Design evaluation method for surveying and mapping load wave band spectral range

Abstract

The invention discloses a method for designing and evaluating a spectral range of a surveying and mapping load wave band, which comprises the following steps of: firstly, transmitting configuration parameters based on radiation under an imaging condition; the entrance pupil radiation characteristic result, the load imaging performance, the image texture characteristic and the spectrum distortion characteristic of the detection wave band designed based on different spectral range parameters at the surveying and mapping load entrance pupil after atmospheric radiation transmission are calculated; and secondly, according to the obtained entrance pupil radiation characteristic result, the imaging performance parameter calculation result, the image texture characteristic parameter calculation result and the spectral distortion characteristic calculation result, evaluating the comprehensive influence of different spectral range parameters on load imaging. According to the invention, the imaging quality and the image information amount of the surveying and mapping load can be evaluated by different spectrum broadening, and the optimal configuration of the spectrum range of the load detection wave band is realized.

Description

Technical Field

[0001] This invention relates to a design and evaluation method for the spectral range of a mapping payload, belonging to the field of mapping payload design and image processing technology. Background Technology

[0002] The combination of panchromatic and multispectral imaging in the design of 3D mapping payloads has been developed and applied in the field of surveying and mapping for many years. How to select and configure the detection band spectral range of the mapping optical payload to more effectively reflect the spectral characteristics and physical details of the target surface, obtain high-resolution and clear-textured imaging data, realize the acquisition of rich information after fusion, and then carry out target feature recognition, classification, information extraction, and land 3D measurement based on the effective texture features of image data is crucial for realizing high-precision 3D mapping and quantitative remote sensing applications. Currently, the optical payloads on high-resolution remote sensing satellites in orbit, both domestically and internationally, exhibit variations in the spectral range settings of their detection bands, particularly in the panchromatic band. Examples show lower-end boundary wavelengths varying between 430 and 500 nm, while higher-end boundary wavelengths range from 680 to 900 nm. However, considering the development trend of all high-resolution space optical remote sensing satellites in orbit, with the advancement and innovation of remote sensing technology and the evolving needs of mapping applications, spatial resolution has gradually improved, leading to optimization and adjustments in the spectral range settings of the panchromatic detection band. For instance, in the development of the French AIRBUS SPOT and Pleiades series, the high-end boundary settings of the panchromatic band range have expanded from 710 nm, 745 nm, and 830 nm to 800 nm, extending towards the near-infrared band. Similarly, the US Landsat optical mapping payload, designed to detect vegetated surfaces, has also been optimized. The contrast with the unvegetated surface is even greater. The panchromatic spectral range of the OLI imager on Landsat-8 / 9 is narrower than that of ETM+ on Landsat-7. The panchromatic spectral range of the optical payloads of the US IKONOS, QuickBird, and Worldview-1 remote sensing satellites is 450–900 nm. The panchromatic spectral range of the US Orbview fifth-generation satellite GeoEye and Worldview-2 / 3 / 4 / Legion mapping payloads is set to 450–800 nm, and all achieve sub-meter spatial resolution. In contrast, the panchromatic spectral range of the domestic Ziyuan-3 series three-line array mapping payloads is set to 500–800 nm, while the panchromatic spectral range of the Gaofen-7 dual-line array mapping payloads has been broadened to 450–900 nm.

[0003] In 2009, Shandong Normal University published "Research on Texture Information Evaluation of Image Fusion Based on CBERS-02B and SPOT-5 Panchromatic Bands," which compared and evaluated the fusion effects of data from different panchromatic band ranges with SPOT-5 multispectral data. In 2010, the Institute of Remote Sensing Applications, Chinese Academy of Sciences, published "Research on Panchromatic Band Broadening of a New Generation High-Resolution Satellite Remote Sensor," analyzing and comparing the atmospheric effects caused by broadening in the 0.50–0.85 μm and 0.45–0.90 μm working bands. In 2013, Beijing 21st Century Space Technology Application Co., Ltd. published "Optimizing the Design of Panchromatic Band Broadening for Spaceborne High-Resolution Optical Cameras from a Fusion Perspective," which analyzed and discussed the quantitative evaluation of spectral distortion in images fused using correlation coefficients and relative overall dimension, and analyzed the selection of panchromatic band broadening from the perspective of spectral fidelity preservation in fused images.

[0004] In the past, image data with different spectral broadenings were often interpreted and compared based on visual effects, but there is a lack of in-depth analysis for application purposes. Summary of the Invention

[0005] The technical problem solved by this invention is to overcome the shortcomings of the prior art and propose a design and evaluation method for the spectral range of a mapping payload. This method integrates multiple performance aspects such as entrance pupil radiation information, system dynamic range, signal-to-noise ratio, on-orbit imaging dynamic MTF factor, and image texture information to evaluate the imaging quality and image information content of the mapping payload under different spectral broadening, and thereby achieves the optimized configuration of the spectral range of the payload detection band.

[0006] The technical solution of this invention is:

[0007] A design evaluation method for the spectral range of a mapping payload includes:

[0008] Based on the radiative transfer configuration parameters under imaging conditions, the radiative information of the probe band after atmospheric radiative transfer under different spectral range parameters is calculated at the entrance pupil of the mapping payload. Based on the obtained radiative information, the radiative contrast and atmospheric scattering ratio corresponding to different spectral range parameters are calculated as the entrance pupil radiative characteristics results.

[0009] Calculate the payload imaging performance under different spectral range parameter designs, including system dynamic range, signal-to-noise ratio, and dynamic MTF influence factor;

[0010] Construct an imaging simulation dataset under different spectral range parameters for the mapping payload; construct an image texture information model based on bidirectional correlation image information content representation, and calculate the image texture characteristics corresponding to different spectral range parameters based on the imaging simulation dataset; construct a spectral distortion characteristic model, and calculate the spectral distortion characteristics corresponding to different spectral range parameters based on the imaging simulation dataset.

[0011] Based on the obtained entrance pupil radiation characteristics, imaging performance parameter calculation results, image texture characteristic parameter calculation results, and spectral distortion characteristic calculation results, the comprehensive impact of different spectral range parameters on payload imaging is evaluated.

[0012] Furthermore, the radiative transfer configuration parameters under imaging conditions include fixed configuration parameters and variable configuration parameters; the fixed configuration parameters include atmospheric mode, aerosol mode, visibility, illumination conditions, observation geometry, observed surface reflectance radiation characteristics, and observation path, while the variable configuration parameters are the spectral range parameters of the detection band.

[0013] Furthermore, an image texture information model based on bidirectional correlation image information content representation is constructed, specifically using the following method:

[0014] Generate gray-level co-occurrence matrices in different directions in the image, and perform the following processing for each single direction:

[0015] Based on the gray-level co-occurrence matrix of the stated direction, extract the positive related information parameter Information+ and the negative related information parameter Information-; perform a full-width row and column traversal of the image with the size of the gray-level co-occurrence matrix to generate several gray-level co-occurrence matrices, and average the information parameter results calculated from all co-occurrence matrices as the information content result of the stated direction of the image;

[0016] By combining the information from various directions, an image texture information model is obtained:

[0017]

[0018] In the formula, n is the number of co-occurrence matrices calculated by image traversal; angle represents the direction of the gray-level co-occurrence matrix, which is angle_1 to angle_num, for a total of num; and These are the positive and negative correlation data for all gray-level co-occurrence matrices at a certain angular direction.

[0019] Furthermore, based on the gray-level co-occurrence matrix in a single direction, the positively correlated information parameter Information+ and the negatively correlated information parameter Information- are extracted:

[0020]

[0021] In the formula, i and j represent the row and column information of an M×N gray-level co-occurrence matrix in the direction of image generation; GLCM_DN i,j is the element value at position (i, j) in the gray-level co-occurrence matrix.

[0022] Furthermore, an imaging simulation dataset was constructed under different spectral range parameter designs for the mapping payload. The dataset includes panchromatic simulated images, multispectral images, and fused images. The construction method is as follows:

[0023] Scenes with multiple types of ground features were selected as observation scenes for imaging simulation. These ground feature types included vegetation, land, water, buildings, and concrete roads.

[0024] The system collects airborne hyperspectral imagery data in the visible spectral range of the observed scene. The spatial resolution of the collected data is less than 0.3m, the spectral resolution is better than 2.5nm, and the spectral range covers at least 400nm to 1100nm.

[0025] Based on airborne hyperspectral imagery data, hyperspectral radiation datasets were obtained by inverting radiation calibration coefficients, and wideband radiation data in the panchromatic spectral range and wideband radiation data in the multispectral range were generated respectively.

[0026] The payload signal and noise information are constructed and simulated from the broadband radiation data. After the signal and noise information are superimposed, they are quantized and converted to obtain panchromatic simulated image data and multispectral image data of different spectral ranges after the imaging link information is transmitted.

[0027] Panchromatic analog image data with different band ranges is fused with multispectral image data to obtain fused image data.

[0028] Furthermore, a spectral distortion characteristic model is constructed, where spectral distortion characteristics are image spectral angular features. The construction method is as follows:

[0029] The data to be analyzed are the initial multispectral simulation data and the multispectral simulation data fused with panchromatic imagery, denoted as MUX. o and MUX f ;

[0030] MUX at the same pixel o and MUX f The gray values ​​of each band are used to form two vectors. The angle between the two vectors is calculated, and the arithmetic mean of the angles between the vectors obtained from all pixels is taken as the spectral angle of the image to be analyzed. That is:

[0031]

[0032] In the formula, θ spec The spectra angle of the image is specified; spec represents the corresponding spectral range setting; line and col are the number of rows and columns of the image data, and line·col equals the total number of pixels in the image.

[0033] Furthermore, the dynamic range and signal-to-noise ratio of the payload imaging system under different spectral range parameter designs are calculated using the following method:

[0034] The entrance pupil radiation energy is transmitted through the imaging of the optical remote sensor, outputting both effective signal and noise. The calculation model is as follows:

[0035]

[0036] In the formula, S band The number of electrons for the output effective signal; N band η is the number of output noise electrons; F is the reciprocal of the relative aperture of the load optical system; η is the vignetting coefficient of the load optical system; τ band L represents the average transmittance of the optical system in that band. band To analyze the apparent radiance of the entrance pupil under certain imaging conditions in the wavelength band, h is Planck's constant; c is the speed of light; λ center To analyze the center wavelength corresponding to the band; A pix Tint is the effective pixel area of ​​the detector; band For band imaging integration time, the number of stages (×TDI) is required for TDI-type detectors; QE band denoted as spectral responsivity in the band, also known as quantum efficiency; Gain is the imaging gain of the payload system; n SHOT n RMS_detector n ADC and n AMP These are photon noise, detector device noise, analog-to-digital conversion noise, and circuit amplification and filtering noise, respectively.

[0037] By calculating the signal and noise under different spectral range parameters, the system dynamic range (DR) and signal-to-noise ratio (SNR) results for different spectral ranges are obtained:

[0038]

[0039] In the formula, N is the number of quantization bits; Sat is the number of electrons output from the detector at full trap.

[0040] Furthermore, the dynamic MTF influence factor (MTF) of payload imaging under different spectral range parameter designs was calculated. band :

[0041] MTF band =MTF aero MTF aberr MTF diff

[0042] In the formula, MTF aero MTF (Medium-Terminal Fiber) in atmospheric aerosols aberr MTF is the optical aberration value. diff For optical diffraction MTF;

[0043] Using the panchromatic detection band as the analysis object and the design values ​​of parameters for different band ranges as variable elements, the MTF was analyzed. band Numerical calculations were performed to obtain the dynamic MTF influence factor results of the load system under parameter designs in different spectral ranges.

[0044] Further, calculate the radiation contrast and the proportion of atmospheric scattering:

[0045]

[0046] In the formula, Contrast t→b Contrast t→b+a `atmos_ratio` and `atmos_ratio` represent the target-to-background contrast, the combined target-to-background and atmospheric scattering contrast, and the proportion of atmospheric scattering in the apparent radiance of the entrance pupil, respectively; L target For the target radiance, L background For background radiance, L atmosphere For atmospheric path radiation, L apparent This represents the total radiant energy at the entrance pupil.

[0047] Furthermore, based on the obtained entrance pupil radiation characteristics, imaging performance parameter calculation results, image texture characteristic parameter calculation results, and spectral distortion characteristic calculation results, the comprehensive impact of different spectral range parameter designs on payload imaging is evaluated. The evaluation method is as follows:

[0048] The payload imaging performance of different spectral range parameters is evaluated. If the payload imaging performance results in the spectral range meet or exceed the requirements of the index system, the spectral range parameter design is deemed to meet the payload imaging requirements. Based on this, the entrance pupil radiation characteristics, image texture characteristics, and image spectral distortion characteristics in different spectral ranges are evaluated.

[0049] The larger the radiation contrast value and the smaller the atmospheric scattering ratio value, the better the spectral range parameter design.

[0050] If the image texture characteristics are richer and the target texture information is richer, then the spectral range parameter design will be better;

[0051] The smaller the spectral angle used to characterize the spectral distortion of an image, the better the design of the spectral range parameter.

[0052] The advantages of this invention compared to the prior art are:

[0053] (1) This invention uses hyperspectral data to simulate a wide-band dataset (single-band original image and fused image), constructs an adaptive information content evaluation model, conducts texture characteristic evaluation based on image quality evaluation criteria, and takes the mean of the calculation results in each direction as the evaluation quantity, which can suppress the influence of its directional components and more effectively reflect the changes in the information characteristics of image data with the design of the band range.

[0054] (2) The quantitative evaluation method for the influence of the band spectral range on image texture described in this invention can verify the consistency and accuracy of the evaluation method in analyzing the influence of spectral broadening changes on the image texture of a certain type of land cover from multiple perspectives by constructing a positive and negative bidirectional related information quantity model that characterizes the image texture features. Attached Figure Description

[0055] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0056] Figure 1 This is a flowchart illustrating the workflow for designing and optimizing the spectral range of the measurement payload detection band in an embodiment of the present invention. Detailed Implementation

[0057] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0058] To address the practical needs of designing and configuring the spectral range parameters of mapping and remote sensing payloads, evaluating and improving the quality of on-orbit image data, and maximizing application effectiveness, this invention proposes a design and evaluation method for the spectral range of mapping payload bands, comprehensively considering imaging radiometric performance and image texture information representation.

[0059] This method first calculates all radiation information of the payload's entrance pupil based on Earth-atmosphere coupled radiation calculations, and calculates radiation characteristics such as multi-parameter contrast. Combining payload design and engineering data, it models and calculates performance results under different spectral bands from multiple perspectives, including system dynamic range (DR), signal-to-noise ratio (SNR), and dynamic mean squared voltage (MTF). Second, it simulates wideband image datasets under different spectral ranges based on hyperspectral data, and then constructs an image texture information representation model to calculate image texture characteristics under different spectral bands. Simultaneously, it calculates spectral angle parameters between spectral bands to represent the degree of spectral distortion of the image under different spectral ranges. Finally, in the actual design scheme of the mapping payload, it conducts a quantitative evaluation of the above characteristics, taking into account the needs of remote sensing applications, satellite design requirements, and payload design constraints. Based on the evaluation results, it selects the optimal configuration scheme for the payload's detection spectral bands.

[0060] This method is as follows Figure 1 As shown, it specifically includes:

[0061] S1: Using the atmospheric radiative transfer model, calculate the entrance pupil radiation characteristics under different spectral range designs.

[0062] a) Establish the input parameter configuration for radiative transfer calculation under typical imaging conditions. The fixed reference configuration includes atmospheric model, aerosol model, visibility, illumination conditions, observation geometry, observed surface reflectance characteristics, and observation path. The variable configuration is the spectral range of the detection band.

[0063] b) Using the atmospheric radiative transfer model, calculate the radiative information at the entrance pupil of the optical mapping payload after atmospheric radiative transfer, including the target radiance L. target Background radiance L background Atmospheric path radiation L atmosphere Isoradiant energy components, total radiation energy L at the entrance pupil apparent apparent radiance, average atmospheric transmittance, and other radiation parameters.

[0064] c) Calculate parameters such as radiation contrast and atmospheric scattering ratio to characterize atmospheric absorption and scattering radiation performance.

[0065]

[0066] Among them, Contrast t→b Contrast t→b+a and atmos_ratio represent the target / background contrast, target / background + atmospheric scattering contrast, and the proportion of atmospheric scattering (i.e. atmospheric path radiation) in the apparent radiance of the entrance pupil, respectively.

[0067] S2: Combining the relevant design parameters and engineering data of the mapping payload, and based on the imaging information transmission model, calculate the imaging performance of the payload designed for different spectral ranges.

[0068] Key links and performance indicators for image quality evaluation of spectral range elements in associated bands in the optical remote sensing imaging link are extracted. Parameters characterizing the response range capability, noise suppression capability, and sharpness of the payload imaging quality—system dynamic range, signal-to-noise ratio, and dynamic MTF influence factor—are selected and calculated.

[0069] a) Calculation of system dynamic range and signal-to-noise ratio

[0070] The entrance pupil radiation energy is transmitted through the imaging of the optical remote sensor, outputting both effective signal and noise. The calculation model is as follows:

[0071]

[0072] Among them, S band The number of electrons for the output effective signal; N bandη is the number of output noise electrons; F is the reciprocal of the relative aperture of the load optical system; η is the vignetting coefficient of the load optical system; τ band L represents the average transmittance of the optical system in that band. band To analyze the apparent radiance of the entrance pupil under certain imaging conditions in the wavelength band; h is Planck's constant; c is the speed of light; λ center To analyze the center wavelength corresponding to the band; A pix Tint is the effective pixel area of ​​the detector; band For band imaging integration time, the number of stages (×TDI) is required for TDI-type detectors; QE band denoted as spectral responsivity in the band, also known as quantum efficiency; Gain is the imaging gain of the payload system; n SHOT n RMS_detector n ADC and n AMP These are photon noise, detector device noise, analog-to-digital conversion noise, and circuit amplification and filtering noise, respectively.

[0073] By calculating the signal and noise under different spectral range parameters, the system dynamic range (DR) and signal-to-noise ratio (SNR) under different spectral ranges can be calculated according to the following mathematical model.

[0074]

[0075] Where N is the number of quantization bits; Sat is the number of electrons output from the detector when it is full-well.

[0076] b) Dynamic MTF Influence Factor

[0077] Based on the optical imaging mechanism, MTF effect factors of correlated spectral range elements were extracted from atmospheric and mapping remote sensing payload systems, mainly including atmospheric aerosol MTF. aero Optical aberration MTF aberr and optical diffraction MTF diff .

[0078] Then, based on the characteristics of MTF cascade, a dynamic MTF influence factor (MTF) for optical loads affected by spectral range at the sampling frequency is established. band Model.

[0079] MTF band =MTF aero MTF aberr MTF diff

[0080] MTF aero For aerosol MTF, the calculation model is as follows:

[0081]

[0082] Where: R is the optical path length, A a and S a r represents the absorption and scattering coefficients of atmospheric aerosols and particles to radiation, respectively. aero denoted as .

[0083] MTF aberr The optical aberration MTF is calculated using the following model:

[0084]

[0085] Among them: W rms denoted as the root mean square of the wavefront difference, A is the wavefront difference calculation parameter, and p is the effective pixel size of the detection device.

[0086] MTF diff For optical diffraction MTF, the calculation model is as follows:

[0087]

[0088]

[0089] Using the panchromatic detection band as the main analysis object and design values ​​for different band ranges as variable factors, the MTF was analyzed. band Numerical calculations were performed to obtain the dynamic MTF influence factor results of the load system under different spectral range designs.

[0090] S3: Calculation of image texture characteristics under different spectral range parameters

[0091] a) Using airborne high-resolution hyperspectral data and design engineering data of the mapping payload, construct an imaging simulation dataset under different spectral ranges of the mapping payload design. The dataset includes panchromatic images, multispectral images, and fused images.

[0092] First, in order to better evaluate the image texture information, a scene with multiple types of ground objects was selected as the observation scene for imaging simulation. Airborne hyperspectral image data in the visible spectrum range was collected. The targets included vegetation, land, water, buildings, cement roads, and specific targets such as square bladed targets. The spatial resolution of the collected data reached <0.3m, the spectral resolution was better than 2.5nm, and the spectral range had to cover at least 400nm to 1100nm.

[0093] Then, the hyperspectral radiation dataset is obtained by inverting the radiation calibration coefficients.

[0094] {L_Hyper i} i＝1:n ={k i} i＝1:n ·{DN i} i＝1:n +{bi} i＝1:n

[0095] Where: L_Hyper is the hyperspectral radiance; k and b are the absolute radiometric calibration coefficients, respectively; DN is the digital output of the hyperspectral image; and the subscript i is the band number of the hyperspectral data (assuming that the hyperspectral data contains n bands, with the numbering from 1 to n).

[0096] By combining the spectral response data of the payload to be simulated with hyperspectral radiative data through convolution resampling calculations, broadband radiative data with different spectral ranges are generated. This example mainly simulates and generates panchromatic and multispectral band data. The panchromatic spectral range includes two types: one with a low end set at 450 nm and a high end varying from 650 nm to 900 nm, and the other with a low end set at 500 nm and a high end varying from 650 nm to 900 nm. The multispectral bands are set to the blue band (450–520 nm), green band (520–590 nm), red band (630–690 nm), and near-infrared band (770–890 nm), which are currently the mainstream wavelengths used by high-resolution optical payloads in orbit.

[0097]

[0098] Where: L_vis is the broadband spectral radiation data to be simulated; R is the spectral response function of the mapping payload at different wavelengths; λ is the wavelength; λ low and λ high These represent the upper and lower limits of the wide-band spectral range to be simulated.

[0099] Based on the signal and noise calculation model in step (2), the scene radiation data L band The payload signal and noise information are constructed and simulated. After the signal and noise information are superimposed, they are quantized and converted to obtain panchromatic and multispectral digital simulation image data after the imaging link information is transmitted.

[0100] Finally, based on the spatial resolution design requirements of different detection bands of the payload, it is generally necessary to downsample the image data generated in step d) to an image with the required spatial resolution, where the spatial resolution of multispectral data is 1 / 4 of that of panchromatic data.

[0101] Thus, we have obtained simulated image datasets and multispectral image data for different spectral ranges in the panchromatic band.

[0102] By fusion of panchromatic images and multispectral images from different band ranges, a fused dataset is obtained through simulation.

[0103] b) Construct an image texture information model to characterize texture properties using an analysis method that extracts texture features from the image gray-level co-occurrence matrix.

[0104] To balance computational speed and texture calculation accuracy, gray levels are appropriately compressed to a suitable bit depth. This example uses 8-bit computation to generate a gray-level co-occurrence matrix (GLCM) with a step size of 1 in four directions (0°, 45°, 90°, 135°) in the image. M×N|(0°，45°，90°，135°) .

[0105] Based on fundamental texture feature parameters for image classification applications, such as second-order angular moment, inverse difference moment, contrast, and gray-level correlation, an information content model characterizing image texture properties is constructed, denoted as TEXTURE_info. + and TEXTURE_info - , representing the positive and negative correlation directions describing the clarity of textures, respectively.

[0106] First, construct a gray-level co-occurrence matrix for a single direction (corresponding to any one of 0°, 45°, 90°, or 135°) to extract the positive information parameter Information. + Information related to negative information parameters - The representation function:

[0107]

[0108] Where: i and j represent the element positions (row and column information) in an M×N gray-level co-occurrence matrix in any direction of image generation; GLCM_DN is the element value at position (i, j) in the gray-level co-occurrence matrix.

[0109] Then, the image is traversed across all rows and columns in the form of a gray-level co-occurrence matrix (M×N) to generate several gray-level co-occurrence matrices. The information parameter results calculated from all co-occurrence matrices are averaged and used as the information content result in a single direction of the image. The numerical calculation results in four angular directions (0°, 45°, 90°, 135°) are then taken as the arithmetic mean and used as the texture characteristic information content (TEXTURE_info) of the ground feature sample image in the positive and negative correlation directions. + and TEXTURE_info - The final representation:

[0110]

[0111] Where: n is the number of co-occurrence matrices calculated by image traversal; angle represents the direction of the gray-level co-occurrence matrix (0°, 45°, 90°, 135°); and These are the average values ​​of the positive or negative correlation information of all gray-level co-occurrence matrices at a certain angular direction.

[0112] c) Using the texture information model constructed in the previous step, calculate the texture information content of simulated images in different spectral ranges.

[0113] To facilitate the analysis of the influence of different spectral ranges on the texture information of different land features, the simulated scene image dataset is cropped based on scene characteristics to generate a small area sample dataset with highly concentrated targets. The single sample dataset image mainly covers the same type of target land features.

[0114] For panchromatic datasets and fused image datasets of various types of samples, the information content of the simulated sample image datasets under different spectral ranges is calculated using the constructed texture information model, and the data results for the horizontal (data of the same sample in different bands) and vertical (data of different samples in the same band) are statistically analyzed respectively.

[0115]

[0116] Wherein: q samples are extracted from the scene, sample represents a sample, k represents the sample number; spec represents the designed spectral range, 1, 2, ..., s represent the sequence numbers of different spectral ranges, and there are s cases; Band represents the band, i represents the band type, including panchromatic bands and various bands in the fused image (such as true color RGB, false color NIRRG, etc.). The fused data mainly analyzed in this example is the true color RGB band image.

[0117] S4: Calculation of image spectral distortion characteristics under different spectral range parameters

[0118] The spectral angular features of the image are selected to describe the spectral distortion characteristics of the image.

[0119] The data to be analyzed are the initial multispectral simulation data and the multispectral simulation data fused with panchromatic imagery, denoted as MUX. o and MUX f MUX the multispectral data of the same pixel. o and MUX f The gray values ​​of each band are used to form two vectors. The angle between the two vectors is calculated, and the arithmetic mean of the angles between the vectors obtained from all pixels is taken as the spectral angle of the image to be analyzed.

[0120]

[0121] Where: θ spec This refers to the image's spectral angle; spec represents the corresponding spectral range setting; Gray MUXo_B1 Gray MUXo_B2 MUX is the initial simulation data for multispectral imaging. o Gray values ​​for each band; Gray MUXf_B1 Gray MUXf_B2The fused multispectral simulation data MUX f The grayscale value of each band; line and col are the number of rows and columns of the image data, and line·col equals the total number of pixels in the image.

[0122] Spectral angle calculations are performed on the original multispectral simulation data of the various land cover type samples generated in step S3 and the multispectral simulation data after fusion with panchromatic data of different spectral ranges.

[0123] S5: Based on the entrance pupil radiation characteristics, imaging performance parameters, image texture characteristics, and spectral distortion characteristics calculated in steps S1 to S4, and in conjunction with satellite platform requirements, mapping and remote sensing payload requirements, and remote sensing application needs, quantitatively evaluate the comprehensive impact of different spectral range designs on payload imaging, and provide an optimized configuration scheme.

[0124] a) First, based on the satellite platform and payload design requirements, evaluate the conformity between the calculated imaging performance results and the requirements of the indicator system.

[0125] Generally, the performance index requirements for image quality proposed by satellite platform and payload design are the highest priority key constraints. Based on the performance impact parameters calculated in the previous steps—imaging dynamic range, signal-to-noise ratio (SNR), and dynamic MTF impact factor—comparative analysis is conducted by comparing them with the performance requirement values ​​in the index system.

[0126] ■If the value meets or exceeds the requirements of the indicator system, the given spectral range design value will be included in other characteristic evaluation units;

[0127] ■ If the performance results are lower than the requirements of the indicator system, return to the scheme design of the spectral range, correct the scheme parameters, and recalculate various influencing characteristics until the imaging performance results meet the requirements of the indicator system.

[0128] b) Then, under the spectral range design parameters that meet the performance index system, the calculation results of other characteristics are evaluated.

[0129] 1) Assessment of entrance pupil radiation characteristics:

[0130] The evaluation criteria for target / background radiation energy contrast, target / (background + atmospheric scattering) radiation energy contrast, and atmospheric transmittance are positive optimization, that is, the larger the numerical result, the better the spectral range design parameters.

[0131] The evaluation criterion for the energy contrast between atmospheric path radiation and apparent radiation is negative optimization, that is, the smaller the numerical result, the better the spectral range design parameters.

[0132] 2) Image texture characteristic evaluation:

[0133] For a single land cover type sample, the positively correlated texture information quantity TEXTURE_info used to characterize texture properties + The evaluation criterion is positive optimization; the larger the numerical result, the richer the ground image information and the clearer the texture details under that spectral range parameter; negative correlation is TEXTURE_info. - The evaluation criterion is negative optimization; the smaller the numerical result, the better the texture characteristics.

[0134] The amount of image texture information varies among samples of different land cover types for the same spectral range design parameters. According to the user requirements of remote sensing applications, for key target types for observation or reconnaissance, spectral range parameters that are conducive to rich texture information of the land cover target should be selected first.

[0135] 3) Image spectral distortion characteristics assessment:

[0136] For a single land cover type sample, the spectral angle, which is used to characterize the spectral distortion characteristics, is also a negative optimization evaluation parameter. The smaller the value, the lower the degree of spectral distortion of the image, and the more preferred this spectral range parameter is.

[0137] Similarly, the spectral distortion characteristics of images with the same spectral range design parameters vary among samples of different land cover types. It is necessary to prioritize the selection of spectral range parameters with small spectral distortion corresponding to the key land cover types in the observation application, based on the user requirements of remote sensing applications.

[0138] c) Among the three characteristic evaluation results in the previous step, texture characteristics and spectral distortion characteristics are given the highest priority. Based on the specific application requirements and the overall satellite development requirements and constraints, the performance of each characteristic is taken into account, and the spectral range parameters of the detection band are optimized.

[0139] This method provides a quantitative assessment of band broadening design from multiple perspectives, including image quality and image information content. Based on this assessment, it designs and configures band spectral range parameters. This not only overcomes the shortcomings of previous methods that relied solely on visual interpretation to assess differences in image data with different spectral broadenings, but also offers richer quantitative evidence compared to traditional methods that design parameters based on inheritance and reference to similar products. This method quantitatively evaluates the differences in spectral broadening from the perspectives of image quality and texture representation in the final image product. It serves as a crucial pre-assessment reference for image products' target recognition and information extraction capabilities, providing an effective and practical method for optimizing the band spectral range design in surveying and remote sensing, as well as extending the verification process to quantitative applications.

[0140] The embodiments described above are merely preferred embodiments of the present invention. Ordinary variations and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of design evaluation for mapping a payload waveband spectral range, characterized by, The method comprises the following steps: Based on the radiation transmission configuration parameters under the imaging condition, the radiation information at the entrance pupil of the mapping load after the atmospheric radiation transmission of the detection waveband under the design of different spectral range parameters is calculated, and based on the obtained radiation information, the radiation contrast and the atmospheric scattering proportion corresponding to different spectral range parameters are calculated as the entrance pupil radiation characteristic results; The imaging performance of the load under the design of different spectral range parameters is calculated, including the system dynamic range, the signal-to-noise ratio and the dynamic MTF influence factor; An imaging simulation data set under the design of different spectral range parameters of the mapping load is constructed; an image texture information model based on the two-way correlation image information quantity representation is constructed, and the image texture characteristics corresponding to different spectral range parameters are calculated based on the imaging simulation data set; a spectral distortion characteristic model is constructed, and the spectral distortion characteristics corresponding to different spectral range parameters are calculated based on the imaging simulation data set; According to the obtained entrance pupil radiation characteristic results, the imaging performance parameter calculation results, the image texture characteristic parameter calculation results and the spectral distortion characteristic calculation results, the comprehensive influence of different spectral range parameters on the load imaging is evaluated.

2. The method of claim 1, wherein, The radiation transmission configuration parameters under the imaging condition include fixed configuration parameters and variable configuration parameters; the fixed configuration parameters include the atmospheric model, the aerosol model, the visibility, the illumination condition, the observation geometry, the observed ground reflected radiation characteristic and the observation path, and the variable configuration parameter is the spectral range parameter of the detection waveband.

3. The method of claim 1, wherein: The image texture information model based on the two-way correlation image information quantity representation is constructed, and the specific method is as follows: In the image, a gray level co-occurrence matrix in different directions is generated, and for each single direction, the following processing is performed: According to the gray level co-occurrence matrix of the direction, the positive correlation information parameter Information+ and the negative correlation information parameter Information- are extracted; The image is traversed in the full-width row and column directions according to the size of the gray level co-occurrence matrix, a plurality of gray level co-occurrence matrices are generated, and the mean values of the information parameter results calculated by all the co-occurrence matrices are taken as the information quantity results of the direction of the image; The image texture information model is obtained by synthesizing the information quantity results of each direction: In the formula, n is the number of the co-occurrence matrix of image traversal calculation; angle represents the direction of the gray co-occurrence matrix, and is angle_1-angle_num, a total of num; and respectively, are the positive correlation and negative correlation information results of all the gray co-occurrence matrices in a certain angle direction.

4. The method of claim 3, wherein: According to the gray level co-occurrence matrix of the single direction, the positive correlation information parameter Information+ and the negative correlation information parameter Information- are extracted: In the formula, i, j represent element row, column information in an MxN size gray level co-occurrence matrix of image generation; GLCM_DN i,j is the element value of position (i, j) in the gray level co-occurrence matrix.

5. The method of claim 1, wherein: An imaging simulation data set under the design of different spectral range parameters of the mapping load is constructed, and the data set includes a panchromatic simulation image, a multispectral image and a fused image, and the construction method is as follows: A scene with multiple types of ground objects is selected as the observation scene for imaging simulation, and the ground object types include vegetation, land, water area, building and cement road; Airborne hyperspectral image data in the visible spectral range is collected, the spatial resolution of the collected data is less than 0.3 m, the spectral resolution is better than 2.5 nm, and the spectral range covers at least 400 nm-1100 nm; Based on the airborne hyperspectral image data, the hyperspectral radiation data set is obtained by inversion using the radiation calibration coefficient, and the wide-band radiation data in the panchromatic wave spectral range and the wide-band radiation data in the multispectral wave range are generated respectively; The load signal and noise information are constructed and simulated based on the wide-band radiation data, and the signal and noise information are superimposed and then quantized and outputted to obtain the full-color simulation image data and the multi-spectral image data of different spectral ranges after the imaging link information transmission; The full-color simulation image data and the multi-spectral image data of different spectral ranges are fused to obtain the fused image data.

6. The method of claim 1, wherein: The spectral distortion characteristic model is constructed, the spectral distortion characteristic is an image spectral angle feature, and the construction method is as follows: The data to be analyzed is selected to be multispectral initial simulation data and multispectral simulation data after fusion with panchromatic image, denoted as MUX o and MUX f ; The MUX of the same pixel point o and the MUX f Each waveband gray value respectively constitutes two vectors, and the included angle between the two vectors is calculated. The arithmetic average of the included angles of all pixel vectors is taken as the spectrum angle of the image to be analyzed, that is: where θ is the image spectral angle spec is the image spectral angle; spec represents the corresponding spectral range setting; line and col are the row and column numbers of the image data, and line·col is equal to the number of all pixels in the image.

7. The method of claim 1, wherein: The load imaging system dynamic range and signal-to-noise ratio under different spectral range parameter designs are calculated, and the calculation method is as follows: The input pupil radiation energy is transmitted through the imaging of the optical remote sensor, and the output effective signal and noise are calculated, and the calculation model is as follows: In the formula, S band is the number of output effective signal electrons; N band is the number of output noise electrons; F is the reciprocal of the relative aperture of the load optical system; η is the vignetting coefficient of the load optical system; τ band is the average transmittance of the waveband optical system; L band is the apparent radiant luminance of the analysis waveband under certain imaging conditions; h is the Planck constant; c is the speed of light; λ center is the corresponding central wavelength of the analysis waveband; A pix is the effective pixel area of the detector device; Tint band is the waveband imaging integration time, and for a TDI type detector device, it needs ×TDI stages; QE band is the spectral response rate of the waveband, also known as quantum efficiency; Gain is the imaging gain of the load system; n SHOT , n RMS_detector , n ADC , and n AMP are respectively the photon noise, the detector device noise, the analog-to-digital conversion noise, and the circuit amplification and filtering noise. By calculating the signal and noise under different spectral range parameters, the system dynamic range DR and signal-to-noise ratio SNR results under different spectral ranges are calculated: In the formula, N is the quantization bit number; Sat is the number of output electrons of the full-well detector.

8. The method of claim 1, wherein: Computing the load imaging dynamic MTF impact factor MTF under different spectral range parameter designs band : MTF band = MTF aero • MTF aberr • MTF diff In the formula, MTF aero is the atmospheric aerosol MTF, MTF aberr is the optical aberration MTF, MTF diff is the optical diffraction MTF; taking the full-color detection waveband as the analysis object and taking different waveband range parameter design values as variable elements, the MTF band Numerical calculation is carried out to obtain the dynamic MTF influence factor results of the load system under different spectral range parameter designs.

9. The method of claim 1, wherein: The radiation contrast and the atmospheric scattering proportion are calculated: In the formula, Contrast t→b , Contrast t→b+a , and atmos_ratio are the target-to-background contrast, the target-to-background and atmospheric scattering combined contrast, and the proportion of atmospheric scattering in the entrance pupil apparent radiance, respectively. target L background L atmosphere L apparent L 10. The method of claim 1, wherein: According to the obtained input pupil radiation characteristic results, the imaging performance parameter calculation results, the image texture characteristic parameter calculation results and the spectral distortion characteristic calculation results, the comprehensive influence of different spectral range parameter designs on the load imaging is evaluated, and the evaluation method is as follows: The load imaging performance of different spectral range parameters is evaluated, if the load imaging performance results under the spectral range meet or are better than the required values of the index system, it is determined that the spectral range parameter design meets the load imaging requirements; on this basis, the input pupil radiation characteristics, the image texture characteristics and the image spectral distortion characteristics under different spectral ranges are evaluated: The greater the radiation contrast value and the smaller the atmospheric scattering proportion value, the better the spectral range parameter design; The richer the target texture information of the image texture characteristics, the better the spectral range parameter design; The smaller the spectral angle used to represent the image spectral distortion characteristics, the better the spectral range parameter design.

Images