep space probe.

[0047] Secondly, based on the error feature space, combined with the manifold learning strategy, the multi-criteria fusion feature selection method is used to screen error features. According to the impact of different errors on the accuracy of measurement data, sensitive feature spaces that meet the threshold or feature spaces that have a great impact on measurement accuracy are selected, redundant error feature spaces are eliminated, and the selected feature spaces are fused to comprehensively characterize the original parameter space, realize the fusion and reduction of the optical sensor measurement error feature space, and establish a multi-source error integrated characterization model.

[0048] Finally, combining the qualitative judgment system of the observability of the autonomous navigation system, and fully considering the impact of the error parameter selection results on computational resource consumption and state estimation accuracy, based on a comprehensive evaluation of the observability of the autonomous navigation system, the error feature space selection results are further determined to obtain a comprehensive error reduction model that balances high accuracy and high efficiency. For example, when too many error parameters are selected, and computational resources are difficult to maintain (i.e., poor observability), it is necessary to use the manifold learning strategy again to reduce the error selection dimension; when too few error parameters are selected, and it is difficult to meet the autonomous navigation accuracy index (i.e., not meeting the observability conditions), it is necessary to add a relatively sensitive error feature space within the tolerance range of computational processing resources.

[0049] S105: On-orbit measurement of the optical sensor is achieved using the error reduction model.

[0050] In embodiments of the present invention, such as Figure 2 Furthermore, an analysis device for on-orbit measurement errors of optical sensors is provided, comprising:

[0051] The acquisition unit 21 is used to acquire and analyze various error parameters that affect the measurement accuracy of the sensor, and form a measurement error feature space;

[0052] The classification unit 22 is used to classify the various error parameters, determine the influence relationship of different error parameters on the measurement accuracy of the sensor, and then construct an error model;

[0053] Construction unit 23 is used to construct the interactive coupling form between error parameters based on the error model and the on-orbit operation of the deep space optical sensor, determine the influence coefficient of error parameters under different conditions, and construct the coupling analytical model between different error parameters by combining the fuzzy comprehensive evaluation method.

[0054] The simplified element 24 is used to optimize the error model using adaptive orthogonal neighborhood, eliminate redundant features in the measurement error feature space, realize the fusion and reduction of error features, and construct an error reduction model.

[0055] Measurement unit 25 is used to perform on-orbit measurement of the optical sensor using the error reduction model.

[0056] The acquisition unit 21 includes:

[0057] This tool is used to analyze the weight of the influence of various error parameters on the accuracy, update frequency and amplitude of measurement data. It uses spectrum analysis technology to trace the root cause of error and establishes quantitative influence functions and transfer matrices between error parameters.

[0058] For error parameters with well-defined mechanisms, analytical models are constructed using physical mechanism functions.

[0059] It is used to extract statistical quantification features from complex error parameters, and combines ground simulation and on-orbit measured data to complete parameter modeling through time series analysis and empirical mode decomposition, thereby realizing the characterization of error parameters;

[0060] Based on the theory of differential manifolds, this method establishes a mapping relationship between the error parameter space and the local homeomorphic Euclidean space, and establishes a two-way mapping. It integrates the quantitative characteristics and dynamic properties of various error parameters into the spatial construction to form a measurement error feature space, thereby realizing a one-to-one correspondence between error parameters and the on-orbit motion state of the detector.

[0061] The classification unit 22 includes:

[0062] This is used to classify the various error parameters according to controllable and uncontrollable, and qualitative and quantitative methods;

[0063] This method is used to analyze the influence of different error factors on the measurement accuracy of the sensor. Combining the generation mechanism and performance characteristics of various error parameters, it adopts the uncertainty error modeling method to establish parametric, semi-parametric and non-parametric models of various error parameter influencing factors.

[0064] The approximately simple element 24 includes:

[0065] Based on the error model, combined with manifold learning strategy and multi-criteria fusion feature selection method, a threshold for accuracy impact is set, and sensitive feature space that meets the threshold requirements and plays a key role in measurement accuracy is selected. Redundant error features are eliminated simultaneously to achieve preliminary simplification of feature dimensions.

[0066] This is used to fuse multiple sensitive feature spaces selected from the sample. The feature fusion algorithm retains the core information of the original error parameter space, ensuring that the fused feature space can fully characterize the key attributes of the original error and establish a multi-source error integrated representation model.

[0067] This is used to combine the observability determination system of autonomous navigation system, balance feature dimension and onboard resource constraints. If the feature dimension is too high after fusion, resulting in computational overload, it is necessary to further reduce the dimension through manifold learning technology. If the feature dimension is insufficient, resulting in the accuracy not meeting the navigation index, sensitive features are added within the range of onboard resources to determine the optimal reduction result that balances error representation accuracy and engineering operation efficiency.

[0068] An analysis device for on-orbit measurement error of an optical sensor employs the above-mentioned method for analyzing on-orbit measurement error of an optical sensor. Its principle and process are the same as those of the method for analyzing on-orbit measurement error of an optical sensor, and will not be repeated here.

[0069] Example:

[0070] This invention provides a method for analyzing on-orbit measurement errors of optical sensors. Specifically, analyzing the mechanism of on-orbit measurement errors of optical sensors is crucial for solving the problem of navigation information availability and is the foundation for achieving high-precision optical autonomous navigation. The accuracy of on-orbit measurement of optical sensors is affected by numerous error factors, with complex and coupled relationships, making it difficult to construct a measurement error model. Furthermore, limitations in onboard computing resources necessitate balancing the accuracy of measurement error characteristics and the efficiency of error processing during model construction. Therefore, it is essential to combine the principles of on-orbit measurement of optical sensors, the engineering background of the deep space exploration environment, onboard computing capabilities, and mathematical methods. Through error factor tracing, hierarchical analysis, and uncertain semi-parametric / non-parametric modeling, the error mechanism of high-precision on-orbit measurement of optical sensors is revealed. Finally, through coupling / correlation analysis and manifold learning strategies, a comprehensive error reduction model representing multi-source errors is established. Specifically, this includes:

[0071] 1) Hierarchical Analysis and Uncertainty Modeling of Error Factors in Optical Sensor Measurement Accuracy: Combining the entire on-orbit measurement process of the optical sensor, this study systematically analyzes various error factors affecting the sensor's measurement accuracy, categorizing them into internal error factors related to sensor design (such as imaging error, image point extraction error, photoelectric sampling noise, and data transmission) and external error factors related to the deep space exploration environment (such as installation error, vibration of rotating components, uneven solar illumination, high-energy particle radiation in space, and accuracy of target celestial ephemeris). Using the error source tracing method, the hierarchy of accuracy-influencing factors is determined, and they are classified according to controllable / uncontrollable, qualitative / quantitative systems and methods. The study investigates the influence relationships of different error factors on the sensor's measurement accuracy. Based on the generation mechanism and characteristics of each error, parametric / semi-parametric / non-parametric uncertainty error modeling methods are employed to establish parametric / semi-parametric / non-parametric models for each error-influencing factor.

[0072] 2) Coupling / Correlation Analysis of Error Models: The system analyzes the progression, coupling, and correlation of various internal and external error factors throughout the measurement process of the optical sensor. Based on the uncertainty modeling of various errors, and combined with the generation mechanism of error factors and the manifestation of error models, it explores the similarity and interactive coupling relationship between optical measurement error models / related parameters. For example, the relationship between the established low-frequency error parameter / semi-parameter model and the temperature field distribution caused by the deep space probe orbit and uneven solar illumination, the relationship between the vibration of rotating parts and image point extraction and imaging errors, and the relationship between reference error, installation error and probe platform deformation, etc.

[0073] 3) Construction of a comprehensive error reduction model for integrated representation of multi-source errors: Based on the analysis of the similarity and coupling of various error models and parameters, the error influencing factors are input into the adaptive orthogonal neighborhood-preserving embedding manifold learning through methods such as spatial transformation / spatial projection. This further optimizes the feature space of the optical sensor measurement error, eliminates redundant features in the measurement error feature space, and achieves the fusion and reduction of error features. Combined with the navigation system observability evaluation system, based on the qualitative judgment of observability, the reduced mixed error is re-modeled, thereby obtaining a comprehensive error reduction model that takes into account both the accuracy of error features and the efficiency of error processing. This lays the foundation for the design of subsequent algorithms such as system error identification and state / parameter complete estimation.

[0074] This invention addresses the challenges of complex deep space exploration environments and multiple sources of external interference. It systematically analyzes the mechanism of on-orbit measurement errors of optical sensors, obtaining a representation of the correlation and coupling between error influencing factors and system state variables. Balancing high-precision error modeling and efficient on-orbit identification, it proposes a method for feature extraction and dimensionality reduction of optical sensor measurement errors based on differential manifolds and sparse constraints. This optimizes the feature space of optical measurement errors and, combined with the observability evaluation system of navigation systems, constructs a comprehensive error reduction model that integrates the representation of multi-source errors based on qualitative judgment of observability. This provides technical support for efficient identification of system errors and complete estimation of navigation state.

[0075] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process may be rearranged without departing from the scope of this disclosure. The appended method claims provide elements of various steps in an exemplary order and are not intended to limit the scope to the specific order or hierarchy described.

[0076] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for analyzing on-orbit measurement errors of an optical sensor, characterized in that, The method includes the following steps: Acquire and analyze various error parameters that affect the measurement accuracy of the sensor to form a measurement error feature space; The various error parameters are classified, and the influence of different error parameters on the measurement accuracy of the sensor is determined. Then, an error model is constructed. Based on the aforementioned error model and combined with the on-orbit operation of the deep space optical sensor, an interactive coupling form between error parameters is constructed, the influence coefficients of error parameters under different conditions are determined, and a coupling analytical model between different error parameters is constructed by combining the fuzzy comprehensive evaluation method. The error model is optimized by adaptive orthogonal neighborhood, and redundant features in the measurement error feature space are eliminated to achieve the fusion and reduction of error features and construct an error reduction model. The aforementioned error reduction model is used to achieve on-orbit measurement of the optical sensor.

2. The method for analyzing on-orbit measurement errors of an optical sensor according to claim 1, characterized in that, The acquisition and analysis of various error parameters affecting the measurement accuracy of the sensor to form a measurement error feature space includes: The influence weights of various error parameters on the accuracy, update frequency and amplitude of measurement data are analyzed. Spectrum analysis technology is used to trace the root causes of errors and to establish quantitative influence functions and transfer matrices among error parameters. For error parameters with well-defined mechanisms, analytical models are constructed using physical mechanism functions; Statistical quantification features are extracted from complex error parameters. Combined with ground simulation and on-orbit measured data, parameter modeling is completed through time series analysis and empirical mode decomposition to achieve the characterization of error parameters. Based on the theory of differential manifolds, a mapping relationship between the error parameter space and the local homeomorphic Euclidean space is established, and a two-way mapping is established. The quantitative characteristics and dynamic properties of various error parameters are integrated into the spatial construction to form a measurement error feature space, realizing a one-to-one correspondence between error parameters and the on-orbit motion state of the detector.

3. The method for analyzing on-orbit measurement errors of an optical sensor according to claim 1, characterized in that, The process involves classifying the various error parameters, determining the influence of different error parameters on the sensor's measurement accuracy, and then constructing an error model, including: The various error parameters are classified according to whether they are controllable or uncontrollable, and qualitative or quantitative methods. This study analyzes the influence of different error factors on the measurement accuracy of the sensor. Combining the generation mechanism and performance characteristics of various error parameters, it adopts the uncertainty error modeling method to establish parametric, semi-parametric and non-parametric models of various error parameter influencing factors.

4. The method for analyzing on-orbit measurement errors of an optical sensor according to claim 3, characterized in that, The step of optimizing the error model using adaptive orthogonal neighborhoods to eliminate redundant features in the measurement error feature space, achieving fusion and reduction of error features, and constructing an error reduction model includes: Based on the aforementioned error model, and combining the manifold learning strategy with the multi-criteria fusion feature selection method, a threshold for the impact of accuracy is set, and sensitive feature spaces that meet the threshold requirements and play a key role in measurement accuracy are selected. Redundant error features are eliminated simultaneously, thus achieving preliminary simplification of feature dimensions. Multiple sensitive feature spaces selected are fused. The core information of the original error parameter space is retained by the feature fusion algorithm to ensure that the fused feature space can fully characterize the key attributes of the original error and establish an error reduction model. By combining the observability determination system of the autonomous navigation system, and balancing the feature dimension and onboard resource constraints, if the feature dimension is too high after fusion, resulting in computational overload, it is necessary to further reduce the dimension through manifold learning technology; if the feature dimension is insufficient, resulting in the accuracy not meeting the navigation index, then sensitive features are added within the range of onboard resources to determine the optimal reduction result that balances error representation accuracy and engineering operation efficiency.

5. An analysis device for on-orbit measurement error of an optical sensor, characterized in that, include: The acquisition unit is used to acquire and analyze various error parameters that affect the measurement accuracy of the sensor, forming a measurement error feature space; A classification unit is used to classify the various error parameters, determine the influence relationship of different error parameters on the measurement accuracy of the sensor, and then construct an error model; The construction unit is used to construct the interactive coupling form between error parameters based on the error model and the on-orbit operation of the deep space optical sensor, determine the influence coefficient of the error parameters under different conditions, and construct the coupling analytical model between different error parameters by combining the fuzzy comprehensive evaluation method. The simplified element is used to optimize the error model using adaptive orthogonal neighborhood, eliminate redundant features in the measurement error feature space, realize the fusion and reduction of error features, and construct an error reduction model; The measurement unit is used to perform on-orbit measurement of the optical sensor using the error reduction model.

6. The analysis device for on-orbit measurement error of an optical sensor according to claim 5, characterized in that, The acquisition unit includes: This tool is used to analyze the weight of the influence of various error parameters on the accuracy, update frequency and amplitude of measurement data. It uses spectrum analysis technology to trace the root cause of error and establishes quantitative influence functions and transfer matrices between error parameters. For error parameters with well-defined mechanisms, analytical models are constructed using physical mechanism functions. It is used to extract statistical quantification features from complex error parameters, and combines ground simulation and on-orbit measured data to complete parameter modeling through time series analysis and empirical mode decomposition, thereby realizing the characterization of error parameters; Based on the theory of differential manifolds, this method establishes a mapping relationship between the error parameter space and the local homeomorphic Euclidean space, and establishes a two-way mapping. It integrates the quantitative characteristics and dynamic properties of various error parameters into the spatial construction to form a measurement error feature space, thereby realizing a one-to-one correspondence between error parameters and the on-orbit motion state of the detector.

7. The analysis device for on-orbit measurement error of an optical sensor according to claim 5, characterized in that, The classification unit includes: This is used to classify the various error parameters according to controllable and uncontrollable, and qualitative and quantitative methods; This method is used to analyze the influence of different error factors on the measurement accuracy of the sensor. Combining the generation mechanism and performance characteristics of various error parameters, it adopts the uncertainty error modeling method to establish parametric, semi-parametric and non-parametric models of various error parameter influencing factors.

8. The analysis device for on-orbit measurement error of an optical sensor according to claim 7, characterized in that, The simple elements include: Based on the error model, combined with manifold learning strategy and multi-criteria fusion feature selection method, a threshold for accuracy impact is set, and sensitive feature space that meets the threshold requirements and plays a key role in measurement accuracy is selected. Redundant error features are eliminated simultaneously to achieve preliminary simplification of feature dimensions. This is used to fuse multiple selected sensitive feature spaces. The feature fusion algorithm retains the core information of the original error parameter space, ensuring that the fused feature space can fully characterize the key attributes of the original error and establish an error reduction model. This is used to combine the observability determination system of autonomous navigation system, balance feature dimension and onboard resource constraints. If the feature dimension is too high after fusion, resulting in computational overload, it is necessary to further reduce the dimension through manifold learning technology. If the feature dimension is insufficient, resulting in the accuracy not meeting the navigation index, sensitive features are added within the range of onboard resources to determine the optimal reduction result that balances error representation accuracy and engineering operation efficiency.

Images