Data processing method of illumination rectangular combined lens for deep space exploration

Abstract

The invention relates to a data processing method of an illumination rectangular combined lens for deep space exploration, which comprises the following steps of: acquiring multispectral imaging data of two-pole regions of the moon, and establishing a lunar soil reflection characteristic database; obtaining reflection characteristic parameters corresponding to different lunar soil types when ice is dug at the two poles of the moon from a lunar soil reflection characteristic database; in the ice digging operation process of the lunar landing vehicle, lunar soil reflectivity data of an operation area are obtained in real time; predicting to obtain target illumination spectrum configuration parameters under different lunar soil conditions; energy consumption and ice digging efficiency of the lunar landing vehicle in the ice digging operation process are obtained; the installation angle and the irradiation direction of the lunar landing vehicle illumination system are adjusted for the first time; and carrying out temperature control on the combined lens assembly in the optimized lunar landing vehicle lighting system. According to the method, by accurately matching the illumination effect with the lunar soil characteristics and considering the energy consumption and efficiency factors, the efficiency and precision of ice digging operation in the two-pole region of the moon are remarkably improved, and important technical support is provided for lunar resource development.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a data processing method for an illuminating rectangular combination lens for deep space exploration. Background Art

[0002] In deep space exploration, the lunar rover needs to dig ice at the lunar poles. Due to the complex and changeable environment on the lunar surface, the reflective properties of the lunar soil also change accordingly. In order to ensure the smooth progress of the ice digging operation, the lunar rover needs to adjust the lighting system in real time according to the reflective properties of the lunar soil. However, the reflective properties of the lunar soil are affected by many factors, such as particle size, mineral composition, and light angle. These factors are coupled with each other, making it difficult for the lunar rover to accurately judge the reflective properties of the lunar soil. In addition, the lighting rectangular combination lens carried by the lunar rover needs to collect lunar soil reflection data in multiple bands, including visible light, near infrared, short-wave infrared, etc. Data from different bands need to be fused to obtain comprehensive information on the reflective properties of the lunar soil. In addition, there are differences in data format, resolution, signal-to-noise ratio, etc. in different bands. How to perform efficient and high-quality data fusion is a technical problem that needs to be solved urgently. At the same time, the extreme environment at the lunar poles places strict requirements on the lighting system of the lunar rover. The temperature on the lunar surface can change from extremely high to extremely low in a very short time, and the temperature difference between day and night is extremely large, which poses a challenge to the stability and reliability of the lighting system. How to ensure the normal operation of the lighting system in harsh environments is also a technical problem that needs to be overcome.

[0003] In summary, the main technical problems faced by the lunar rover when digging ice at the lunar poles include: the complex and variable reflectivity of the lunar soil, the efficient fusion of multi-band data, and the stability and reliability of the lighting system in harsh environments. Summary of the invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a data processing method for a rectangular combination lens for deep space exploration. The data processing method for a rectangular combination lens for deep space exploration significantly improves the efficiency and accuracy of ice excavation operations in the lunar polar regions by accurately matching the lighting effect with the characteristics of the lunar soil and taking into account energy consumption and efficiency factors, providing important technical support for the development of lunar resources.

[0005] The data processing method of the illumination rectangular combination lens for deep space exploration described in the present invention comprises the following steps: S1. Obtain multispectral imaging data of the lunar polar regions, perform preprocessing and calculate phase angles to obtain lunar soil reflectivity and bidirectional reflectivity characteristics in different bands. By comparing and analyzing the spectral data of samples with preset particle sizes, establish the corresponding relationship between reflectivity and lunar soil particle size, and establish a lunar soil reflectivity characteristics database; S2. Obtaining reflection characteristic parameters corresponding to different lunar soil types when digging ice at the lunar poles from the lunar soil reflection characteristic database, and obtaining a corresponding spectrum configuration scheme for the illumination rectangular combination lens in the lunar rover illumination system; S3. During the lunar rover's ice excavation operation, the lunar soil reflectivity data of the operation area is obtained in real time, and compared with the data in the lunar soil reflectivity characteristic database, and the spectrum configuration scheme of the lunar rover's lighting system is dynamically adjusted according to the type and particle size distribution characteristics of the lunar soil in the current operation area; S4. Based on the lunar soil reflectivity data collected during the lunar rover's ice excavation operation, the target lighting spectrum configuration parameters under different lunar soil conditions are predicted and applied to the dynamic adjustment process of the lunar rover's lighting system; S5, obtaining the energy consumption and ice-digging efficiency of the lunar rover during the ice-digging operation, and performing correlation analysis with the target illumination spectrum configuration parameters, selecting a target wavelength point within a preset visible light range, calculating the light intensity configuration of each wavelength, and obtaining an updated spectrum configuration scheme; S6. According to the terrain and lighting conditions of the lunar polar regions and the updated spectrum configuration scheme, the installation angle and illumination direction of the lunar rover lighting system are adjusted once, and according to the bidirectional reflection data in the lunar soil reflection characteristic database, the installation angle of the lunar rover lighting system is adjusted twice; S7. Temperature control is performed on the combined lens assembly in the optimized lunar rover lighting system. The temperature status of the combined lens is monitored in real time through a preset thermal management system, and the combined lens whose temperature exceeds a preset temperature threshold range is cooled or heated.

[0006] Preferably, the step S1 specifically includes: Using a satellite-borne multispectral detector to obtain lunar surface reflectance data from the lunar polar regions, performing radiometric calibration on the lunar surface reflectance data according to the spectral response curve of the multispectral detector to obtain lunar surface reflectance data; For the lunar surface reflectivity data, the solar incidence angle and the detection angle are calculated according to the attitude parameters of the lunar orbit satellite platform, and the lunar surface bidirectional reflection calculation formula is used to perform angle normalization to obtain the reflectivity data under standard geometric conditions; Performing band combination operation on the reflectance data under the standard geometric conditions, and extracting spectral characteristic parameters by least square fitting, wherein the spectral characteristic parameters include spectral slope, absorption band depth and band ratio; A mapping relationship between the spectral characteristic parameters and the lunar surface particle size is established, a lunar surface particle size inversion model is constructed using a support vector regression method, and particle size mapping of the lunar polar regions is performed using the lunar surface particle size inversion model.

[0007] Preferably, the step S2 specifically includes: The spectrum configuration scheme includes the light intensity distribution ratio of each band and the lighting angle setting parameters; Performing spectral scanning on the lunar soil sample to obtain the spectral reflectance curve and absorption peak wavelength of the lunar soil sample in the visible light band, near infrared band and short-wave infrared band; Calculate the band transmittance of the combined lens using a ray tracing program according to the spectral reflectance curve to obtain spectral transmittance data of the combined lens in different bands; Using a Bayesian optimization algorithm to process the spectral reflectance data and the spectral transmittance data to obtain a spectral intensity distribution scheme; According to the spectral transmittance data and the spectral intensity distribution scheme, a Monte Carlo ray tracing method is used to calculate the light intensity distribution diagram of the illumination beam on the lunar soil surface to obtain a light intensity distribution uniformity index; Generate the optimal lighting angle parameters based on the light intensity distribution uniformity index.

[0008] Preferably, the step S3 specifically includes: Obtain lunar soil surface reflectance data according to the band calibration curve, and perform spectral noise reduction and normalization processing on the lunar soil surface reflectance data to obtain spectral characteristic parameters including spectral slope, spectral depth and band ratio; The similarity between the spectral characteristic parameters and the standard data in the pre-established lunar soil reflectance characteristic database is calculated, and the lunar soil type identification result and the mineral content value are obtained by using the k-nearest neighbor algorithm; According to the lunar soil type identification results and the mineral content values, a recurrent neural network is used to obtain the power adjustment coefficient of each band of light sources, and the transmitted light intensity of each band of light sources is modulated by a liquid crystal spectral modulation array.

[0009] Preferably, the step S4 specifically includes: Acquire lunar soil reflectivity data collected by a lunar rover multispectral camera, and perform dark current correction and spectral calibration on the lunar soil reflectivity data to obtain a lunar soil reflectivity spectrum data set; The lunar soil reflectance spectrum data set is standardized by using a maximum and minimum value normalization method, and spectral characteristic parameters of spectral slope, absorption band depth and band ratio are extracted based on the standardized data; Constructing a deep neural network model according to the spectral feature parameters, mapping the spectral feature parameters to a latent feature space through an encoding layer of the deep neural network, and converting the latent features into spectral configuration parameters through a decoding layer of the deep neural network; A lighting component control instruction sequence is generated according to the spectrum configuration parameters, the control instruction sequence including a band number, a power coefficient and a modulation voltage parameter, and the control instruction sequence is sent to a lighting controller via a serial communication interface.

[0010] Preferably, the step S5 specifically includes: According to the data collected by the current and voltage monitoring unit, the real-time power values ​​of the lighting components and the mining components are calculated, and the energy consumption data of the lighting components and the mining components are obtained through time integration operation; Using a depth sensor to record excavation operation parameters to obtain excavation depth data, integrating the energy consumption data and the excavation depth data to obtain an energy consumption data set per unit volume of ice layer; Constructing a mapping function between spectrum configuration parameters and unit energy consumption according to the energy consumption data set, and determining a weight coefficient of the mapping function by a least square method; A set of wavelength sampling points is obtained from the lunar soil reflectivity curve, and the initial value of light intensity is iteratively optimized using the gradient descent method. If the deviation between the predicted energy consumption and the measured value exceeds the deviation threshold, the light intensity parameter is updated.

[0011] Preferably, the step S6 specifically includes: The lunar surface point cloud data is obtained by using the LiDAR scanner, and the lunar surface elevation distribution map is obtained by projecting it onto the horizontal plane through coordinate transformation; Performing least squares surface fitting on the lunar surface elevation distribution map to obtain a terrain feature parameter set including slope value, aspect angle and curvature value; A multispectral camera is used to monitor the light intensity in the working area, and the light shielding area is calculated by combining the terrain feature parameter set and the azimuth and altitude angle output by the sun position calculator; According to the light shielding area and the terrain feature parameter set, a local search algorithm is used to optimize the posture angle of the lighting component to obtain the optimal installation angle of the lighting component.

[0012] Preferably, the step S7 specifically includes: The temperature sensor arrays arranged in the central area and the edge area of ​​the lens collect the temperature values ​​of the temperature measuring points, and the temperature values ​​of the temperature measuring points are obtained by synchronous sampling by the data acquisition module; Performing neural network spatial interpolation calculation according to the temperature values ​​of the temperature measurement points, and the output layer of the neural network generates lens surface temperature distribution data; A temperature field uniformity index value is calculated based on the lens surface temperature distribution data, and the temperature field uniformity index value is used to determine the area with drastic temperature changes through temperature gradient calculation; A fuzzy rule base is queried according to the temperature field uniformity index value of the area with drastic temperature changes, and the fuzzy rule base outputs the power value of the semiconductor refrigerator and the power value of the heating wire to the temperature closed-loop controller.

[0013] The data processing method of the illumination rectangular combination lens for deep space exploration described in the present invention has the following advantages: The data processing method of the rectangular combined lens for deep space exploration described in the present invention obtains multi-spectral imaging data of the lunar polar regions, establishes a lunar soil reflectance characteristic database, and designs a spectral configuration scheme of the lunar rover lighting system based on the database; during the ice excavation operation, the lunar soil reflectance data of the operation area is obtained in real time, and the lighting system configuration is dynamically adjusted to match the lunar soil reflectance characteristics; at the same time, based on the energy consumption and ice excavation efficiency data, the spectral configuration parameters are optimized, and the lighting system installation angle is adjusted according to the terrain and lighting conditions; and the combined lens assembly in the lunar rover lighting system is temperature controlled to ensure the stability and reliability of the lunar rover lighting system. The present invention significantly improves the efficiency and accuracy of ice excavation operations in the lunar polar regions by accurately matching the lighting effect with the lunar soil characteristics and taking into account the energy consumption and efficiency factors, providing important technical support for the development of lunar resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flow chart of a data processing method of a rectangular combination lens for deep space exploration according to the present invention. DETAILED DESCRIPTION

[0015] like Figure 1 As shown, the data processing method of the deep space exploration illumination rectangular combination lens of the present invention comprises the following steps: S1. Obtain multispectral imaging data of the lunar polar regions, perform preprocessing and calculate phase angles to obtain lunar soil reflectivity and bidirectional reflection characteristics in different bands. Establish the corresponding relationship between reflectivity and lunar soil particle size by comparing and analyzing the spectral data of preset particle size samples, and establish a lunar soil reflection characteristic database. The lunar soil reflectivity in different bands is obtained by performing radiation correction and atmospheric correction preprocessing on the acquired multispectral imaging data, and the bidirectional reflection characteristics of the lunar soil are obtained by calculating the phase angle according to the observation angle and the illumination angle. S2. Obtain the reflection characteristic parameters corresponding to different lunar soil types when digging ice at the lunar poles from the lunar soil reflection characteristic database, and obtain the corresponding spectrum configuration scheme of the illumination rectangular combination lens in the lunar rover illumination system; S3. During the lunar rover's ice excavation operation, the lunar soil reflectivity data of the operating area is obtained in real time, and compared with the data in the lunar soil reflectivity characteristic database. According to the type and particle size distribution characteristics of the lunar soil in the current operating area, the spectrum configuration scheme of the lunar rover's lighting system is dynamically adjusted; S4. Based on the lunar soil reflectivity data collected during the lunar rover's ice excavation operation, the target lighting spectrum configuration parameters under different lunar soil conditions are predicted and applied to the dynamic adjustment process of the lunar rover's lighting system; S5. Obtain the energy consumption and ice-digging efficiency of the lunar rover during the ice-digging operation, and perform correlation analysis with the target lighting spectrum configuration parameters, select the target wavelength point within the preset visible light range, calculate the light intensity configuration of each wavelength, and obtain an updated spectrum configuration scheme; S6. Adjust the installation angle and illumination direction of the lunar rover lighting system based on the terrain and illumination conditions of the lunar polar regions and the updated spectrum configuration scheme, and adjust the installation angle of the lunar rover lighting system based on the bidirectional reflection data in the lunar soil reflection characteristic database; S7. Temperature control is performed on the combined lens assembly in the optimized lunar rover lighting system. The temperature status of the combined lens is monitored in real time through a preset thermal management system, and the combined lens whose temperature exceeds a preset temperature threshold range is cooled or heated.

[0016] Furthermore, in this embodiment, step S1 specifically includes: The lunar surface reflectance data is obtained from the lunar polar regions using a satellite-borne multispectral detector. The lunar surface reflectance data is radiometrically calibrated according to the spectral response curve of the multispectral detector to obtain the lunar surface reflectance data. For the lunar surface reflectivity data, the solar incidence angle and detection angle are calculated according to the attitude parameters of the lunar orbit satellite platform, and the lunar surface bidirectional reflection calculation formula is used to normalize the angle to obtain the reflectivity data under standard geometric conditions; The reflectance data under standard geometric conditions are subjected to band combination operation, and the spectral characteristic parameters are extracted by least square fitting. The spectral characteristic parameters include spectral slope, absorption band depth and band ratio. The mapping relationship between the spectral characteristic parameters and the lunar surface particle size is established, and the lunar surface particle size inversion model is constructed using the support vector regression method. The particle size inversion model is used to map the lunar polar regions. Here is an example: The satellite-borne multispectral detector is used to obtain lunar surface reflectance data including visible light band, near infrared band and short-wave infrared band from the lunar polar regions. The raw data is radiometrically calibrated according to the spectral response curve of the multispectral detector, and the atmospheric attenuation correction is performed on the data of each band using the Lambert-Beer atmospheric transmission equation to obtain the lunar surface reflectance data. According to the lunar orbit satellite platform attitude parameters and lunar orbit position parameters, the solar incidence angle and the multi-spectral detector detection angle are calculated for each sampling point, and the lunar surface bidirectional reflection calculation formula is used to normalize the angle to obtain the reflectivity data under the standard geometric conditions of the lunar surface; By performing band combination calculation on the reflectance data under standard geometric conditions of the lunar surface, the lunar surface reflectance spectrum curve is obtained, and the spectrum curve is fitted by least squares to extract the spectrum characteristic parameters, which include spectrum slope, absorption band depth and band ratio; A spectrum sampler is used to measure the spectra of preset lunar surface samples of different particle sizes. The wavelength of the measured data is calibrated by the built-in calibration function of the spectrometer. The spectrum data is standardized by the area normalization method to obtain a lunar soil reflectance characteristic database. Based on the lunar soil reflectance database, the mapping relationship between spectral characteristic parameters and lunar surface particle size is established, and the lunar surface particle size inversion model is constructed using the support vector regression method. The model input is the spectral characteristic parameters, and the output is the estimated value of the lunar surface particle size. Particle size mapping was performed in the lunar polar regions based on the lunar surface particle size inversion model. The inversion results were cross-validated using a test set from the lunar surface sample standard spectral library. The inversion accuracy was evaluated using the root mean square error. The acquisition of lunar surface reflectance spectrum data involves multiple links. The lunar surface reflectance measurement is affected by many factors. In the process of obtaining lunar surface reflectance, the spectral response characteristics of the multi-spectral detector will affect the original data. The spectral response curve of the detector in different bands shows different characteristics. The spectral response coefficient of the detector is obtained through spectral calibration experiments, which corresponds to the visible light band of 380-780 nanometers, the near infrared band of 780-2500 nanometers, and the short-wave infrared band of 2500-8000 nanometers. Radiometric calibration is performed based on the spectral response coefficient to eliminate the impact of the detector response characteristics on the data. In the process of the atmosphere transmitting the reflected light from the lunar surface, atmospheric molecules and aerosols will cause attenuation and scattering of the light transmission. The Lambert-Beer atmospheric transmission equation describes the attenuation of the atmosphere on light. The atmospheric attenuation correction model is established through the measured atmospheric optical thickness data to obtain the true reflectivity of the lunar surface. The lunar surface reflectivity measurement is also affected by the observation geometry. Changes in the solar incidence angle and the detector observation angle will lead to differences in reflectivity data. When the lunar orbit satellite platform is in orbit, the attitude parameters of the satellite platform are obtained through star sensors and gyroscopes. The attitude parameters include roll angle, pitch angle and yaw angle. The observation angle of the detector to the lunar surface is calculated according to the satellite orbit position and attitude parameters. At the same time, the solar incidence angle is calculated according to the position of the moon and the sun. The reflectivity data is standardized using the lunar surface bidirectional reflection calculation formula to eliminate the influence of observation geometric conditions on the reflectivity. The extraction of lunar surface spectral characteristic parameters is an important means to characterize the composition and granularity characteristics of lunar surface materials. The lunar surface reflectance spectrum curve is obtained by performing band combination calculation on the reflectance data under standard geometric conditions. For lunar surface samples, their reflectance spectrum curves show unique absorption features in different bands. For example, ilmenite has absorption features in the 415-nanometer and 650-nanometer bands, and pyroxene has absorption features in the 900-nanometer and 2000-nanometer bands. By performing least square fitting on the spectral curve, the characteristic parameters of spectral slope, absorption band depth and band ratio are extracted. These three characteristic parameters are closely related to the composition and granularity characteristics of lunar surface materials. The lunar surface sample spectrum measurement experiment provides basic data for particle size inversion. The ASD field spectrometer is used to measure the spectrum of lunar surface samples with different particle sizes. The sample particle size ranges from 45 microns to 500 microns. The wavelength of the measured data was calibrated by the built-in calibration function of the spectrometer, and the spectral data was standardized by the area normalization method to establish a standard spectral library of lunar surface samples. Based on the spectral library data, the relationship between the lunar surface particle size and the spectral characteristic parameters was analyzed. It was found that as the particle size increases, the slope of the spectral curve gradually decreases and the depth of the absorption band gradually increases. This relationship provides a theoretical basis for the inversion of lunar surface particle size. The support vector regression model inputs spectral feature parameters, optimizes model parameters through training data, and finally outputs the estimated value of lunar surface particle size. The model is verified by cross-validation method, and the root mean square error is less than 50 microns.

[0017] Furthermore, in this embodiment, step S2 specifically includes: The spectrum configuration scheme includes the light intensity distribution ratio of each band and the lighting angle setting parameters; Conduct spectral scanning on lunar soil samples to obtain the spectral reflectance curves and absorption peak wavelengths of lunar soil samples in the visible light band, near infrared band and short-wave infrared band; The band transmittance of the combined lens is calculated using a ray tracing program according to the spectral reflectance curve, and the spectral transmittance data of the combined lens in different bands are obtained; The spectral reflectance data and spectral transmittance data were processed using the Bayesian optimization algorithm to obtain the spectral intensity distribution scheme; Based on the spectral transmittance data and spectral intensity distribution scheme, the Monte Carlo ray tracing method is used to calculate the light intensity distribution map of the illumination beam on the lunar soil surface and obtain the light intensity distribution uniformity index; Generate the optimal lighting angle parameters according to the light intensity distribution uniformity index; Here is an example: Based on the lunar soil reflectance database of the lunar polar regions, lunar soil types are classified according to their particle size and mineral composition. Reflectance characteristic curves are extracted for each type of lunar soil in the visible light band, near infrared band, and short-wave infrared band to obtain the spectral reflectance and absorption peak wavelength of each type of lunar soil. For the rectangular combined lens used for lunar rover lighting, a ray tracing program is used to calculate the transmission trajectory of light in each incident direction. By adjusting the surface curvature and thickness distribution of the lens, spectral transmittance curves of different bands are obtained, and the band transmittance data set of the combined lens is established. According to the lunar soil spectral reflectance data and the combined lens band transmittance data, the illumination spectrum optimization objective function is established. The function input is the light intensity distribution ratio of each band, and the output is the lunar soil reflection signal intensity. The optimal spectrum configuration scheme is solved by the Bayesian optimization algorithm. The Monte Carlo ray tracing method is used to calculate the scattering distribution of the illumination beam on the lunar soil surface based on the band transmittance and light intensity distribution ratio of the combined lens, and obtain the light intensity distribution uniformity index; According to the light intensity distribution uniformity index, the lighting angle optimization function is established. The function input is the tilt angle and rotation angle of the combined lens, and the output is the standard deviation of the light intensity in the lighting area. The optimal lighting angle parameters are solved by the gradient descent method. The calculation formula of the lighting angle optimization function is as follows:

[0018] in, represents the lighting angle optimization function, represents the tilt angle of the combined lens, represents the rotation angle of the combined lens, represents the number of sampling points in the lighting area, Indicates The light intensity at each sampling point is Represents the average light intensity of all sampling points. This formula calculates the standard deviation of the light intensity in the lighting area and is used to evaluate the lighting uniformity; Based on the optimal spectrum configuration scheme and lighting angle parameters, the installation position coordinates and posture angles of the combined lens are calculated, and the spatial layout parameter table of the lighting component is generated, including the three-dimensional coordinate values ​​of the lens center and the Euler angle values; The lunar soil reflectance database contains the spectral reflectance characteristics of different types of lunar soil on the lunar surface, which are classified based on the composition and particle size of the lunar soil. The particle size range of the lunar soil covers 45 microns to 500 microns, and the mineral composition mainly includes pyroxene, plagioclase and iron-titanium oxide. The lunar soil exhibits different spectral reflectance characteristics in the visible, near-infrared and short-wave infrared bands. For example, the pyroxene lunar soil has obvious absorption characteristics in the 900-nanometer and 2000-nanometer bands, with the reflectivity reduced to 20% and 25% respectively, while the reflectivity in the 750-nanometer band can reach 65%. The lunar rover lighting system uses a rectangular combined lens for beam modulation. The combined lens is composed of a refractive lens and a diffraction grating. The spatial modulation of the beam is achieved through the curvature and thickness distribution of the lens surface. Ray tracing calculations show that when the angle between the incident light and the normal line of the lens surface is within 30 degrees, the transmittance of the lens in the visible light band remains above 85%, while in the near-infrared band, as the incident angle increases, the transmittance shows a significant attenuation characteristic, and when the incident angle reaches 25 degrees, the transmittance decreases to 65%; The illumination spectrum optimization design is based on the lunar soil reflection characteristics and lens transmission characteristics. By adjusting the light intensity distribution ratio of different bands, the signal-to-noise ratio of the lunar soil reflection signal can be improved. The Bayesian optimization algorithm predicts the intensity of reflected signals under different spectral configurations by constructing a Gaussian process regression model. The optimization results show that the optimal light intensity distribution ratios for the visible light band, near infrared band, and short-wave infrared band are 50%, 35%, and 15%, respectively. The spatial distribution characteristics of the light beam directly affect the uniformity of the lunar soil reflection signal. The Monte Carlo ray tracing method simulates the scattering process of 1 million light rays on the lunar soil surface. The calculation results show that when the side length of the illuminated area is 500 mm, the difference in light intensity between the central area and the edge area reaches 45%. By adjusting the illumination angle of the combined lens, the uniformity of the light intensity distribution can be improved. When the lens tilt angle is set to 15 degrees and the rotation angle is set to 45 degrees, the standard deviation of the light intensity in the illuminated area is reduced to 12%; The spatial layout of the lighting assembly involves the coordinated configuration of multiple lens units, and the position and posture of each lens unit need to be precisely controlled; The lens installation position is expressed in a three-dimensional rectangular coordinate system, with the origin at the center of the lunar rover, the x-axis pointing to the front of the rover, the y-axis pointing to the left side of the rover, and the z-axis pointing vertically upward; The lens attitude is expressed by Euler angles, which rotate around the z-axis, y-axis, and x-axis in sequence, corresponding to the yaw angle, pitch angle, and roll angle, respectively; In the layout scheme obtained through optimization calculation, the four lens units are installed at a distance of 200 mm from the origin, forming a symmetrical distribution, and the Euler angle configuration ensures overlapping coverage of the lighting area.

[0019] Furthermore, in this embodiment, step S3 specifically includes: The lunar soil surface reflectance data is obtained according to the band calibration curve, and the lunar soil surface reflectance data is subjected to spectral noise reduction and normalization processing to obtain spectral characteristic parameters including spectral slope, spectral depth and band ratio; The similarity between the spectral characteristic parameters and the standard data in the pre-established lunar soil reflectance characteristic database is calculated, and the lunar soil type identification result and mineral content value are obtained through the k-nearest neighbor algorithm; According to the lunar soil type identification results and mineral content values, the power adjustment coefficient of each band light source is obtained by using a recurrent neural network, and the transmitted light intensity of each band light source is modulated by a liquid crystal spectrum modulation array. Here is an example: The multispectral camera carried by the lunar rover is used to scan the working area in real time to obtain raw multispectral image data. The image data is preprocessed using dark current correction and spectral calibration methods. The reflectivity data of the lunar soil surface in the visible light, near infrared and short-wave infrared bands are extracted according to the band calibration curve. Perform spectral noise reduction and normalization on the lunar soil reflectivity data, resample the reflectivity curve according to the preset wavelength interval, and extract spectral characteristic parameters including spectral slope, spectral depth and band ratio; The similarity between the spectral characteristic parameters and the standard data in the lunar soil reflectance characteristics database was calculated, the lunar soil type was identified using the k-nearest neighbor algorithm, and the lunar soil mineral content was quantitatively inverted based on the spectral absorption characteristics. For the identified lunar soil type and mineral content, a regression neural network is used to calculate the optimal light intensity distribution ratio. The input layer of the neural network contains the lunar soil type feature vector and mineral content vector, and the output layer is the power adjustment coefficient of the light source in each band. Generate lighting configuration instructions according to the light source power adjustment coefficient, modulate the transmitted light intensity of each band light source through the liquid crystal spectrum modulation array of the combined lens, and establish a configuration data table including the band number, light intensity coefficient and modulation voltage; The light intensity feedback control unit is used to collect the reflected signal intensity of the lunar soil surface in real time, calculate the deviation between the measured light intensity and the target light intensity, and dynamically compensate the control voltage of the spectral modulation array through the proportional integral controller; The multispectral camera carried by the lunar rover uses an array CCD sensor with a spectral range of 380-2500 nanometers, including 8 characteristic bands, including 3 visible light bands, 3 near-infrared bands, and 2 short-wave infrared bands; During the operation of the camera, the dark current noise will change with the temperature. Under the condition that the temperature of the lunar surface varies from -180 to 120 degrees Celsius, the dark current noise can reach 5% of the full scale. Inserting a light shield in the optical path to collect dark field images, obtain dark current distribution data, establish a temperature-dark current lookup table, and achieve dark current correction; The spectral calibration uses an external spectral calibration source. The spectral radiance of the calibration source fluctuates less than 2% in the range of 380-780 nanometers. The spectral calibration establishes the corresponding relationship between the CCD output signal and the radiance. The lunar soil reflectivity data was processed by spectral noise reduction and normalization to obtain a standard spectral curve, which was then resampled at a wavelength interval of 20 nanometers to extract spectral characteristic parameters. The spectral slope describes the trend of reflectivity changing with wavelength. In the ilmenite-dominated area, the spectral slope in the visible light band is significantly higher than that in other areas; The spectral depth characterizes the intensity of the absorption feature. Pyroxene minerals have characteristic absorption valleys at 900 nanometers and 2000 nanometers, and the absorption depth is positively correlated with the mineral content. The band ratio uses the reflectance ratio of adjacent bands, which is insensitive to changes in lighting conditions. The lunar soil type identification uses the k-nearest neighbor algorithm, which measures the similarity between the spectral characteristic parameters and the database samples based on the Euclidean distance, and the k value is set to 5. The database contains spectral data of typical lunar soil samples, with 100 samples of each type, covering different particle sizes and composition ratios; Mineral content inversion is based on spectral absorption characteristics, and a quantitative relationship between absorption band depth and mineral content is established. Spectral configuration optimization uses a three-layer neural network, the input layer contains 24 nodes, corresponding to the one-hot coding of lunar soil types and the content values ​​of major minerals, the hidden layer uses 48 nodes, and the output layer has 8 nodes corresponding to the power adjustment coefficients of light sources in each band; The network training uses samples from the lunar soil reflectance database to achieve uniform light intensity distribution by minimizing the variance of the reflected signal. The liquid crystal spectrum modulation array of the combined lens is composed of 8 independently controlled liquid crystal units, each unit corresponds to a spectral band, and the light intensity modulation is achieved by adjusting the voltage applied to the liquid crystal unit. The voltage range is 0-10 volts, and the modulation depth is greater than 90%; The light intensity feedback control adopts a proportional-integral controller with a sampling period of 20 milliseconds, a proportional coefficient of 0.5, an integral time constant of 100 milliseconds, and a control accuracy better than 3%.

[0020] Furthermore, in this embodiment, step S4 specifically includes: Obtain lunar soil reflectivity data collected by the lunar rover's multispectral camera, perform dark current correction and spectral calibration on the lunar soil reflectivity data, and obtain a lunar soil reflectance spectrum data set; The maximum and minimum normalization method was used to standardize the lunar soil reflectance spectrum data set. Based on the standardized data, the spectral characteristic parameters of spectral slope, absorption band depth and band ratio were extracted. A deep neural network model is constructed based on the spectral feature parameters, the spectral feature parameters are mapped to the latent feature space through the encoding layer of the deep neural network, and the latent features are converted into spectral configuration parameters through the decoding layer of the deep neural network; Generate a lighting component control instruction sequence according to the spectrum configuration parameters, the control instruction sequence including a band number, a power coefficient and a modulation voltage parameter, and send the control instruction sequence to a lighting controller through a serial communication interface; Here is an example: According to the lunar soil reflectivity data collected by the lunar rover's multispectral camera, the raw data was subjected to dark current correction and spectral calibration. The median filtering method was used to remove abnormal sampling points with a signal-to-noise ratio lower than the preset threshold. The lunar soil reflectivity data was classified and sorted according to the band number to establish a lunar soil reflectance spectrum data set. For the reflectance spectrum data set, the maximum and minimum normalization method was used to standardize the data, and the spectral characteristic parameters including spectral slope, absorption band depth and band ratio were extracted; The lunar soil reflectivity data were grouped according to the operation time sequence, the power coefficient of each band light source and the lens modulation voltage of the lighting system at the corresponding time were extracted as the annotation data, and the K-fold cross-validation method was used to divide the training data set and the validation data set. A deep neural network consisting of an encoding layer and a decoding layer is constructed. The encoding layer maps spectral feature parameters to a latent feature space, and the decoding layer converts the latent features into spectral configuration parameters. The network parameters are optimized using a mean square error loss function. Based on the validation data set, the relative error between the model prediction result and the actual configuration parameters is calculated, and the features of the samples whose errors exceed the preset range are re-extracted, and the training data set is updated; The trained neural network is used to predict the spectral configuration of the lunar soil reflectivity data collected in real time, and the prediction results are smoothed using a sliding mean window of length 5. Generate a control instruction sequence for the lighting component according to the smoothed prediction result, the control instruction sequence including three control parameters: band number, power coefficient and modulation voltage, and send it to the lighting controller through a serial communication interface; The light intensity sensor is used to collect the irradiance distribution of the illuminated area in real time, and the deviation between the actual configuration parameters and the predicted parameters is calculated. When the deviation exceeds the threshold, the model parameters are triggered to update online. The collection process of lunar soil reflectivity data is affected by various noise sources, among which dark current noise fluctuates with temperature, which is particularly significant in the extreme temperature environment on the lunar surface. When the temperature rises from -180 degrees Celsius to 120 degrees Celsius, the dark current noise can reach 5% of the sensor's full scale, which is corrected by acquiring a dark field image before each acquisition; The spectral calibration uses an external spectral calibration source. The spectral radiance of the calibration source fluctuates less than 2% in the range of 380-2500 nanometers. The mapping relationship between the sensor output signal and the radiance is established. In data preprocessing, a median filter with a window length of 5 is used to remove impulse noise. When the signal-to-noise ratio of a sampling point is less than 6 decibels, it is determined to be an abnormal point; The maximum and minimum value normalization maps the data to the range of 0-1 to eliminate the differences in the scales of data in different bands. The spectral feature extraction includes three parameters: spectral slope, absorption band depth and band ratio. The spectral slope describes the rate of change of reflectivity with wavelength. The slope value reaches 0.8 in the ilmenite-dominated area, while it is only 0.3 in the feldspar-dominated area. The spectrum configuration parameters of the lighting system include the power coefficient and lens modulation voltage of 8 bands, the power coefficient range is 0-1, and the modulation voltage range is 0-10V; The neural network adopts an encoder-decoder structure. The encoding layer contains three fully connected layers with 64, 32, and 16 nodes respectively. The ReLU activation function is used, and the decoding layer is symmetrically designed. The training adopts the stochastic gradient descent method with a batch size of 32, the initial value of the learning rate is 0.01, and it decays by 50% every 50 rounds; The model validation uses 5-fold cross validation to calculate the relative error between the predicted results and the actual configuration parameters. When the power coefficient error exceeds 10% or the modulation voltage error exceeds 5%, the corresponding samples are re-extracted. The sliding mean window length is set to 5, and the prediction results are time-series smoothed to suppress parameter fluctuations. The control instructions are transmitted through a serial interface with a baud rate of 115200. Each group of instructions contains 24 bytes of data, of which the band number occupies 1 byte, and the power coefficient and modulation voltage each occupy 2 bytes; The lighting effect monitoring uses a 16-channel light intensity sensor array, with sensors evenly distributed in the lighting area at 50 mm intervals, and a sampling frequency of 50 Hz; When the root mean square error between the actual configuration parameters and the predicted parameters exceeds the preset threshold of 0.1, the model's online update mechanism is triggered, and the network parameters are retrained using the most recent 1,000 sets of samples. The update cycle does not exceed 60 seconds to ensure that the model adapts to the dynamic changes in lunar soil properties.

[0021] Furthermore, in this embodiment, step S5 specifically includes: According to the data collected by the current and voltage monitoring unit, the real-time power values ​​of the lighting components and the mining components are calculated, and the energy consumption data of the lighting components and the mining components are obtained through time integration operation; A depth sensor is used to record the excavation operation parameters to obtain the excavation depth data, and the energy consumption data and the excavation depth data are integrated to obtain the energy consumption data set per unit volume of ice layer; A mapping function between spectral configuration parameters and unit energy consumption is constructed according to the energy consumption data set, and the weight coefficient of the mapping function is determined by the least square method; The wavelength sampling point set is obtained from the lunar soil reflectivity curve, and the initial value of light intensity is iteratively optimized using the gradient descent method. If the deviation between the predicted energy consumption and the measured value exceeds the deviation threshold, the light intensity parameter is updated: Here is an example: According to the current and voltage data collected by the lunar rover's power monitoring unit, the real-time power values ​​of the lighting components and the excavation components are calculated at fixed time intervals, and the cumulative energy consumption data of each component is obtained through time integration calculation; Use depth sensors and area scanners to record excavation parameters, obtain excavation depth curves and operation area data, calculate excavation volume by depth integration, and establish a data set of energy consumption per unit volume of ice. The energy consumption data set is normalized, abnormal data points beyond the preset range are removed, and the support vector regression method is used to construct the mapping function between the spectrum configuration parameters and unit energy consumption; The mining operation optimization objective function is established based on the mapping function. The function input includes the spectral wavelength and the corresponding light intensity value, and the output is the predicted unit energy consumption index. The least squares method is used to determine the objective function weight coefficient. For the preset visible light band range, sampling is performed at fixed wavelength intervals, a set of wavelength sampling points is obtained based on the lunar soil reflectivity curve, and the initial value of the light intensity of each sampling point is calculated through the spectral calibration curve; The gradient descent method is used to optimize the initial value of light intensity, and the energy consumption is predicted through objective function calculation. The deviation between the predicted value and the measured value is determined, and the light intensity parameters with deviations exceeding the threshold are iteratively updated. Generate an updated spectrum configuration scheme based on the iterative optimization results, including three configuration parameters: wavelength number, wavelength value and corresponding light intensity, and establish a spectrum configuration data table; Use a spectrum analyzer to verify the configuration scheme online, calculate the correlation coefficient between the actual output spectrum and the target spectrum, and trigger parameter correction when the correlation coefficient is lower than the threshold; Energy consumption monitoring during lunar rover operations involves multiple components, among which the lighting component and the excavation component are the main energy consumption units; Power monitoring uses high-precision current sensors and voltage sampling circuits with a sampling frequency of 100 Hz, a current measurement range of 0-20 amps, and a voltage measurement range of 0-48 volts; Through real-time power calculation, it is found that the power fluctuation range of the lighting component is 200-800 watts, and the power fluctuation range of the mining component is 1000-3000 watts; The energy consumption is obtained through integration operation, and the integration time window is set to 1 second. The depth sensor adopts the principle of laser ranging, with a measurement range of 0-500 mm and a resolution of 0.1 mm. The excavation depth curve is obtained through continuous scanning; The area scanner is based on the principle of structured light, projects a grid of light spots on the working area, and obtains the working area data through image processing. The area measurement accuracy is better than 1 square centimeter. The excavation volume is calculated by integrating the depth curve, and the energy consumption per unit volume shows that when the lunar soil type is basalt, the lighting energy consumption accounts for about 15% and the excavation energy consumption accounts for about 85%; The energy consumption data is normalized using the maximum and minimum value method to map the data to the 0-1 interval. The abnormal data is determined based on the principle of 3 times the standard deviation. When the data deviates from the mean by more than 3 times the standard deviation, it is determined as an abnormal point. Support vector regression uses radial basis kernel function, and the kernel parameters are determined by cross-validation. The training data contains 1000 sets of samples, and the verification accuracy reaches 90%; The spectrum optimization objective function comprehensively considers the two indicators of energy consumption and efficiency. The weight coefficient is determined by the least squares method. The weight of the energy consumption indicator is 0.6, and the weight of the efficiency indicator is 0.4. The spectrum optimization objective function is as follows:

[0022] in, represents the spectral optimization objective function, Indicates the energy consumption index. represents the efficiency index, 0.6 and 0.4 are the weight coefficients of energy consumption and efficiency, respectively, which are determined by the least square method; In the visible light band of 380-780 nanometers, sampling is performed at intervals of 20 nanometers to obtain 20 wavelength sampling points; The spectral calibration curve is obtained from a standard light source, and the spectral radiance fluctuation of the calibration source is less than 2%; The light intensity optimization adopts the gradient descent method with adaptive step size. The initial step size is set to 0.1. When the prediction error is less than 5% for three consecutive iterations, the step size is reduced by 50%; The optimization results show that the light intensity in the 450-550 nanometer band should account for 60%, which corresponds to the higher reflectivity of lunar soil in this band. The spectrum configuration data table contains three parameters: wavelength number, wavelength value and light intensity. The wavelength number is 1-20, the wavelength value is 380-780 nanometers, and the light intensity is a normalized value between 0-1. Spectral verification uses a spectrum analyzer to monitor the output spectrum in real time, with a sampling period of 100 milliseconds and a spectral resolution of 2 nanometers; The correlation coefficient is calculated using the Pearson correlation coefficient. When the correlation coefficient is lower than 0.95, parameter correction is triggered. The correction process includes three steps: re-collecting spectral data, updating calibration parameters, and adjusting control parameters. The entire correction process takes no more than 1 second.

[0023] Furthermore, in this embodiment, step S6 specifically includes: The lunar surface point cloud data is obtained by using the LiDAR scanner, and the lunar surface elevation distribution map is obtained by projecting it onto the horizontal plane through coordinate transformation; The least squares surface fitting is performed on the lunar surface elevation distribution map to obtain a set of terrain characteristic parameters including slope value, aspect angle and curvature value; A multispectral camera is used to monitor the light intensity in the working area, and the light blocking area is calculated by combining the terrain feature parameter set and the azimuth and altitude angles output by the sun position calculator; According to the parameter set of light blocking area and terrain characteristics, the local search algorithm is used to optimize the posture angle of the lighting component to obtain the optimal installation angle of the lighting component; Here is an example: The three-dimensional point cloud data of the lunar surface is obtained by the laser radar scanner carried by the lunar rover. The point cloud data is projected onto the horizontal plane through coordinate transformation. The elevation change curve is obtained by terrain undulation calculation, and the lunar surface elevation distribution map is established. The terrain slope and aspect parameters are calculated based on the lunar surface elevation distribution map, and the terrain surface is fitted using the least squares method to extract a set of terrain characteristic parameters including slope value, aspect angle and curvature value. A multispectral camera is used to monitor the light intensity in the working area, and the sun azimuth and altitude angle are obtained according to the sun position calculator, and the light shielding area is calculated in combination with the terrain feature parameter set; According to the distribution of light-blocked areas, a convolutional neural network is used to calculate the light intensity compensation of the working area. The network input is the terrain feature parameter set, and the output is the compensated light intensity value. According to the compensation light intensity value and the terrain feature parameter set, the azimuth angle and pitch angle of the lighting component are solved by the steepest descent method to establish the initial attitude angle parameters of the lighting component; Extract the bidirectional reflection distribution function from the lunar soil reflection characteristic database, and establish the reflection light intensity prediction function by combining the spectral configuration parameters of the lighting component. The local search algorithm is used to optimize the posture angle parameters of the lighting component, and the reflected light intensity prediction function is used as the optimization target to obtain the optimal installation angle of the lighting component. Adjust the lighting component posture according to the optimal installation angle, monitor the lighting area in real time through the light intensity distribution sensor array, and trigger posture fine-tuning when the light intensity uniformity is lower than the threshold; The lunar surface topography was measured using a lidar scanner with an emission wavelength of 1064 nanometers, a scanning frequency of 100 Hz, and an angular resolution of 0.1 degrees; The density of the 3D point cloud data obtained by scanning reaches 1,000 points per square meter. The point cloud data includes spatial coordinate values ​​and reflection intensity values; The point cloud data is projected onto the horizontal plane using coordinate transformation, and an elevation grid with a resolution of 10 cm is calculated. The elevation variation range reaches 50 cm within the working area. The extraction of terrain parameters is the basis for optimizing the lighting angle. The slope describes the angle between the ground surface and the horizontal plane. The slope range in the working area is 0-30 degrees. The slope aspect indicates the azimuth of the maximum slope, ranging from 0 to 360 degrees. The curvature reflects the degree of terrain concavity and convexity, and is calculated using the second-order derivative. A positive value indicates convex terrain, and a negative value indicates concave terrain. The elevation data was fitted with a surface using the least squares method, with a fitting residual error of less than 1 cm. Light intensity was monitored using an 8-channel multispectral camera with a wavelength range of 380-1100 nanometers. The calculation of the sun's position takes into account the lunar orbital parameters, including an orbital period of 27.32 days, a rotation period of 27.32 days, and an orbital inclination of 5.145 degrees; In the lunar polar regions, the solar altitude angle varies from 0 to 1.5 degrees, resulting in a significant terrain shadow effect. The light-blocked area accounts for 30% of the operating area, and the light intensity in the blocked area is only 10% of the direct area. The convolutional neural network adopts a 5-layer structure. The input layer contains terrain feature parameters and sun position parameters. Three convolutional layers are used to extract spatial features. The last two fully connected layers output the compensated light intensity. The network training used 1,000 sets of illumination data, and the verification accuracy reached 95%. The uniformity of illumination intensity distribution after compensation was increased to 85%; The lighting component posture optimization adopts the steepest descent method, with an azimuth search range of 0-360 degrees and an elevation search range of 0-90 degrees. The step size of each iteration is adaptively adjusted. When the objective function improvement of three consecutive iterations is less than 1%, it is considered to have reached the local optimal solution. Experiments show that the optimized lighting angle reduces the coefficient of variation of the light intensity distribution to 0.15; The bidirectional reflectance distribution function describes the reflectance characteristics of the lunar soil surface. The reflectance intensity varies nonlinearly with the observation angle and the incident angle. Under vertical incidence conditions, the reflectivity can reach up to 0.4, while it drops to 0.1 under grazing incidence conditions. Combined with the spectral configuration parameters, a prediction model for reflected light intensity was established with a prediction accuracy better than 90%; Light intensity distribution monitoring uses a 4x4 sensor array with sensors spaced 50 mm apart to cover the entire operating area; The light intensity data collected in real time is used to calculate the uniformity index. When the uniformity is lower than 0.8, the attitude fine-tuning mechanism of the lighting component is activated, and the adjustment accuracy is better than 0.1 degrees. The stability of the lighting effect is maintained through closed-loop control to adapt to the dynamic changes in lunar surface conditions.

[0024] Furthermore, in this embodiment, step S7 specifically includes: The temperature sensor arrays set in the center and edge areas of the lens collect the temperature values ​​of the temperature measuring points, and the temperature values ​​of the temperature measuring points are obtained by synchronous sampling by the data acquisition module; According to the temperature values ​​of the temperature measurement points, the neural network spatial interpolation calculation is performed, and the output layer of the neural network generates the lens surface temperature distribution data; The temperature field uniformity index value is calculated based on the lens surface temperature distribution data, and the temperature field uniformity index value determines the area with drastic temperature changes through temperature gradient calculation; According to the temperature field uniformity index value of the area with drastic temperature changes, the fuzzy rule base is queried, and the fuzzy rule base outputs the power value of the semiconductor refrigerator and the power value of the heating wire to the temperature closed-loop controller; Here is an example: The temperature sensor arrangement position is determined according to the structural parameters of the combined lens, and the temperature sensor array is arranged in the central area and the edge area of ​​the lens, and the temperature value of each temperature measurement point is synchronously sampled through the data acquisition module; A neural network is used to perform spatial interpolation on discrete temperature data collected by the temperature sensor array. The network input layer is the coordinate value and temperature value of the temperature measurement point, the hidden layer contains a spatial feature extraction unit, and the output layer generates the lens surface temperature distribution data. The temperature field uniformity is calculated based on the lens surface temperature distribution data, the temperature distribution dispersion is evaluated using the standard deviation, and the area with drastic temperature changes is determined by temperature gradient calculation; For the temperature field distribution characteristics, a temperature closed-loop controller is established based on the fuzzy rule base. The controller input includes the temperature uniformity index and the temperature gradient value, and the output includes the semiconductor cooler power value and the heating wire power value. Obtaining the upper and lower temperature limits from a preset temperature threshold range, and when the measured temperature exceeds the threshold range, sending a control instruction sequence including the actuator number, power value and action time to the actuator; The Kalman filter is used to suppress the noise of the temperature data fed back by the actuator, the temperature change trend is calculated by the curve fitting method, and a performance indicator set including the temperature overshoot value and the adjustment cycle is established; Optimize temperature control parameters according to the performance index set, use an adaptive algorithm to adjust the proportional coefficient and integral time of the controller, and generate an updated control parameter configuration table; The control parameter configuration table is downloaded to the actuator through the serial communication interface, and the actuator adjusts the output power of the refrigerator and the heater in real time according to the configuration parameters; The temperature control of the combined lens involves multiple technical links, among which the arrangement of temperature sensors is the key foundation; On the surface of the 200 mm diameter lens, a 5x5 array of temperature sensors is arranged, with a sensor spacing of 40 mm; Each sensor uses PT100 platinum resistance, with a temperature measurement range of -200 to 150 degrees Celsius, a measurement accuracy of 0.1 degrees Celsius, and a response time of less than 1 second; The data acquisition module uses a 24-bit ADC with a sampling frequency of 100 Hz, and all sensors are sampled synchronously; The temperature spatial interpolation uses a three-layer neural network, where the input layer contains the spatial coordinates and temperature value of the sensor, a total of 4 parameters; The hidden layer uses 32 neurons to extract spatial features through the ReLU activation function, and the output layer generates a 100x100 temperature distribution matrix with a spatial resolution of 2 mm; The network training uses standard temperature field data collected by thermal imagers, with 1,000 sets of training samples, and the verification accuracy is better than 0.2 degrees Celsius; The temperature field uniformity is evaluated using the standard deviation index. Under normal working conditions, the temperature standard deviation should be less than 2 degrees Celsius; The temperature gradient was calculated using the Sobel operator, and areas with gradient values ​​exceeding 5 °C / cm were marked as hot spots; The fuzzy controller contains 25 rules, the input variables are temperature uniformity and maximum gradient value, and the output variables are cooling power and heating power; The power adjustment range is 0-100 watts, with a minimum adjustment step of 1 watt. The temperature threshold range is set according to the lens material characteristics, with an upper limit temperature of 50 degrees Celsius and a lower limit temperature of -20 degrees Celsius. The control instruction adopts 32-byte format, including 8-bit actuator number, 16-bit power value and 8-bit action time; The actuator contains 8 semiconductor coolers and 8 heating wires, which are arranged on the edge and back of the lens respectively. The cooling coefficient of the semiconductor cooler is 1.2, and the maximum temperature difference is 60 degrees Celsius. The heating wire is made of nickel-chromium alloy with a resistance of 10 ohms / meter. During the temperature control process, the measurement noise variance of the Kalman filter is set to 0.01 and the process noise variance is 0.001; The temperature change trend is fitted by a third-order polynomial with a fitting interval of 60 seconds. The performance indicators include temperature overshoot and adjustment time. The overshoot is limited to less than 10% and the adjustment time is required to be less than 30 seconds. The adaptive control algorithm dynamically adjusts PID parameters according to performance indicators, with a proportional coefficient range of 1-10, an integral time range of 1-100 seconds, and a differential time range of 0.1-1 second; The control parameter configuration table is transmitted to the actuator via the RS485 interface, with a baud rate of 115200 and a communication cycle of 100 milliseconds; The actuator has a built-in 32-bit microcontroller, which uses PWM to control power output, with a PWM frequency of 20 kHz and a duty cycle resolution of 0.1%; During the closed-loop control process, the uniformity of the temperature field was improved by 80%, the maximum temperature gradient was reduced by 70%, and the system stability and dynamic response performance were significantly improved.

[0025] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "lateral, vertical, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the device or element referred to must have a specific direction or be constructed and operated in a specific direction. Therefore, they cannot be understood as limiting the scope of protection of the present invention.

[0026] For those skilled in the art, various other corresponding changes and deformations can be made according to the technical solutions and concepts described above, and all of these changes and deformations should fall within the protection scope of the claims of the present invention.

Claims

1. A data processing method for a rectangular combined lens for deep space exploration, characterized in that: The following steps are involved: S1. Obtain multispectral imaging data of the lunar polar regions, perform preprocessing and calculate phase angles to obtain lunar soil reflectivity and bidirectional reflectivity characteristics in different bands. By comparing and analyzing the spectral data of samples with preset particle sizes, establish the corresponding relationship between reflectivity and lunar soil particle size, and establish a lunar soil reflectivity characteristics database; S2. Obtaining reflection characteristic parameters corresponding to different lunar soil types when digging ice at the lunar poles from the lunar soil reflection characteristic database, and obtaining a corresponding spectrum configuration scheme for the illumination rectangular combination lens in the lunar rover illumination system; S3. During the lunar rover's ice excavation operation, the lunar soil reflectivity data of the operation area is obtained in real time, and compared with the data in the lunar soil reflectivity characteristic database, and the spectrum configuration scheme of the lunar rover's lighting system is dynamically adjusted according to the type and particle size distribution characteristics of the lunar soil in the current operation area; S4. Based on the lunar soil reflectivity data collected during the lunar rover's ice excavation operation, the target lighting spectrum configuration parameters under different lunar soil conditions are predicted and applied to the dynamic adjustment process of the lunar rover's lighting system; S5, obtaining the energy consumption and ice-digging efficiency of the lunar rover during the ice-digging operation, and performing correlation analysis with the target illumination spectrum configuration parameters, selecting a target wavelength point within a preset visible light range, calculating the light intensity configuration of each wavelength, and obtaining an updated spectrum configuration scheme; S6. According to the terrain and lighting conditions of the lunar polar regions and the updated spectrum configuration scheme, the installation angle and illumination direction of the lunar rover lighting system are adjusted once, and according to the bidirectional reflection data in the lunar soil reflection characteristic database, the installation angle of the lunar rover lighting system is adjusted twice; S7. Temperature control is performed on the combined lens assembly in the optimized lunar rover lighting system. The temperature status of the combined lens is monitored in real time through a preset thermal management system, and the combined lens whose temperature exceeds a preset temperature threshold range is cooled or heated.

2. The data processing method of the deep space exploration illumination rectangular compound lens according to claim 1, characterized in that: The step S1 specifically includes: Using a satellite-borne multispectral detector to obtain lunar surface reflectance data from the lunar polar regions, and performing radiometric calibration on the lunar surface reflectance data according to the spectral response curve of the multispectral detector to obtain lunar surface reflectance data; For the lunar surface reflectivity data, the solar incidence angle and the detection angle are calculated according to the attitude parameters of the lunar orbit satellite platform, and the lunar surface bidirectional reflection calculation formula is used to perform angle normalization to obtain the reflectivity data under standard geometric conditions; Performing band combination operation on the reflectance data under the standard geometric conditions, and extracting spectral characteristic parameters by least square fitting, wherein the spectral characteristic parameters include spectral slope, absorption band depth and band ratio; A mapping relationship between the spectral characteristic parameters and the lunar surface particle size is established, a lunar surface particle size inversion model is constructed using a support vector regression method, and particle size mapping of the lunar polar regions is performed using the lunar surface particle size inversion model.

3. The data processing method of the deep space exploration illumination rectangular compound lens according to claim 1, characterized in that: The step S2 specifically includes: The spectrum configuration scheme includes the light intensity distribution ratio of each band and the lighting angle setting parameters; Performing spectral scanning on the lunar soil sample to obtain the spectral reflectance curve and absorption peak wavelength of the lunar soil sample in the visible light band, near infrared band and short-wave infrared band; Calculate the band transmittance of the combined lens using a ray tracing program according to the spectral reflectance curve to obtain spectral transmittance data of the combined lens in different bands; Using a Bayesian optimization algorithm to process the spectral reflectance data and the spectral transmittance data to obtain a spectral intensity distribution scheme; According to the spectral transmittance data and the spectral intensity distribution scheme, a Monte Carlo ray tracing method is used to calculate the light intensity distribution diagram of the illumination beam on the lunar soil surface to obtain a light intensity distribution uniformity index; Generate the optimal lighting angle parameters based on the light intensity distribution uniformity index.

4. The data processing method of the deep space exploration illumination rectangular compound lens according to claim 1, characterized in that: The step S3 specifically includes: Obtain lunar soil surface reflectivity data according to the band calibration curve, and perform spectral noise reduction and normalization processing on the lunar soil surface reflectivity data to obtain spectral characteristic parameters including spectral slope, spectral depth and band ratio; The similarity between the spectral characteristic parameters and the standard data in the pre-established lunar soil reflectance characteristic database is calculated, and the lunar soil type identification result and the mineral content value are obtained by using the k-nearest neighbor algorithm; According to the lunar soil type identification results and the mineral content values, a recurrent neural network is used to obtain the power adjustment coefficient of each band of light sources, and the transmitted light intensity of each band of light sources is modulated by a liquid crystal spectral modulation array.

5. The data processing method of the deep space exploration illumination rectangular compound lens according to claim 1, characterized in that: The step S4 specifically includes: Acquire lunar soil reflectivity data collected by a lunar rover multispectral camera, and perform dark current correction and spectral calibration on the lunar soil reflectivity data to obtain a lunar soil reflectivity spectrum data set; The lunar soil reflectance spectrum data set is standardized by using a maximum and minimum value normalization method, and spectral characteristic parameters of spectral slope, absorption band depth and band ratio are extracted based on the standardized data; Constructing a deep neural network model according to the spectral feature parameters, mapping the spectral feature parameters to a latent feature space through an encoding layer of the deep neural network, and converting the latent features into spectral configuration parameters through a decoding layer of the deep neural network; A lighting component control instruction sequence is generated according to the spectrum configuration parameters, the control instruction sequence including a band number, a power coefficient and a modulation voltage parameter, and the control instruction sequence is sent to a lighting controller via a serial communication interface.

6. The data processing method of the deep space exploration illumination rectangular compound lens according to claim 1, characterized in that: The step S5 specifically includes: According to the data collected by the current and voltage monitoring unit, the real-time power values ​​of the lighting components and the mining components are calculated, and the energy consumption data of the lighting components and the mining components are obtained through time integration operation; Using a depth sensor to record excavation operation parameters to obtain excavation depth data, integrating the energy consumption data and the excavation depth data to obtain an energy consumption data set per unit volume of ice layer; Constructing a mapping function between spectrum configuration parameters and unit energy consumption according to the energy consumption data set, and determining a weight coefficient of the mapping function by a least square method; A set of wavelength sampling points is obtained from the lunar soil reflectivity curve, and the initial value of light intensity is iteratively optimized using the gradient descent method. If the deviation between the predicted energy consumption and the measured value exceeds the deviation threshold, the light intensity parameter is updated.

7. The data processing method of the deep space exploration illumination rectangular compound lens according to claim 1, characterized in that: The step S6 specifically includes: The lunar surface point cloud data is obtained by using the LiDAR scanner, and the lunar surface elevation distribution map is obtained by projecting it onto the horizontal plane through coordinate transformation; Performing least squares surface fitting on the lunar surface elevation distribution map to obtain a terrain feature parameter set including slope value, aspect angle and curvature value; A multispectral camera is used to monitor the light intensity in the working area, and the light shielding area is calculated by combining the terrain feature parameter set and the azimuth and altitude angle output by the sun position calculator; According to the light shielding area and the terrain feature parameter set, a local search algorithm is used to optimize the posture angle of the lighting component to obtain the optimal installation angle of the lighting component.

Images