Simulation method for yaw calibration data of flyback type multi-slit satellite-borne hyperspectral camera

Abstract

The invention relates to yaw calibration data simulation of a hyperspectral camera, and provides a simulation method for yaw calibration data of a flyback type multi-slit satellite-borne hyperspectral camera in order to solve the problems that an existing yaw calibration data simulation mode for the hyperspectral camera is relatively simple, the accuracy of obtained data is relatively low and the like. Comprising slit yaw trajectory simulation, yaw calibration data sampling and yaw data noise simulation. Wherein the slit yaw trajectory simulation comprises construction of a bottom image pixel coordinate system, construction of a camera coordinate system, conversion of a slit position and the bottom image pixel coordinate system and generation of a slit bottom image projection coordinate trajectory; the yaw calibration data sampling comprises yaw calibration data space dimension sampling and yaw calibration data spectrum dimension sampling; the yaw calibration data sampling comprises yaw calibration data space dimension sampling and yaw calibration data spectrum dimension sampling; according to the simulation method, the precision and accuracy of simulation data are effectively improved, and effective support is provided for yaw calibration algorithm research and calibration precision evaluation of the hyperspectral camera.

Description

Technical Field

[0001] This invention relates to the simulation of yaw calibration data from a hyperspectral camera, specifically to a simulation method for yaw calibration data from a retrace-type multi-slit spaceborne hyperspectral camera. Background Technology

[0002] Hyperspectral remote sensing is a technology that uses sensors mounted on satellites, aircraft, or drones to acquire information about the reflection or radiation of ground objects within tens to hundreds of consecutive narrow spectral bands. The core advantage of hyperspectral remote sensing lies in its extremely high spectral resolution (down to the nanometer level), enabling it to capture subtle spectral features of ground objects and is widely used in environmental monitoring, mineral exploration, and agricultural yield estimation. Conventional hyperspectral remote sensing cameras acquire ground object data from a single slit per frame, obtaining the spectral information of that slit object through a spectroscopic component. Based on the imaging mechanism of hyperspectral cameras, to ensure sufficient signal-to-noise ratio for each spectral band, spaceborne hyperspectral cameras typically employ a backscan compensation imaging mode. Furthermore, as... Figure 3 As shown, multi-slit imaging technology can also effectively improve the signal-to-noise ratio of hyperspectral camera images.

[0003] Yaw calibration is a commonly used relative radiometric correction technique for spaceborne remote sensing cameras, such as... Figure 2 As shown, pushbroom imaging is the conventional imaging method for spaceborne remote sensing cameras, while yaw calibration, by rotating the yaw direction of the spaceborne remote sensing camera by 90 degrees, can achieve the effect of each pixel of the camera detector imaging the same ground object, and thus the relative radiometric correction coefficient can be calculated. The advantage of yaw calibration is that it can generate relative radiometric correction coefficients of different brightness levels with fewer imaging attempts.

[0004] Traditional panchromatic or multispectral remote sensing cameras employ a passive pushbroom calibration mode for yaw calibration. Due to satellite orbit design and control capabilities, passive pushbroom calibration ensures that the camera's detector projection on the ground follows a straight line, resulting in simple yaw calibration data patterns. However, yaw calibration for scan-back hyperspectral remote sensing cameras cannot guarantee a straight-line projection of the camera slits on the ground due to the pitch angle during the scan process. Therefore, the acquired hyperspectral camera yaw calibration data is "unconventional." Furthermore, the multi-slit design means that the center of each slit is not at the center of the camera's optical axis, further complicating the yaw calibration data patterns of this type of hyperspectral camera.

[0005] In order to effectively analyze and study the yaw calibration data processing algorithm of the retrace-type multi-slit spaceborne hyperspectral camera, and to effectively evaluate the yaw calibration accuracy, it is necessary to conduct accurate simulation of the yaw calibration data and obtain real and effective simulation data. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing simulation methods for yaw calibration data of hyperspectral cameras, which are relatively simple and have low data accuracy. Therefore, this invention proposes a simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera.

[0007] The design concept of this invention is as follows: Based on the design principle of the retrace-type multi-slit spaceborne hyperspectral camera, the detector noise mode, the characteristics of spaceborne retrace imaging, and the distribution of objects at the calibration site, the invention effectively simulates the yaw calibration real-world imaging data of the hyperspectral camera using information such as satellite attitude and orbit parameters, camera pointing coordinates, and slit installation positions, thus providing effective support for the research of yaw calibration algorithms and the evaluation of calibration accuracy for this type of camera.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A simulation method for yaw calibration data from a backscan type multi-slit spaceborne hyperspectral camera, characterized by the following steps:

[0010] S1, Slit Yaw Trajectory Simulation

[0011] The slit yaw trajectory simulation includes constructing a base map pixel coordinate system, constructing a camera coordinate system, transforming the slit position to the base map pixel coordinate system, and generating the slit base map projection coordinate trajectory.

[0012] S1.1 Construct a base map pixel coordinate system on the hyperspectral base map data of the yaw calibration field based on the slit projection range of the hyperspectral camera;

[0013] S1.2 Construct a camera coordinate system and obtain the installation position of the hyperspectral camera slit in the camera coordinate system;

[0014] S1.3. Based on the installation position of the hyperspectral camera slit in the camera coordinate system and the satellite attitude and orbit parameters, calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for each single frame slit during the yaw calibration process.

[0015] S1.4 Calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for all frames during the yaw calibration process, forming a sequence of slit base map projection coordinate trajectories;

[0016] S2, Yaw Calibration Data Sampling

[0017] The yaw calibration data sampling includes spatial dimension sampling and spectral dimension sampling of yaw calibration data.

[0018] S2.1. Based on the pixel coordinates of the slit projection endpoints of each frame, perform yaw calibration data spatial dimension sampling for each spectral band;

[0019] S2.2 Based on the spectral response function of the hyperspectral camera, spectral sampling of the yaw calibration data is performed on each spectral band based on the spatial dimension sampling of the yaw calibration data to obtain the yaw imaging data of each frame.

[0020] S3, Yaw Data Noise Simulation

[0021] The yaw calibration data sampling includes spatial dimension sampling and spectral dimension sampling of yaw calibration data.

[0022] S3.1 Model the detector photon noise and detector electronic noise;

[0023] S3.2 Add detector photon noise and detector electronic noise to all frames of yaw imaging data to perform yaw calibration data noise simulation.

[0024] Furthermore, step S1.1 specifically includes:

[0025] The origin of the base image pixel coordinate system is set at the upper left corner of the hyperspectral base image. The X-axis of the base image pixel coordinate system points horizontally, the Y-axis points vertically, and the unit of the base image pixel coordinate system is pixels.

[0026] Furthermore, step S1.2 specifically includes:

[0027] The location of the intersection of the image plane where the hyperspectral camera slit is located and the optical axis is set as the origin of the camera coordinate system. The unit of the camera coordinate system is meters. The Xc axis of the camera coordinate system points to the direction of the satellite orbit flight, the Zc axis of the camera coordinate system points to the direction of the camera optical axis, and the Yc axis of the camera coordinate system is determined according to the right-hand rule.

[0028] Furthermore, step S1.3 specifically includes:

[0029] S1.3.1 Based on the installation position of the hyperspectral camera slit in the camera coordinate system and the satellite attitude and orbit parameters, calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for each single frame slit during the yaw calibration process.

[0030] The satellite attitude and orbit parameters include the satellite orbital altitude H, the pitch angle Pitch_i of a single frame during the yaw calibration retrace imaging process, and the longitude coordinates C_i_Lon and latitude coordinates C_i_Lat of the hyperspectral camera optical axis pointing position C_i of a single frame, i∈1~N, where N is the number of imaging frames in the yaw calibration process.

[0031] S1.3.2. Calculate the longitude conversion coefficient Lon2P and latitude conversion coefficient Lat2P between the four corner points in the base map pixel coordinate system, as follows:

[0032]

[0033] Where HSI_W and HSI_H represent the width of the hyperspectral base map data in the X-axis direction and the height in the Y-axis direction of the base map pixel coordinate system, respectively; P1_Lon, P2_Lon, P3_Lon, and P4_Lon represent the longitude coordinates of the four corner points; and P1_Lat, P2_Lat, P3_Lat, and P4_Lat represent the latitude coordinates of the four corner points.

[0034] S1.3.3 Calculate the base map pixel coordinates P_C_i(P_C_ix, P_C_iy) of the hyperspectral camera optical axis pointing position C_i, as shown in the following formula:

[0035] P1x = P1_Lon × Lon2P

[0036] P1y = P1_Lat × Lat2P

[0037] P_C_ix=C_i_Lon×Lon2P-P1x

[0038] P_C_iy=C_i_Lat×Lat2P-P1y

[0039] Where: (P1x, P1y) are the base map pixel coordinates of the top left corner P1 point of the hyperspectral base map data;

[0040] S1.3.4 Calculate the satellite orbit and satellite direction vector P_V_i in the base map pixel coordinate system of the current i-th frame, as follows:

[0041] P_V_i=P_C_(i+1)-P_C_i

[0042] Where: P_C_(i+1) is the base map pixel coordinate of the hyperspectral camera optical axis pointing position in the (i+1)th frame;

[0043] S1.3.5 Calculate the distance Lh_i from the origin of the camera coordinate system in the current i-th frame to the position C_i pointed to by the camera optical axis, as follows:

[0044]

[0045] S1.3.6 Calculate the projection distance P_Ds_i between the slit and the optical axis in the base image pixel coordinate system, as follows:

[0046]

[0047] Where Res is the spatial resolution of the base image pixel coordinate system; Ds is the distance between the slit and the optical axis; and F is the focal length of the hyperspectral camera.

[0048] S1.3.7 Calculate the normalized normal vector P_Vs_i and the projection midpoint P_Cs_i of the slit in the base image pixel coordinate system, as follows:

[0049]

[0050] P_Cs_i=P_C_i+P_Ds_i×P_Vs_i

[0051] Where: norm is the magnitude of the vector; P_V_ix is ​​the normal vector of the x-axis, and P_V_iy is the normal vector of the y-axis;

[0052] S1.3.8 Calculate the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints in the current i-th frame, as follows:

[0053]

[0054]

[0055] Where: P_Ls_i is the projection length of the slit in the i-th frame of the current image in the pixel coordinate system of the base image; Ls is the slit length.

[0056] Further, step S2.1 specifically includes:

[0057] S2.1.1 Based on the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints of each frame, sample and construct the yaw imaging data OneBand of each slit for each frame on the hyperspectral base map.

[0058] S2.1.2 Construct the projection coordinates P_List_i of the base map pixel coordinate system for all pixels corresponding to the slit in the current frame based on the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints, as follows:

[0059]

[0060] Where: Ns is the number of spatial dimension pixels of the hyperspectral camera, and P_List_i j Let be the projected coordinates of the j-th spatial pixel in the i-th frame's base image pixel coordinate system;

[0061] S2.1.3 Calculate the data value of each cell position in OneBand using the two-dimensional interpolation algorithm interp2, as shown in the following formula:

[0062] OneBand j =interp2(HSI,HSI_X,HSI_Y,P_List_ix j ,P_List_iy j j∈1~Ns

[0063] Where: HSI_X and HSI_Y are the X and Y coordinate matrices of all pixel locations in the hyperspectral base map data, and P_List_ix j and P_List_iy j For P_List_i j The X and Y axis coordinates;

[0064] S2.1.4. Based on the number of spectral segments B of the hyperspectral base map data, perform yaw calibration data spatial dimension sampling for each spectral segment according to the methods in S2.1.1 to S2.1.3.

[0065] Furthermore, the two-dimensional interpolation algorithm described in step S2.1.3 is a bilinear or bicubic spline interpolation method.

[0066] Furthermore, step S2.2 specifically includes:

[0067] The spatial dimension sampling data of yaw calibration data from all frames during the yaw calibration process are merged to obtain the original yaw imaging data AllBandsRaw for each slit. Based on the spectral response function (SRF) of the hyperspectral camera, the spectral dimension of the yaw calibration data is sampled for each spectral band to construct the simulated yaw imaging data AllBands, as shown in the following equation:

[0068] AllBands(z;y,x)=AllBandsRaw(1~B;y,x)*SRF(z)

[0069] z∈1~b

[0070] Where b is the number of spectral bands of the hyperspectral camera, y and x are the spatial coordinates of the yaw imaging data, and * represents convolution calculation.

[0071] Further, in step S3.1, the detector noise includes detector photon noise, detector electronic noise and dark level noise. Dark level noise in the hyperspectral base map data is removed, and detector photon noise and detector electronic noise are modeled.

[0072] Step S3.1 is as follows:

[0073] S3.1.1 The detector photon noise follows a Poisson distribution of the DN values ​​in the hyperspectral base map data, and the calculation formula is as follows:

[0074] N Photo (DN) = Possion(η × DN)

[0075] Where, N photo (DN) represents the detector photon noise; Possion represents the random noise generated based on the DN value of the input hyperspectral base map data, which follows a Poisson distribution; η represents the detector quantum efficiency.

[0076] S3.1.2 The detector's electronic noise follows a Gaussian distribution, and the calculation formula is as follows:

[0077] N Read =Gaussian(μ,σ)

[0078] Where, N Read δ represents the detector's electronic noise; Gaussian represents the generated random noise that follows a Gaussian distribution, μ is the expectation, and δ is the standard deviation.

[0079] Furthermore, step S3.2 specifically includes:

[0080] In the simulated yaw imaging data AllBands of each slit, a single frame data Frame is extracted, where the size of Frame is [Ns, b]. Detector photon noise and detector electronic noise are added to the Frame to obtain the yaw positioning simulation data AllBandsNoise for that frame, as shown in the following formula:

[0081] FrameNoise = Frame + N Photo (Frame)+N Read

[0082] The above method was used to obtain all frames of yaw positioning simulation data (AllBandsNoise) with detector photon noise and detector electronic noise, thus completing the yaw calibration data noise simulation.

[0083] The beneficial effects of this invention are:

[0084] [1] Compared with the traditional, simple-mode simulation of yaw calibration data of passive pushbroom panchromatic or multispectral spaceborne remote sensing cameras, the simulation method of the present invention for yaw calibration data of retrace-type multi-slit spaceborne hyperspectral cameras can effectively simulate yaw data of retrace-type, multi-slit complex modes of spaceborne hyperspectral cameras, effectively improving the accuracy and precision of the simulation data, and providing effective support for the research of yaw calibration algorithms and the evaluation of calibration accuracy of hyperspectral cameras.

[0085] [2] By comprehensively considering information such as satellite attitude and orbit parameters, hyperspectral camera pointing coordinates, and slit installation position, this invention can accurately simulate the projection imaging range of the slit on the ground object in each frame during the yaw calibration process of the hyperspectral camera, thereby ensuring the accuracy of the yaw calibration data simulation.

[0086] [3] Based on the parameters of the hyperspectral camera, the present invention adopts the spatial and spectral dimensions of the hyperspectral base map data, which can effectively simulate the real-scene yaw calibration data obtained under the actual on-orbit yaw imaging state of the hyperspectral camera.

[0087] [4] This invention models the noise of the hyperspectral camera detector and superimposes it on the yaw calibration simulation data. This can effectively simulate the nonlinear relative response inconsistency of the hyperspectral camera, and thus evaluate the accuracy of the relative radiometric correction based on the yaw data. Attached Figure Description

[0088] Figure 1 This is a flowchart illustrating a simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to the present invention.

[0089] Figure 2 In the background technology, pushbroom imaging is a conventional imaging method for spaceborne remote sensing cameras;

[0090] Figure 3 This refers to the imaging modes of multi-slit hyperspectral cameras in the background technology;

[0091] Figure 4 This is a schematic diagram illustrating the construction of the base image pixel coordinate system in an embodiment of the present invention;

[0092] Figure 5 This is a schematic diagram of the camera coordinate system construction in an embodiment of the present invention;

[0093] Figure 6 This is a schematic diagram illustrating the construction of pixel coordinates at the slit projection endpoints in an embodiment of the present invention.

[0094] Figure 7 This is a schematic diagram of the slit base map projection coordinate trajectory in an embodiment of the present invention;

[0095] Figure 8 This is a schematic diagram of spatial dimension sampling of yaw calibration data in an embodiment of the present invention;

[0096] Figure 9 This is a schematic diagram of spectral dimension sampling of yaw calibration data in an embodiment of the present invention;

[0097] Figure 10 This is a schematic diagram of the yaw calibration simulation data of the hyperspectral camera in an embodiment of the present invention;

[0098] (a) is a schematic diagram of the yaw calibration simulation data of the first slit hyperspectral camera;

[0099] (b) is a schematic diagram of the yaw calibration simulation data of the second slit hyperspectral camera;

[0100] Figure 11 This is a schematic diagram of the noise simulation results of yaw calibration data in an embodiment of the present invention;

[0101] (a) is a schematic diagram of the noise simulation results of the first slit yaw calibration data;

[0102] (b) is a schematic diagram of the noise simulation results of the second slit yaw calibration data; Detailed Implementation

[0103] like Figure 1 As shown, a simulation method for yaw calibration data from a retrace-type multi-slit spaceborne hyperspectral camera includes the following steps:

[0104] S1, Slit Yaw Trajectory Simulation

[0105] The slit yaw trajectory simulation includes constructing a base map pixel coordinate system, constructing a camera coordinate system, transforming the slit position to the base map pixel coordinate system, and generating the slit base map projection coordinate trajectory.

[0106] S1.1 Construct a base map pixel coordinate system on the hyperspectral base map data of the yaw calibration field based on the slit projection range of the hyperspectral camera;

[0107] like Figure 4 As shown, the origin of the base image pixel coordinate system is set at the upper left corner of the hyperspectral base image. The X-axis of the base image pixel coordinate system points horizontally, the Y-axis points vertically, and the unit of the base image pixel coordinate system is pixels.

[0108] The four corner points located in the hyperspectral base map data, namely the upper left, upper right, lower right, and lower left corners, are defined as P1, P2, P3, and P4, respectively. The longitude coordinates corresponding to the four corner points P1, P2, P3, and P4 are P1_Lon, P2_Lon, P3_Lon, and P4_Lon, respectively, and the latitude coordinates are P1_Lat, P2_Lat, P3_Lat, and P4_Lat, respectively.

[0109] S1.2 Construct a camera coordinate system, with the unit being meters, and obtain the installation position of the hyperspectral camera slit within the camera coordinate system;

[0110] like Figure 5 As shown, the position of the intersection of the image plane where the hyperspectral camera slit is located and the optical axis is set as the origin of the camera coordinate system. The unit of the camera coordinate system is meters. The Xc axis of the camera coordinate system points to the direction of the satellite orbit flight, the Zc axis of the camera coordinate system points to the direction of the camera optical axis, and the Yc axis of the camera coordinate system is determined according to the right-hand rule.

[0111] S1.3, such as Figure 6 As shown, based on the installation position of the hyperspectral camera slit in the camera coordinate system and the satellite attitude and orbit parameters, the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system of each single frame slit during the yaw calibration process are calculated.

[0112] S1.3.1 Based on the installation position of the hyperspectral camera slit in the camera coordinate system and the satellite attitude and orbit parameters, calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for each single frame slit during the yaw calibration process.

[0113] The satellite attitude and orbit parameters include the satellite orbital altitude H, in meters; the pitch angle Pitch_i of a single frame during the yaw calibration retrace imaging process, in degrees, i∈1~N; and the longitude coordinates C_i_Lon and latitude coordinates C_i_Lat of the hyperspectral camera optical axis pointing position C_i of a single frame, both in degrees, i∈1~N, where N is the number of imaging frames in the yaw calibration process.

[0114] S1.3.2. Calculate the longitude conversion coefficient Lon2P and latitude conversion coefficient Lat2P between the four corner points in the base map pixel coordinate system, as follows:

[0115]

[0116]

[0117] Where HSI_W and HSI_H represent the width of the hyperspectral base map data in the X-axis direction and the height in the Y-axis direction of the base map pixel coordinate system, respectively; P1_Lon, P2_Lon, P3_Lon, and P4_Lon represent the longitude coordinates of the four corner points; and P1_Lat, P2_Lat, P3_Lat, and P4_Lat represent the latitude coordinates of the four corner points.

[0118] S1.3.3 Calculate the base map pixel coordinates P_C_i(P_C_ix, P_C_iy) of the hyperspectral camera optical axis pointing position C_i, as shown in the following formula:

[0119] P1x = P1_Lon × Lon2P

[0120] P1y = P1_Lat × Lat2P

[0121] P_C_ix=C_i_Lon×Lon2P-P1x

[0122] P_C_iy=C_i_Lat×Lat2P-P1y

[0123] Where: (P1x, P1y) are the base map pixel coordinates of the top left corner P1 point of the hyperspectral base map data;

[0124] S1.3.4 Calculate the satellite orbit and satellite direction vector P_V_i in the base map pixel coordinate system of the current i-th frame, as follows:

[0125] P_V_i=P_C_(i+1)-P_C_i

[0126] Where: P_C_(i+1) is the base map pixel coordinate of the hyperspectral camera optical axis pointing position in the (i+1)th frame;

[0127] S1.3.5 Calculate the distance Lh_i from the origin of the camera coordinate system in the current i-th frame to the position C_i pointed to by the camera optical axis, as follows:

[0128]

[0129] S1.3.6 Calculate the projection distance P_Ds_i between the slit and the optical axis in the base image pixel coordinate system, as follows:

[0130]

[0131] Where Res is the spatial resolution of the base image pixel coordinate system; Ds is the distance between the slit and the optical axis; and F is the focal length of the hyperspectral camera.

[0132] S1.3.7 Calculate the normalized normal vector P_Vs_i and the projection midpoint P_Cs_i of the slit in the base image pixel coordinate system, as follows:

[0133]

[0134] P_Cs_i=P_C_i+P_Ds_i×P_Vs_i

[0135] Where: norm is the magnitude of the vector; P_V_ix is ​​the normal vector of the x-axis, and P_V_iy is the normal vector of the y-axis;

[0136] S1.3.8 Calculate the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints in the current i-th frame, as follows:

[0137]

[0138] Where: P_Ls_i is the projection length of the slit in the i-th frame of the current image in the pixel coordinate system of the base image; Ls is the slit length.

[0139] This allows us to obtain the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for each single-frame slit during the yaw calibration process.

[0140] S1.4 Calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for all frames during the yaw calibration process, forming a sequence of slit base map projection coordinate trajectories;

[0141] S2, Yaw Calibration Data Sampling

[0142] The yaw calibration data sampling includes spatial dimension sampling and spectral dimension sampling of yaw calibration data.

[0143] S2.1 Based on the pixel coordinates of the slit projection endpoints in each frame, sample and construct yaw imaging data of a single spectral band for each frame of each slit in the hyperspectral base map pixel coordinate system. Based on the number of spectral bands in the hyperspectral base map data, perform yaw calibration data spatial dimension sampling for each spectral band.

[0144] S2.1.1 Based on the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints of each frame, sample and construct the yaw imaging data OneBand of each slit for each frame on the hyperspectral base map.

[0145] S2.1.2, such as Figure 8 As shown, the projection coordinates P_List_i of the base map pixel coordinate system for all pixels corresponding to the slit in the current frame are constructed based on the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints, as follows:

[0146]

[0147] Where: Ns is the number of spatial dimension pixels of the hyperspectral camera, and P_List_i j Let be the projected coordinates of the j-th spatial pixel in the i-th frame's base image pixel coordinate system;

[0148] S2.1.3 Calculate the data value of each cell position in OneBand using the two-dimensional interpolation algorithm interp2, as shown in the following formula:

[0149] OneBand j =interp2(HSI,HSI_X,HSI_Y,P_List_ix j ,P_List_iy j j∈1~Ns

[0150] Where: HSI_X and HSI_Y are the X and Y coordinate matrices of all pixel locations in the hyperspectral base map data, and P_List_ix j and P_List_iy j For P_List_i j The X and Y axis coordinates;

[0151] Based on the number of spectral bands B in the hyperspectral base map data, yaw calibration data spatial dimension sampling is performed on each spectral band;

[0152] The two-dimensional interpolation algorithm is either bilinear or bicubic spline interpolation.

[0153] S2.2. Based on the spectral response function of the hyperspectral camera, perform spectral dimension sampling of the yaw calibration data for each spectral band according to the methods in S2.1.1 to S2.1.3.

[0154] The spatial dimension sampling data of yaw calibration data from all frames during the yaw calibration process are merged to obtain the original yaw imaging data AllBandsRaw for each slit. Based on the spectral response function (SRF) of the hyperspectral camera, the spectral dimension of the yaw calibration data is sampled for each spectral band to construct the simulated yaw imaging data AllBands, as shown in the following equation:

[0155] AllBands(z;y,x)=AllBandsRaw(1~B;y,x)*SRF(z)

[0156] z∈1~b

[0157] Where b is the number of spectral bands of the hyperspectral camera, y and x are the spatial coordinates of the yaw imaging data, * represents convolution calculation, and the spectral curve after yaw calibration spectral dimension sampling is shown in the figure. Figure 9 As shown.

[0158] S3, Yaw Data Noise Simulation

[0159] The yaw data noise simulation includes the construction of a detector nonlinear noise model and the simulation of yaw calibration data noise.

[0160] S3.1 Model the detector photon noise and detector electronic noise;

[0161] Detector noise includes detector photon noise, detector electronic noise, and dark level noise. Dark level noise is removed from the hyperspectral base map data, and detector photon noise and detector electronic noise are modeled and simulated.

[0162] S3.1.1 The detector photon noise follows a Poisson distribution of the DN values ​​in the hyperspectral base map data, and the calculation formula is as follows:

[0163] N Photo (DN) = Possion(η × DN)

[0164] Where, N photo (DN) represents the detector photon noise; Possion represents the random noise generated based on the DN value of the input hyperspectral base map data, which follows a Poisson distribution; η represents the detector quantum efficiency.

[0165] S3.1.2 The detector's electronic noise follows a Gaussian distribution, and the calculation formula is as follows:

[0166] N Read =Gaussian(μ,σ)

[0167] Where, N Read δ represents the detector's electronic noise; Gaussian represents the generated random noise that follows a Gaussian distribution, μ is the expectation, and δ is the standard deviation.

[0168] S3.2 Add detector photon noise and detector electronic noise to all frames of yaw imaging data to perform yaw calibration data noise simulation;

[0169] In the simulated yaw imaging data AllBands of each slit, a single frame data Frame is extracted, where the size of Frame is [Ns, b]. Detector photon noise and detector electronic noise are added to the Frame to obtain the yaw positioning simulation data AllBandsNoise for that frame, as shown in the following formula:

[0170] FrameNoise = Frame + N Photo (Frame)+N Read

[0171] The above method was used to obtain all frames of yaw positioning simulation data (AllBandsNoise) with detector photon noise and detector electronic noise, thus completing the yaw calibration data noise simulation.

[0172] In this embodiment, the yaw calibration data of a given retrace-type spaceborne hyperspectral camera are simulated through the above simulation steps. The hyperspectral camera adopts a double-slit grating dispersive spectrometer system, and its spectral range covers visible light to near-infrared.

[0173] S1, Slit Yaw Trajectory Simulation

[0174] The hyperspectral base map data HSI uses the publicly available Chikusei hyperspectral dataset, with a width HSI_W of 2335 pixels, a height HSI_H of 2517 pixels, a spectral band number B of 128, and a spatial resolution Res of 2.5 meters. The latitude and longitude coordinates of the four corner points P1, P2, P3, and P4 in the hyperspectral base map data are (36.3230°, 139.9755°), (36.3236°, 140.0405°), (36.2668°, 140.0412°), and (36.2663°, 139.9762°), respectively.

[0175] The parameters of the hyperspectral camera are as follows: slit length Ls = 0.006 m, slit distance Ds = 0.006 m, focal length F = 6 m.

[0176] The yaw calibration process is set to perform a total of N frames of imaging, where N is 1079 frames of imaging.

[0177] The satellite orbital altitude H is 500,000m, and the pitch angle changes from -3° to +3° during the retrace process.

[0178] The slit base map projection coordinate trajectory calculated according to the method in step S1.3 is as follows: Figure 7 As shown.

[0179] S2, Yaw Calibration Data Sampling

[0180] The hyperspectral camera has 200 spatial pixels (Ns) and 16 spectral bands.

[0181] According to the calculation method in step S2, the hyperspectral camera yaw calibration simulation data obtained after sampling the spatial and spectral dimensions of the yaw calibration data is as follows: Figure 10 As shown.

[0182] S3, Yaw Data Noise Simulation

[0183] The detector's quantum efficiency η is 0.7, and the detector's electronic noise follows a Gaussian distribution with an expected value μ of 0 and a standard deviation δ of 100.

[0184] Based on the calculation method in step S3, the simulation results of yaw calibration data noise are as follows: Figure 11 As shown.

Claims

1. A simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera, characterized in that, Includes the following steps: S1, Slit Yaw Trajectory Simulation S1.1 Construct a base map pixel coordinate system on the hyperspectral base map data of the yaw calibration field based on the slit projection range of the hyperspectral camera; S1.2 Construct a camera coordinate system and obtain the installation position of the hyperspectral camera slit in the camera coordinate system; S1.

3. Based on the installation position of the hyperspectral camera slit in the camera coordinate system and the satellite attitude and orbit parameters, calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for each single frame slit during the yaw calibration process. S1.4 Calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for all frames during the yaw calibration process, forming a sequence of slit base map projection coordinate trajectories; S2, Yaw Calibration Data Sampling S2.

1. Based on the pixel coordinates of the slit projection endpoints of each frame, perform yaw calibration data spatial dimension sampling for each spectral band; S2.2 Based on the spectral response function of the hyperspectral camera, spectral sampling of the yaw calibration data is performed on each spectral band based on the spatial dimension sampling of the yaw calibration data to obtain the yaw imaging data of each frame. S3, Yaw Data Noise Simulation S3.1 Model the detector photon noise and detector electronic noise; S3.2 Add detector photon noise and detector electronic noise to all frames of yaw imaging data to perform yaw calibration data noise simulation.

2. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 1, characterized in that, Step S1.1 is as follows: The origin of the base image pixel coordinate system is set at the upper left corner of the hyperspectral base image. The X-axis of the base image pixel coordinate system points horizontally, the Y-axis points vertically, and the unit of the base image pixel coordinate system is pixels.

3. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 2, characterized in that, Step S1.2 specifically includes: The location of the intersection of the image plane where the hyperspectral camera slit is located and the optical axis is set as the origin of the camera coordinate system. The unit of the camera coordinate system is meters. The Xc axis of the camera coordinate system points to the direction of the satellite orbit flight, the Zc axis of the camera coordinate system points to the direction of the camera optical axis, and the Yc axis of the camera coordinate system is determined according to the right-hand rule.

4. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 3, characterized in that, Step S1.3 specifically includes: S1.3.1 Based on the installation position of the hyperspectral camera slit in the camera coordinate system and the satellite attitude and orbit parameters, calculate the pixel coordinates of the slit projection endpoints in the base map pixel coordinate system for each single frame slit during the yaw calibration process. The satellite attitude and orbit parameters include the satellite orbital altitude H, the pitch angle Pitch_i of a single frame during the yaw calibration retrace imaging process, and the longitude coordinates C_i_Lon and latitude coordinates C_i_Lat of the hyperspectral camera optical axis pointing position C_i of a single frame, i∈1~N, where N is the number of imaging frames in the yaw calibration process. S1.3.

2. Calculate the longitude conversion coefficient Lon2P and latitude conversion coefficient Lat2P between the four corner points in the base map pixel coordinate system, as follows: Where HSI_W and HSI_H represent the width of the hyperspectral base map data in the X-axis direction and the height in the Y-axis direction of the base map pixel coordinate system, respectively; P1_Lon, P2_Lon, P3_Lon, and P4_Lon represent the longitude coordinates of the four corner points; and P1_Lat, P2_Lat, P3_Lat, and P4_Lat represent the latitude coordinates of the four corner points. S1.3.3 Calculate the base map pixel coordinates P_C_i(P_C_ix, P_C_iy) of the hyperspectral camera optical axis pointing position C_i, as shown in the following formula: P1x = P1_Lon × Lon2P P1y = P1_Lat × Lat2P P_C_ix=C_i_Lon×Lon2P-P1x P_C_iy=C_i_Lat×Lat2P-P1y Where: (P1x, P1y) are the base map pixel coordinates of the top left corner P1 point of the hyperspectral base map data; S1.3.4 Calculate the satellite orbit and satellite direction vector P_V_i in the base image pixel coordinate system of the current i-th frame, as follows: P_V_i=P_C_(i+1)-P_C_i Where: P_C_(i+1) is the base map pixel coordinate of the hyperspectral camera optical axis pointing position in the (i+1)th frame; S1.3.5 Calculate the distance Lh_i from the origin of the camera coordinate system in the current i-th frame to the position C_i pointed to by the camera optical axis, as follows: S1.3.6 Calculate the projection distance P_Ds_i between the slit and the optical axis in the base image pixel coordinate system, as follows: Where Res is the spatial resolution of the base image pixel coordinate system; Ds is the distance between the slit and the optical axis; and F is the focal length of the hyperspectral camera. S1.3.7 Calculate the normalized normal vector P_Vs_i and the projection midpoint P_Cs_i of the slit in the base image pixel coordinate system, as follows: P_Cs_i=P_C_i+P_Ds_i×P_Vs_i Where: norm is the magnitude of the vector; P_V_ix is ​​the normal vector of the x-axis, and P_V_iy is the normal vector of the y-axis; S1.3.8 Calculate the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints in the current i-th frame, as follows: Where: P_Ls_i is the projection length of the slit in the i-th frame of the current image in the pixel coordinate system of the base image; Ls is the slit length.

5. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 4, characterized in that, Step S2.1 is as follows: S2.1.1 Based on the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints in each frame, sample and construct the OneBand yaw imaging data of a single spectral band for each frame of each slit on the hyperspectral base map. S2.1.2 Construct the projection coordinates P_List_i of the base map pixel coordinate system for all pixels corresponding to the slit in the current frame based on the pixel coordinates P_S1_i and P_S2_i of the slit projection endpoints, as follows: Where: Ns is the number of spatial dimension pixels of the hyperspectral camera, and P_List_i j Let be the projected coordinates of the j-th spatial pixel in the i-th frame's base image pixel coordinate system; S2.1.3 Calculate the data value of each cell position in OneBand using the two-dimensional interpolation algorithm interp2, as shown in the following formula: OneBand j =interp2(HSI,HSI_X,HSI_Y,P_List_ix j ,P_List_iy j ) j∈1~Ns Where: HSI_X and HSI_Y are the X and Y coordinate matrices of all pixel locations in the hyperspectral base map data, and P_List_ix j and P_List_iy j For P_List_i j The X and Y axis coordinates; S2.1.

4. Based on the number of spectral segments B of the hyperspectral base map data, perform yaw calibration data spatial dimension sampling for each spectral segment according to the methods in S2.1.1 to S2.1.

3.

6. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 5, characterized in that: The two-dimensional interpolation algorithm described in step S2.1.3 is a bilinear or bicubic spline interpolation method.

7. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 6, characterized in that, Step S2.2 specifically includes: The spatial dimension sampling data of yaw calibration data from all frames during the yaw calibration process are merged to obtain the original yaw imaging data AllBandsRaw for each slit. Based on the spectral response function (SRF) of the hyperspectral camera, the spectral dimension of the yaw calibration data is sampled for each spectral band to construct the simulated yaw imaging data AllBands, as shown in the following equation: AllBands(z;y,x)=AllBandsRaw(1~B;y,x)*SRF(z) z∈1~b Where b is the number of spectral bands of the hyperspectral camera, y and x are the spatial coordinates of the yaw imaging data, and * represents convolution calculation.

8. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 7, characterized in that: In step S3.1, the detector noise includes detector photon noise, detector electronic noise, and dark level noise. Dark level noise in the hyperspectral base map data is removed, and detector photon noise and detector electronic noise are modeled. Step S3.1 specifically includes: S3.1.1 The detector photon noise follows a Poisson distribution of the DN values ​​in the hyperspectral base map data, and the calculation formula is as follows: N Photo (DN)=Possion(η×DN) Where, N photo (DN) represents the detector photon noise; Possion represents the random noise generated based on the DN value of the input hyperspectral base map data, which follows a Poisson distribution; η represents the detector quantum efficiency. S3.1.2 The detector's electronic noise follows a Gaussian distribution, and the calculation formula is as follows: N Read =Gaussian(μ,σ) Where, N Read denoted as detector electronic noise; Gaussian represents the generated random noise that follows a Gaussian distribution; μ is the expectation; and δ is the standard deviation.

9. The simulation method for yaw calibration data of a retrace-type multi-slit spaceborne hyperspectral camera according to claim 8, characterized in that, Step S3.2 specifically includes: In the simulated yaw imaging data AllBands of each slit, a single frame data Frame is extracted, where the size of Frame is [Ns, b]. Detector photon noise and detector electronic noise are added to the Frame to obtain the yaw positioning simulation data AllBandsNoise for that frame, as shown in the following formula: FrameNoise=Frame+N Photo (Frame)+N Read The above method was used to obtain all frames of yaw positioning simulation data (AllBandsNoise) with detector photon noise and detector electronic noise, thus completing the yaw calibration data noise simulation.

Images