A rapid radiometric calibration method based on multi-wavelength and multi-mode in tropical rainforests

Abstract

The present invention discloses a multi-wavelength, multi-mode rapid radiation calibration method based on tropical rainforests, belonging to the technical field of radiation calibration. The present invention only collects three echo data to obtain calibration constants for all time-width and bandwidth combinations. By quantizing observable parameters and eliminating their influence on different wavelengths and modes, the obtained calibration constants and noise coefficients can represent correction parameters for all wavelengths, thereby reducing data acquisition time, acquisition cycle, and data analysis volume, and improving calibration timeliness. The radiation calibration achieves high accuracy, with radiation accuracy exceeding 0.43dB for the same wavelength and exceeding 0.69dB for generalized wavelengths.

Description

Technical Field

[0001] The present invention relates to the technical field of radiation calibration, and in particular to a multi-wavelength and multi-mode rapid radiation calibration method based on tropical rain forests. Background Art

[0002] Spaceborne synthetic aperture radar (SAR) satellites offer all-day, all-weather observation capabilities and hold broad application prospects in emergency photography and military applications. Traditional large satellites are well-calibrated for radiometric measurements. However, small, lightweight commercial SAR satellites, due to their small size and the relatively poor accuracy of their attitude control and positioning equipment, generally lack high-precision radiometric calibration.

[0003] External calibration techniques require a ground-based standard calibrator. Based on the calibrator's characteristics, they can be categorized into three main types: active calibrators, passive angle reflectors, and natural scene (tropical rainforest) The first two require on-site coordination, requiring rented sites and on-site testing, which is costly. Using a natural scene eliminates the need for on-site coordination and is therefore less expensive.

[0004] Existing external calibration techniques typically only test a small number of wavebands, resulting in large computational overhead, poor generalization, insufficient theoretical foundation for generalization, and low accuracy. To address this issue, a multi-waveband, multi-mode rapid radiometric calibration method based on tropical rainforests is proposed. Summary of the Invention

[0005] The technical problem to be solved by the present invention is: how to solve the problems of existing external calibration technology, which usually only tests a small number of wave positions, has the problems of large amount of calculation of collected data, weak generalization ability, insufficient theoretical basis for promotion, and low accuracy. A multi-wavelength and multi-mode rapid radiation calibration method based on tropical rainforest is provided, which uses a small amount of wave position data analysis to solve the problem of obtaining calibration constants for all wave positions in modes such as strip, beam, TOPS, etc., reduces the workload of data collection and analysis, and meets the needs of rapid calibration.

[0006] The present invention solves the above technical problems through the following technical solutions, which include the following steps:

[0007] S1: Develop a calibration plan

[0008] The calibration satellite was used to collect echo data of the 4-band TOPS calibration field at the same incident angle three times. The calibration field was the Amazon rainforest.

[0009] S2: Imaging Processing

[0010] The echo data collected by the calibration satellite is imaged using the omiga-K algorithm;

[0011] S3: Parameter calculation and acquisition

[0012] Calculate the slant range, incident angle, and transmit power during echo data acquisition, and obtain radar receiving gain control parameters, imaging processing bandwidth, wave position number, and antenna pattern;

[0013] S4: Data screening

[0014] The images obtained by imaging processing are grouped according to wave positions, and the data validity is analyzed, that is, the imaging results are screened according to the screening conditions and the imaging results that do not meet the screening conditions are eliminated;

[0015] S5: Data Correction

[0016] Calculate various influencing factors and use all influencing factors to pre-process the screened imaging results to obtain corrected amplitude data, where the imaging results are single-view complex data;

[0017] S6: Get calibration parameters

[0018] The corrected amplitude data is subjected to linear regression analysis with the Beta data corrected by the Sentinel-1 SAR satellite radiation, and the calibration constant K and noise coefficient P are output. n ;

[0019] S7: Extrapolation of calibration parameters at different wave positions

[0020] The calibration constant K and noise factor P n Extrapolation of other wave position images applied in TOPS mode;

[0021] S8: Extrapolation of calibration parameters for different modes

[0022] Calculate the imaging processing influence factors in strip and beam modes, replace the imaging processing influence factors in TOPS mode, and make the calibration constant K and noise coefficient P output in S6 n Able to extrapolate wave position images applied in strip and spotlight modes.

[0023] Furthermore, in step S3, the parameters are specifically calculated as follows:

[0024] The slope distance calculation method is as follows:

[0025] First calculate the closest slope distance:

[0026]

[0027] The closest slope distance R near It is calculated from the accurate time delay t and light speed c obtained by the calibration within the payload slant range;

[0028] Then, based on the range resolution, calculate the slant distance of each pixel:

[0029] R index =R near +ΔR*index

[0030] Among them, ΔR represents the range resolution of the SAR image, R index is the slope distance of the index-th pixel;

[0031] The angle of incidence is calculated as follows:

[0032]

[0033] in, For radar perspective, N is the wave position number, is the satellite side swing angle, 0.01 represents the angular interval between wave positions, R e is the radius of the earth, h is the altitude of the satellite;

[0034] The formula for calculating the transmit power is as follows:

[0035] P v =P t *prf*τ

[0036] Among them, P t is the peak power, τ is the pulse duration, and prf is the pulse repetition frequency.

[0037] Furthermore, in step S3, the parameters are specifically obtained as follows:

[0038] Radar receiving gain control parameters: directly obtained from radar imaging parameters and expressed as MGC; imaging processing bandwidth, wave position number, and antenna radiation pattern are known parameters.

[0039] Furthermore, in step S5, the specific process is as follows:

[0040] S51: Convert complex data into amplitude data:

[0041]

[0042] Wherein, b1 represents the real part of the complex image; b2 represents the imaginary part of the complex image;

[0043] S52: Calculate the radiation power impact factor:

[0044]

[0045] S53: Calculate the MGC impact factor:

[0046]

[0047] S54: Calculate the antenna pattern impact factor:

[0048]

[0049] Among them, R cent is the center slant distance, G() represents the antenna pattern;

[0050] S55: Computational imaging processing impact factors:

[0051]

[0052] Among them, f s Indicates the sampling bandwidth, N burst Indicates the number of pulses per burst

[0053] S56: Calculate the resolution impact factor:

[0054]

[0055] Wherein, ΔR represents the range resolution of the SAR image; ΔA represents the azimuth resolution of the SAR image;

[0056] S57: Calculate the slope distance influence factor:

[0057]

[0058] S58: Based on the above influencing factors, the corrected amplitude data, i.e. DN value, is obtained. The calculation formula is as follows:

[0059]

[0060] Furthermore, in step S6, the calibration constant K and the noise coefficient P are output. n The specific process is as follows:

[0061] S61: collecting pixel areas of different brightness in the same area of ​​the calibration image and the reference image, and calculating the mean of the corrected amplitude data in the calibration image and the mean of the Beta data in the reference image, where the calibration image is the corrected amplitude data and the reference image is the Beta image of the Sentinel-1 SAR satellite after radiometric correction;

[0062] S62: Use the least squares method to perform linear fitting. The input X-axis data is the data collected by the calibration satellite, the corrected amplitude data in the calibration image, and the Y-axis data is the Beta data in the reference image.

[0063] S63: The intersection of the fitting line and the Y axis is used as the noise coefficient Pn, and the slope of the fitting line is used as the calibration constant K.

[0064] Furthermore, in step S8, the calculation formula of the imaging processing influence factor in the stripe mode is as follows:

[0065]

[0066] Among them, L a Indicates the antenna azimuth length, V st represents the satellite speed, τ represents the pulse width, λ represents the central wavelength, R represents the slant range, prf represents the pulse repetition frequency, and f s Indicates the sampling bandwidth.

[0067] Furthermore, in step S8, the calculation formula of the imaging processing influence factor in the beamforming mode is as follows:

[0068]

[0069] Where τ represents the pulse width, f s represents the sampling bandwidth, and 32768 is the number of pulses accumulated in the azimuth direction.

[0070] Compared with the existing technology, the present invention has the following advantages: the multi-wavelength and multi-mode rapid radiation calibration method based on tropical rainforests can obtain calibration constants for all time-width and bandwidth combinations by collecting echo data only three times; by quantizing the observable parameters, their influence on different wavelengths and modes is eliminated, so that the obtained calibration constants and noise coefficients can represent the correction parameters of all wavelengths, thereby reducing data collection time, collection cycle, data analysis volume, and improving calibration timeliness; cross-validation with a satellite in the same frequency band (Sentinel-1 SAR satellite) is adopted to reduce the difficulty of obtaining the true value; the radiation calibration accuracy is high, the radiation accuracy at the same wavelength is better than 0.43dB, and the radiation calibration accuracy at the extended wavelength is better than 0.69dB. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 This is a flow chart of the multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rain forests. ， ” refers to DN ‘2 ;

[0072] Figure 2 : This is a diagram of radiation calibration errors at different wave positions in an embodiment of the present invention, where the X-axis represents the viewing angle*10, the Y-axis represents the error, and the blue dots represent error data points. DETAILED DESCRIPTION

[0073] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. However, the protection scope of the present invention is not limited to the following embodiment.

[0074] like Figure 1 As shown, this embodiment provides a technical solution: a multi-wavelength and multi-mode rapid radiometric calibration method based on a tropical rainforest, comprising the following steps:

[0075] Step 1: Develop a calibration plan

[0076] The calibration satellite collects echo data three times in a four-band TOPS calibration field at the same angle of incidence. The calibration field is the Amazon rainforest. The mission planning module calculates the acquisition time based on the Amazon rainforest location, satellite orbit, and payload performance parameters.

[0077] Step 2: Get the true sigma value of the calibration field

[0078] Download the IW mode SLC product data acquired by the Sentinel-1 satellite over the Amazon rainforest region during the same period (before and after one month), perform radiometric correction processing using the SNAP software, and output the true sigma value of the Amazon rainforest region as the input for step ten.

[0079] Step 3: Imaging Processing

[0080] The echo data collected by the calibration satellite are imaged using the omiga-K algorithm.

[0081] Step 4: Parameter calculation and acquisition

[0082] Calculate the slant range, incident angle, and average transmit power during echo data acquisition, and obtain radar receive gain control parameters, imaging processing bandwidth, wave position number, and antenna pattern;

[0083] The slope distance calculation method is as follows:

[0084] First calculate the closest slope distance:

[0085]

[0086] The closest slope distance R near It is calculated from the accurate time delay t and light speed c obtained by the calibration within the payload slant range;

[0087] Then, based on the range resolution, calculate the slant distance of each pixel:

[0088] R index =R near +ΔR*index

[0089] Among them, ΔR represents the range resolution of the SAR image, R index is the slant distance of the index-th pixel (calculated from the closest distance);

[0090] The angle of incidence is calculated as follows:

[0091] According to the wave position number N, satellite side swing angle Calculate radar viewing angle using other parameters

[0092]

[0093] in, is the satellite side swing angle (±26 degrees), N∈(-200,200), and 0.01 represents the angular interval between wave positions.

[0094] Then according to the satellite height h, the earth radius R e The same parameters are used to calculate the incident angle θ using the triangle sine theorem:

[0095]

[0096] Because left-sided vision is a negative number, so when calculating the incident angle θ, Take the absolute value.

[0097] The method for obtaining MGC is as follows:

[0098] Radar receiver gain control parameters are directly obtained from radar imaging parameters and are represented by MGC.

[0099] The formula for calculating the average transmit power is as follows:

[0100] P v =P t *prf*τ

[0101] According to the peak power P t , pulse time width τ, pulse repetition frequency prf and other parameters to calculate the average transmission power.

[0102] The imaging processing bandwidth, wave position number, and antenna pattern are known parameters.

[0103] Step 5: Data screening

[0104] The images obtained by imaging processing are grouped according to wave positions, and the data validity is analyzed, that is, the imaging results are screened according to the screening conditions and the imaging results that do not meet the screening conditions are eliminated;

[0105] The screening conditions are as follows:

[0106] (1) Echo: loss rate 0%;

[0107] (2) Shooting parameter consistency detection: MGC value, time width (duration of pulse transmission), bandwidth (frequency range of linear frequency modulation of pulse) are within the valid range, and attitude stability (roll angle and pitch angle remain unchanged);

[0108] (3) The imaging area is within the Amazon calibration area;

[0109] (4) The verification data are consistent with the collection data date: the difference is less than 1 month;

[0110] (5) The data collection weather was sunny;

[0111] (6) The difference between the data-based and attitude-based Doppler estimates is less than 5 Hz;

[0112] (7) Image equivalent noise coefficient is better than -20dB;

[0113] Step 6: Data Correction

[0114] Calculate various influencing factors (correction factors) and use all influencing factors to preprocess the screened imaging results (single-view complex data SLC) to obtain the corrected amplitude data;

[0115] The specific processing process is as follows:

[0116] 1. Convert complex data into amplitude data:

[0117]

[0118] Wherein, b1 represents the real part of the complex image; b2 represents the imaginary part of the complex image;

[0119] 2. Calculate the radiation power impact factor (used to eliminate the radiation power impact in radiation calibration):

[0120]

[0121] 3. Calculate the MGC impact factor (used to eliminate the MGC impact in radiation calibration):

[0122]

[0123] 4. Antenna pattern correction

[0124] Extract the antenna pattern corresponding to the imaging wave position from the database, extract the correction value corresponding to the 3dB angle of the SAR payload; calculate the center slant range R cent , the center of the image corresponds to the highest point of the directional pattern; the correction value corresponding to different slant distances is calculated according to the following formula:

[0125]

[0126] Where R represents the slant range of the current pixel and G() represents the antenna pattern.

[0127] 5. Calculate the imaging processing influence factor (used to eliminate the influence of imaging processing in radiometric calibration):

[0128] The imaging processing mainly involves the accumulation of pulse pressure in the range direction and correlation in the azimuth direction. Under different parameters such as time width, satellite speed and slant range, the accumulated energy varies greatly, so this part needs to be eliminated in the radiation calibration.

[0129]

[0130] Where τ represents the pulse width, f s Indicates the sampling bandwidth, N burst Indicates the number of pulses per burst.

[0131] 6. Calculate the resolution impact factor (used to eliminate the impact of resolution in radiometric calibration):

[0132]

[0133] Wherein, ΔR represents the range resolution of the SAR image; ΔA represents the azimuth resolution of the SAR image;

[0134] 7. Calculate the slant range influence factor (used to eliminate the influence of slant range in radiometric calibration):

[0135]

[0136] 8. Combining the above influencing factors, we can get the corrected DN value:

[0137]

[0138] Step 7: Obtain calibration parameters

[0139] The sixth step preprocessing result DN ‘2 Perform linear regression analysis with the Beta data of the Sentinel-1 SAR satellite after radiation correction, and output the calibration constant K and noise coefficient P n .

[0140] Output scaling constant K and noise figure P n The specific process is as follows:

[0141] 1. Pixels of varying brightness were collected from the same areas, such as water, forest, and urban areas, on the calibration image and the reference image. The mean of the corrected amplitude data in the calibration image and the mean of the beta data in the reference image were calculated. The calibration image is the corrected amplitude data, and the reference image is the beta image from the Sentinel-1 SAR satellite after radiometric correction.

[0142] 2. Use the least squares method for linear fitting. The input X-axis data is the data collected by the calibration satellite, the corrected amplitude data in the calibration image, and the Y-axis data is the Beta data in the reference image.

[0143] 3. The intersection of the fitting line and the Y-axis is taken as the noise coefficient Pn, and the slope of the fitting line is taken as the calibration constant K.

[0144] Step 8: Extrapolation of calibration parameters at different wave positions

[0145] In the sixth step above, the parameters related to the wave position have been separated from the calibration constants, the calibration constant K and the noise coefficient P n It is applicable to other wave position images under this mode, which is also the innovation of the present invention.

[0146] Step 9: Extrapolation of calibration parameters for different modes

[0147] The difference in image brightness in different modes is mainly affected by the azimuth energy accumulation. In the sixth step, the calculation formula of the influencing factor of TOPS mode imaging processing is given. In order to ensure that the calibration constant K and noise coefficient P in the seventh step can still be used in strip and beam modes n , a different formula is needed to eliminate the effects of imaging processing.

[0148] (1) Formula for the influence factor of stripe pattern imaging processing

[0149] The imaging processing mainly involves the accumulation of pulse pressure in the range direction and correlation in the azimuth direction. Under different parameters such as time width, satellite speed and slant range, the accumulated energy varies greatly, so this part needs to be eliminated in the radiation calibration.

[0150]

[0151] Where, L a Indicates the antenna azimuth length, V st represents the satellite speed, τ represents the pulse width, λ represents the central wavelength, R represents the slant range, prf represents the pulse repetition frequency, and f s Indicates the sampling bandwidth.

[0152] (2) Formula for the influencing factor of spotlight imaging processing

[0153] In order to ensure the consistency of radiation, the number of pulses accumulated in the azimuth direction is fixed to 32768. Therefore, the formula is as follows:

[0154]

[0155] Step 10: Algorithm Accuracy Evaluation

[0156] The algorithm accuracy was evaluated by comparing the radiometrically corrected images with the Sentinel-1 SAR satellite images. SAR images of urban and grassland areas were collected at the same incident angle and compared with the radiometrically corrected Sentinel-1 SAR satellite images. The error (root mean square error, RMSE) at the same wave position was less than 0.43 dB, and the radiometric error at the generalized wave position was less than 0.69 dB. Figure 2 .

[0157] In summary, the multi-wavelength and multi-mode rapid radiation calibration method based on tropical rainforests in the above embodiment uses a small amount of wavelength data analysis to solve the problem of obtaining calibration constants for all wavelengths of modes such as strip, beam, and TOPS, reducing the workload of data collection and analysis and meeting the needs of rapid calibration; the variable parameters involved include MGC, bandwidth, time width, slant range, satellite speed, range resolution, azimuth resolution, antenna radiation pattern and other parameters, which can effectively incorporate parameters affecting radiation calibration into the calibration constant calculation, thereby solving the theoretical basis for the promotion of different wavelengths.

[0158] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest, characterized in that: The following steps are involved: S1: Develop a calibration plan The calibration satellite was used to collect echo data of the 4-band TOPS calibration field at the same incident angle three times. The calibration field was the Amazon rainforest. S2: Imaging Processing The echo data collected by the calibration satellite is imaged using the omiga-K algorithm; S3: Parameter calculation and acquisition Calculate the slant range, incident angle, and transmit power during echo data acquisition, and obtain radar receiving gain control parameters, imaging processing bandwidth, wave position number, and antenna pattern; S4: Data screening The images obtained by imaging processing are grouped according to wave positions, and the data validity is analyzed, that is, the imaging results are screened according to the screening conditions and the imaging results that do not meet the screening conditions are eliminated; S5: Data Correction Calculate various influencing factors and use all influencing factors to pre-process the screened imaging results to obtain corrected amplitude data, where the imaging results are single-view complex data; S6: Get calibration parameters The corrected amplitude data is subjected to linear regression analysis with the Beta data corrected by the Sentinel-1 SAR satellite radiation, and the calibration constant K and noise coefficient P are output. n ; S7: Extrapolation of calibration parameters at different wave positions The calibration constant K and noise factor P n Extrapolation of other wave position images applied in TOPS mode; S8: Extrapolation of calibration parameters for different modes Calculate the imaging processing influence factors in strip and beam modes, replace the imaging processing influence factors in TOPS mode, and make the calibration constant K and noise coefficient P output in S6 n Able to extrapolate wave position images applied in strip and spotlight modes.

2. The multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest according to claim 1, characterized in that: In step S3, the parameters are specifically calculated as follows: The slope distance calculation method is as follows: First calculate the closest slope distance: The closest slope distance R near It is calculated from the accurate time delay t and light speed c obtained by the calibration within the payload slant range; Then, based on the range resolution, calculate the slant distance of each pixel: R index =R near +ΔR*index Among them, ΔR represents the range resolution of the SAR image, R index is the slope distance of the index-th pixel; The angle of incidence is calculated as follows: in, For radar perspective, N is the wave position number, is the satellite side swing angle, 0.01 represents the angular interval between wave positions, R e is the radius of the earth, h is the altitude of the satellite; The formula for calculating the transmit power is as follows: P v =P t *prf*τ Among them, P t is the peak power, τ is the pulse duration, and prf is the pulse repetition frequency.

3. The multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest according to claim 2, characterized in that: In step S3, the parameters are specifically obtained as follows: Radar receiving gain control parameters: directly obtained from radar imaging parameters and expressed as MGC; imaging processing bandwidth, wave position number, and antenna radiation pattern are known parameters.

4. The multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest according to claim 2, characterized in that: In step S5, the specific process is as follows: S51: Convert complex data into amplitude data: Wherein, b1 represents the real part of the complex image; b2 represents the imaginary part of the complex image; S52: Calculate the radiation power impact factor: S53: Calculate the MGC impact factor: Wherein, MGC represents the radar receiving gain control parameter; S54: Calculate the antenna pattern impact factor: Among them, R cent is the center slant distance, G() represents the antenna pattern; S55: Computational imaging processing impact factors: Among them, f s Indicates the sampling bandwidth, N burst Indicates the number of pulses per burst S56: Calculate the resolution impact factor: Wherein, ΔR represents the range resolution of the SAR image; ΔA represents the azimuth resolution of the SAR image; S57: Calculate the slope distance influence factor: Where R represents the slant distance; S58: Based on the above influencing factors, the corrected amplitude data, i.e. DN value, is obtained. The calculation formula is as follows:

5. The multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest according to claim 4 is characterized in that: In step S6, the calibration constant K and the noise coefficient P are output. n The specific process is as follows: S61: collecting pixel areas of different brightness in the same area of ​​the calibration image and the reference image, and calculating the mean of the corrected amplitude data in the calibration image and the mean of the Beta data in the reference image, where the calibration image is the corrected amplitude data and the reference image is the Beta image of the Sentinel-1 SAR satellite after radiometric correction; S62: Use the least squares method to perform linear fitting. The input X-axis data is the data collected by the calibration satellite, the corrected amplitude data in the calibration image, and the Y-axis data is the Beta data in the reference image. S63: The intersection of the fitting line and the Y axis is used as the noise coefficient Pn, and the slope of the fitting line is used as the calibration constant K.

6. The multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest according to claim 1, characterized in that: In step S8, the calculation formula of the imaging processing influence factor in the stripe mode is as follows: Among them, L a Indicates the antenna azimuth length, V st represents the satellite speed, τ represents the pulse width, λ represents the central wavelength, R represents the slant range, prf represents the pulse repetition frequency, and f s Indicates the sampling bandwidth.

7. The multi-wavelength and multi-mode rapid radiometric calibration method based on tropical rainforest according to claim 1, characterized in that: In step S8, the calculation formula of the imaging processing influence factor in the beamforming mode is as follows: Where τ represents the pulse width, f s represents the sampling bandwidth, and 32768 is the number of pulses accumulated in the azimuth direction.

Images