An improved calibration method and calibration device for the sky tilt curve of a microwave radiometer

Abstract

The present invention discloses an improved calibration method for the sky tilt curve of a microwave radiometer, which comprises the following steps: Step 1, assuming the noise temperature of the reference source, designing a non-linear system model of the microwave radiometer; Step 2, measuring and calculating the brightness temperature at different zenith angles; Step 3, obtaining the total atmospheric attenuation at each zenith angle by using the brightness temperature at each zenith angle and the average atmospheric radiation temperature; Step 4, performing linear regression on the total atmospheric attenuation and the secant value of the zenith angle, and carrying out the calibration process and other steps. The calibration method and device disclosed by the present invention solve the problems that the ground-based microwave radiometer is affected by the average atmospheric temperature and system non-linearity during the calibration of the sky tilt curve. It can improve the calibration success rate of the ground-based microwave radiometer during long-term unattended operation, thereby ensuring the detection accuracy during long-term operation.

Description

Technical Field

[0001] The present invention belongs to the field of atmospheric microwave remote sensing, and particularly relates to an improved calibration method and device for the sky tilt curve of a microwave radiometer in this field, which can improve the calibration accuracy of the microwave radiometer under unattended conditions. Background Art

[0002] A microwave radiometer is a device that measures the radiation brightness temperature of the atmosphere through passive remote sensing and inversely obtains atmospheric parameters such as the atmospheric temperature profile and relative humidity profile through methods such as neural networks. The microwave radiometer needs to be calibrated to determine the quantitative relationship between the received voltage signal and the radiation brightness temperature. Calibration is a prerequisite for using the microwave radiometer for measurement, and the calibration accuracy directly affects the measurement accuracy of the microwave radiometer for the atmospheric radiation brightness temperature (generally, the measurement accuracy is required to be within 1K), and further affects its inversion accuracy for environmental parameters.

[0003] Currently, the well-known calibration methods for microwave radiometers include: two-point absolute calibration method, four-point calibration method, sky tilt curve (Tipping-Curve) calibration method, etc. Among them, the sky tilt curve calibration method calculates the relatively accurate zenith brightness temperature using the relationship of the total atmospheric attenuation of the microwave radiometer at different zenith angles and uses it as a calibration source for calibration. The sky tilt curve calibration is a calibration method that does not require an additional reference source, and it has the advantages of convenience and speed compared with the two-point absolute calibration and four-point calibration methods. However, the existing sky tilt curve calibration algorithm does not fully consider the model error and system nonlinearity effects, which easily leads to calibration failure. Therefore, designing an improved sky tilt curve calibration method and device that considers the model error and system nonlinearity effects has important practical value. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an improved calibration method and device for the sky tilt curve of a microwave radiometer.

[0005] The present invention adopts the following technical solutions:

[0006] An improved calibration method for the sky tilt curve of a microwave radiometer, the improvement lies in that it includes the following steps:

[0007] Step 1, assume that the noise temperature of the reference source is T N , then the nonlinear system model of the microwave radiometer is:

[0008] V a =G(T sys +T a ) α

[0009] V a+N =G(T sys+T a +T N ) α

[0010] In the above formula, α is the nonlinear coefficient of the system, T a is the input radiation bright temperature of the antenna port, T sys is the noise temperature of the receiver, G is the total gain of the system, V a is the output voltage value of the system after square-law detection, V a+N is the output voltage value of the system after adding noise;

[0011] Define γ as the correlation coefficient between the linear regression curve and the observation data of each zenith angle during the calibration of the sky tilt curve:

[0012]

[0013] In the above formula, τ i is the total atmospheric attenuation corresponding to the i-th zenith angle before regression, is the average value of τ i ; τ i ' is the total atmospheric attenuation corresponding to the i-th zenith angle after regression, is the average value of τ i ', i = 1, 2, 3, 4, 5…;

[0014] Define σ as the root mean square error between the linear regression curve and the observation data of each zenith angle during the calibration of the sky tilt curve:

[0015]

[0016] In the above formula, Δτ i is the error value of the total atmospheric attenuation corresponding to the i-th zenith angle;

[0017] Select the measured zenith angle θ i according to the magnitude of the zenith atmospheric bright temperature value;

[0018] Step 2, measure and calculate the bright temperature T a (θ):

[0019]

[0020] In the above formula, V a (θ) is the measured voltage value at the zenith angle θ i , G0 is the preset total gain of the system, T sys0 is the preset noise temperature of the receiver;

[0021] Step 3, use the bright temperature of each zenith angle and the average atmospheric radiation temperature to find the total atmospheric attenuation of each zenith angle:

[0022]

[0023] In the above formula, T m is the average atmospheric radiation temperature, and T sky is the brightness temperature at each zenith angle;

[0024] After linear regression, a new brightness temperature T sky '(0°) is calculated:

[0025] T sky '(0°) = 2.73·e -τ'(0°) + Tm(0°)·[1 - e -τ'(0°)

[0026] Step 4: Perform a linear regression on the total atmospheric attenuation and the secant value of the zenith angle. If the correlation coefficient γ ≥ 0.999 and the root mean square error |σ| ≤ σ min , the calibration is successful, and the new calibration coefficients G' and T' are calculated through the following formula SYS and the calibration process ends:

[0027]

[0028]

[0029] In the above formula, V H is the voltage value of the measured blackbody reference source, T H is the absolute temperature of the measured blackbody reference source, and V sky is the voltage value when measuring the zenith;

[0030] Step 5: If the correlation coefficient γ ≥ 0.999 and the root mean square error σ > σ min , then adjust the new T m value in increments or decrements of 0.1 K and continuously repeat Steps 3 and 4 until σ ≤ σ min , and then calculate the new calibration coefficients G' and T' SYS and end the calibration process;

[0031] Step 6: If the correlation coefficient γ < 0.999 or the root mean square error σ > σ min always holds, then modify the non - linear system model in Step 1 to:

[0032]

[0033] In the above formula, V H+N is the voltage value of the measured blackbody reference source after adding noise;

[0034] Repeat Steps 3 - 5 until the calibration is successful. If the conditions are still not met after repeating a certain number of times, the calibration is determined to fail.​

[0035] Further, the average atmospheric radiation temperature in step 3 is calculated by a ground temperature and humidity statistical model.

[0036] Further, the certain number of times in step 6 is set to 100 times.

[0037] An improved calibration device for the sky tilt curve of a microwave radiometer, which is used to implement the above calibration method, is improved in that: an additional internal noise source is added to the receiver as an internal reference source to realize system nonlinear correction.

[0038] The beneficial effects of the present invention are:

[0039] The calibration method and device disclosed by the present invention solve the problems that the ground-based microwave radiometer is affected by the average atmospheric temperature and system nonlinearity during the sky tilt curve calibration. It can improve the calibration success rate of the ground-based microwave radiometer during long-term unattended operation, thereby ensuring the detection accuracy during long-term operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a schematic flow chart of the calibration method disclosed by the present invention;

[0041] Figure 2 is a correlation coefficient diagram obtained by the calibration method disclosed by the present invention;

[0042] Figure 3 is a comparison diagram of the statistical results of the original method and the method of the present invention;

[0043] Figure 4 is a block diagram of the composition of the calibration device disclosed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0045] As an all-weather atmospheric observation device, the ground-based microwave radiometer needs to have the capabilities of unattended observation and automatic calibration. The sky tilt curve calibration method can achieve continuous high-precision calibration without other external auxiliary reference sources such as liquid nitrogen, and the calibration result is controllable (it can be judged whether the calibration is accurate through the correlation coefficient, etc.). Therefore, the sky tilt curve calibration algorithm is generally used for the automatic calibration of the radiometer. The existing sky tilt curve calibration algorithms usually assume that the atmosphere is stable and horizontally uniformly distributed, and such an ideal assumption is prone to calibration failure.

[0046] Example 1. This example discloses an improved calibration method for the sky tilt curve of a microwave radiometer, which takes into account the atmospheric model error and the influence of system nonlinearity, and can improve the calibration success rate of the microwave radiometer. As Figure 1 shown, it includes the following steps:

[0047] Step 1. Assume that the noise temperature of the reference source is T N , then the nonlinear system model of the microwave radiometer is:

[0048] V a = G(T sys + T a ) α

[0049] V a+N = G(T sys + T a + T N ) α

[0050] In the above formula, α is the nonlinear coefficient of the system, T a is the input radiation brightness temperature of the antenna port, T sys is the noise temperature of the receiver, G is the total gain of the system, V a is the output voltage value of the system after square-law detection, V a+N is the output voltage value of the system after adding noise;

[0051] Define γ as the correlation coefficient between the linear regression curve and the observation data of each zenith angle during the sky tilt curve calibration:

[0052]

[0053] In the above formula, τ i is the total atmospheric attenuation corresponding to the i-th zenith angle before regression, is the average value of τ i ; τ i ' is the total atmospheric attenuation corresponding to the i-th zenith angle after regression, is the average value of τ i ', i = 1, 2, 3, 4, 5...;

[0054] Define σ as the root mean square error between the linear regression curve and the observation data of each zenith angle during the sky tilt curve calibration:

[0055]

[0056] In the above formula, Δτ i is the error value of the total atmospheric attenuation corresponding to the i-th zenith angle;

[0057] The parameters γ and σ are the key indicators to define whether the calibration is successful.

[0058] Reasonably select the zenith angle θ for measurement according to the magnitude of the zenith atmospheric brightness temperature value i ;

[0059] Step 2, measure and calculate the brightness temperature T a (θ):

[0060]

[0061] In the above formula, V a (θ) is the measured voltage value at the zenith angle θ i , G0 is the preset total system gain, and T sys0 is the preset receiver noise temperature;

[0062] Step 3, use the brightness temperature at each zenith angle and the average atmospheric radiation temperature (which can be calculated through statistical models such as ground temperature and humidity) to obtain the total atmospheric attenuation at each zenith angle:

[0063]

[0064] In the above formula, T m is the average atmospheric radiation temperature, and T sky is the brightness temperature at each zenith angle;

[0065] After linear regression, calculate the new brightness temperature T sky '(0°):

[0066] Tsky'(0°) = 2.73e-τ'(0°) + Tm(0°)·[1 - e-τ'(0°)]

[0067] Step 4, perform linear regression on the total atmospheric attenuation and the secant value of the zenith angle. If the correlation coefficient γ ≥ 0.999 and the root mean square error σ ≤ σ min , then the calibration is successful, and calculate the new calibration coefficients G' and T' through the following formula SYS and end the calibration process:

[0068]

[0069]

[0070] In the above formula, V H is the voltage value of the measured blackbody reference source, T H is the absolute temperature of the measured blackbody reference source, and V sky is the voltage value when measuring the zenith;

[0071] Step 5, if the correlation coefficient γ ≥ 0.999 and the root mean square error σ > σ min, then adjust and calculate the new T in increments or decrements of 0.1K m value, and continuously repeat steps 3 and 4 until σ ≤ σ min After that, calculate the new calibration coefficients G' and T' SYS and end the calibration process;

[0072] Step 6, if the correlation coefficient γ < 0.999 or the root mean square error σ > σ always min , then modify the non-linear system model in step 1 to:

[0073]

[0074] In the above formula, V H+N is the voltage value of the measured blackbody reference source after adding noise;

[0075] Repeat steps 3 - 5 until the calibration is successful. If the conditions are still not met after repeating a certain number of iterations (which can be set to 100 times), it is determined that the calibration fails.

[0076] The calibration results obtained by using the calibration method of the present invention in the Qingdao area from 2017 to 2018 were statistically analyzed. During the statistics, the measurement results on cloudy and rainy days were removed, and only 152 sets of calibration results during the sounding period were retained. The correlation coefficients obtained by the calibration method of the present invention are as Figure 2 shown. It can be seen from the figure that the correlation coefficient after improvement has been significantly improved. The statistical results obtained by using the linear calibration model (original method) and the non-linear calibration model (calibration method of the present invention) are compared as Figure 3 shown. By using the calibration method of the present invention, the proportion of γ ≥ 0.999 reaches 98%, and the calibration success rate is much higher than the original about 60%.

[0077] This embodiment also discloses an improved calibration device for the sky tilt curve of a microwave radiometer, which is used to implement the above calibration method. As Figure 4 shown, an additional internal noise source is added to the receiver as an internal reference source to achieve system non-linear correction.

Claims

1. An improved calibration method for the sky tilt curve of a microwave radiometer, characterized in that, It includes the following steps: Step 1, assume that the noise temperature of the reference source is T N , then the non-linear system model of the microwave radiometer is as follows: V a = G(T sys + T a ) α V a+N = G(T sys + T a + T N ) α In the above formula, α is the nonlinear coefficient of the system, T a is the input radiant brightness temperature of the antenna port, T sys is the noise temperature of the receiver, G is the total gain of the system, V a is the output voltage value of the system after square-law detection, V a+N is the output voltage value of the system after adding noise; Define γ as the correlation coefficient between the linear regression curve and the observed data of each zenith angle during the calibration of the sky tilt curve: In the above formula, τ i is the total atmospheric attenuation corresponding to the i-th zenith angle before regression, is the average value of τ i ; τ i ' is the total atmospheric attenuation corresponding to the i-th zenith angle after regression, is the average value of τ i ', where i = 1, 2, 3, 4, 5…; Define σ as the root mean square error between the linear regression curve and the observed data of each zenith angle during the calibration of the sky tilt curve: In the above formula, Δτ i is the error value of the total atmospheric attenuation corresponding to the i-th zenith angle; Select the zenith angle θ to be measured according to the magnitude of the brightness temperature value of the zenith atmosphere i ; Step 2, measure and calculate the brightness temperature T a (θ): In the above formula, V a (θ) is the measured voltage value at the zenith angle θ i , G0 is the preset total system gain, and T sys0 is the preset receiver noise temperature; In step 3, use the brightness temperature and the average atmospheric radiation temperature of each zenith angle to calculate the total atmospheric attenuation of each zenith angle: In the above formula, T m is the average atmospheric radiation temperature, and T sky is the brightness temperature at each zenith angle; A new brightness temperature T is calculated after linear regression sky '(0°): T sky '(0°) = 2.73·e -τ'(0°) + Tm(0°)·[1 - e -τ'(0°) ​ Step 4: Perform a linear regression on the total atmospheric attenuation and the secant value of the zenith angle. If the correlation coefficient γ ≥ 0.999 and the root mean square error |σ| ≤ σ min , the calibration is successful. Calculate the new calibration coefficients G' and T' using the following formula SYS and end the calibration process: In the above formula, V H is the voltage value of the measured blackbody reference source, and T H is the absolute temperature of the measured blackbody reference source, and V sky is the voltage value when measuring the zenith; Step 5, if the correlation coefficient γ ≥ 0.999 and the root mean square error |σ| > σ min , then adjust and calculate a new T value in increments or decrements of 0.1K m , and continuously repeat Steps 3 and 4 until |σ| ≤ σ min , then calculate the new calibration coefficients G' and T' SYS and end the calibration process; Step 6, if the correlation coefficient γ < 0.999 or the root mean square error |σ| > σ all the time min , then modify the non-linear system model in Step 1 to be: In the above formula, V H+N is the voltage value of the measured blackbody reference source after adding noise; Repeat steps 3 - 5 until the calibration is successful. If the conditions are still not met after repeating a certain number of times, it is determined that the calibration fails.

2. The improved calibration method for the sky tilt curve of a microwave radiometer according to claim 1, characterized in that: The average atmospheric radiation temperature in step 3 is calculated through the ground temperature and humidity statistical model.

3. The improved calibration method for the sky tilt curve of a microwave radiometer according to claim 1, characterized in that: Set the certain number of times in step 6 to 100 times.

4. An improved sky tilt curve calibration device for a microwave radiometer, which is used to implement the calibration method described in claim 1, and is characterized in that: Add an additional internal noise source in the receiver as an internal reference source to achieve system nonlinear correction.

Images