Microwave radiometer cold space reconstruction calibration method based on bilateral cold space observations

Abstract

This invention discloses a microwave radiometer cold-sky reconstruction calibration method based on dual-sided cold-sky observations, belonging to the technical field of on-orbit calibration and data processing for spaceborne microwave remote sensing payloads. The method includes: acquiring a first-sided cold-sky observation sample to be processed, a first-sided cold-sky modeling observation sample, and a second-sided reference cold-sky field of view set; calculating the first-sided cold-sky mean, the second-sided reference cold-sky mean, and the internal dispersion of the second-sided cold-sky; constructing a first-sided cold-sky prediction model to calculate the predicted mean of the first-sided cold-sky; statistically analyzing background distribution characteristics; constructing anomaly judgment thresholds and judgment conditions; performing first-sided cold-sky anomaly discrimination and second-sided reference cold-sky alarm discrimination on the first-sided cold-sky observation sample to be processed, determining whether to reconstruct based on the discrimination results, and outputting the corresponding identifier. Therefore, this invention utilizes the second-sided reference cold-sky to reconstruct the first-sided cold-sky when the first-sided cold-sky is anomaly, improving the recoverability of the cold reference and the retention rate of observation data during the affected period.

Description

Technical Field

[0001] This invention relates to a cold-space reconstruction calibration method for microwave radiometers based on dual-sided cold-space observations, belonging to the technical field of on-orbit calibration and data processing of spaceborne microwave remote sensing payloads. Background Technology

[0002] Spaceborne microwave radiometers are widely used for atmospheric temperature and humidity detection, and the accuracy of their observed brightness temperature directly affects subsequent inversion and operational effectiveness. For scanning microwave radiometers, a two-point calibration method combining thermal and cold references is typically used to convert the instrument's raw observations into antenna temperature or brightness temperature products. The thermal reference is generally provided by the instrument's internal thermal load, while the cold reference is typically provided by cold-space observations. Due to the stable radiation background and low equivalent brightness temperature of cold-space, cold-space observations have long served as an important cryogenic reference for on-orbit calibration.

[0003] For low-Earth orbit satellite platforms, the cold space field of view may be affected by lunar intrusion under specific orbital configurations and scanning geometry conditions. After lunar radiation enters the cold space field of view, it will cause the cold reference observations to deviate from the true deep space state, thereby affecting the calibration gain and bias, and propagating the error to subsequent Earth observation brightness temperature products, which manifests as brightness temperature anomalies, time series discontinuities, and a decrease in long-term stability.

[0004] Current technologies for handling cold-air contamination typically employ methods such as contaminated sample removal, shielding of abnormal time periods, or correction methods based on geometric relationships and empirical parameters. These methods are generally geared towards single-sided cold-air calibration structures: on the one hand, when the primary cold reference is continuously contaminated, samples from the corresponding time periods often have to be discarded, making it difficult to balance data preservation and calibration continuity; on the other hand, existing methods primarily focus on identifying and correcting contaminated cold air, failing to fully utilize the redundant constraints between bilateral cold-air observations to directly reconstruct the contaminated primary cold air. Therefore, current technologies still lack a microwave radiometer cold-air reconstruction calibration method based on bilateral cold-air observations. Summary of the Invention

[0005] This invention addresses the problems of cold reference distortion, sample loss, unrecoverable data, and insufficient calibration continuity in existing microwave radiometer cold space calibration under lunar contamination conditions. It provides a microwave radiometer cold space reconstruction calibration method based on dual-sided cold space observations. When the cold space on the operational calibration side is contaminated, the contaminated cold reference is reconstructed using the second-sided reference cold space and used for subsequent calibration correction.

[0006] To achieve the above-mentioned technical objectives, the present invention will adopt the following technical solution:

[0007] A microwave radiometer cold-space reconstruction calibration method based on bilateral cold-space observations includes the following steps:

[0008] Step S1: Obtain the first-side cold air observation sample to be processed and the full-scan observation data; construct the first-side cold air modeling observation sample under the specified modeling period based on the first-side cold air observation sample to be processed; select the second-side reference cold air field of view set based on the full-scan observation data;

[0009] Step S2: Construct cold air features for each scanning cycle based on the observation samples of the first-side cold air model and the set of the second-side reference cold air field of view; the cold air features include the mean of the first-side cold air, the mean of the second-side reference cold air, and the internal dispersion of the second side;

[0010] Step S3: Linearly fit the first-side cold air mean value and the second-side reference cold air mean value under the same scanning period obtained in step S2 to construct the first-side cold air prediction model. Then, under the specified modeling period, calculate the corresponding first-side cold air prediction mean value based on each second-side reference cold air mean value, and calculate the corresponding first-side cold air prediction value based on the observation values ​​of each field of view of the second-side reference cold air in the second-side reference cold air field of view set.

[0011] Step S4: Based on the mean value of the first side cold air and the mean value of the second side reference cold air obtained in step S2, and the mean value of the first side cold air prediction and the predicted value of each field of view of the first side cold air obtained in step S3, the background distribution characteristics of the first side cold air modeling observation sample and the second side reference cold air field of view set are statistically analyzed in each scanning cycle.

[0012] Step S5: Construct anomaly detection threshold and detection conditions based on background distribution characteristics and the internal dispersion of the second side;

[0013] Step S6: Based on the anomaly detection threshold and detection conditions, each scanning cycle of the observation sample to be processed in the first side cold air is detected, and based on the detection results, it is determined whether the observation sample of the corresponding scanning cycle needs to be reconstructed and replaced. After the process is completed, the subsequent calibration process begins.

[0014] As a further improvement of the present invention, in step S1, when constructing the first side cold air modeling observation sample, based on the first side cold air observation sample to be processed, the observation samples of each field of view of the first side cold air under the specified modeling period are obtained for each channel and each orbit type, so as to construct the first side cold air modeling observation sample.

[0015] When selecting the second-side reference cold air field of view set, based on full-scan observation data, from the non-Earth observation area opposite to each field of view of the first-side cold air, for each channel and each orbit type, a set of continuous scan positions that meet the requirements of internal consistency, temporal stability, and stability of the relationship with the mean of the first-side cold air are selected within a specified modeling period as the second-side reference cold air field of view set. This is obtained through the following steps:

[0016] During the specified modeling period, the average observation value is calculated for each scan position of the full scan observation data, and the Earth observation area and non-Earth observation area are identified by combining the scan position distribution; in the non-Earth observation area opposite to each field of view of the first side cold air, candidate areas are determined; for the candidate continuous scan position set in the candidate area, the internal consistency, temporal stability and the stability of the relationship with the mean of the first side cold air are evaluated respectively, and among the candidate continuous scan position sets that meet the requirements of internal consistency and temporal stability, the set with the smallest residual dispersion after establishing a benchmark relationship with the mean of the first side cold air is selected as the second side reference cold air field of view set.

[0017] As a further improvement of the present invention, in step S2, the mean value of the first side cold air, the mean value of the second side reference cold air, and the internal dispersion of the second side are calculated by the following formulas:

[0018] ;

[0019] ;

[0020] ;

[0021] In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below; For the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Observations from each field of view This represents the field of view of the first side of the cold air. For field of view index, ; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air average; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Each field of view observation value; For the first The first channel, the first Orbital-like structures in the scan cycle The internal dispersion of the second side below.

[0022] As a further improvement of the present invention, in step S3, the first-side cold air prediction model is expressed as:

[0023] ;

[0024] ;

[0025] In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast mean; Indicates the slope of a linear mapping; Represents the intercept of a linear mapping; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air average; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Predicted values ​​for each field of view; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Observations from each field of view.

[0026] As a further improvement of the present invention, in step S4, the statistically obtained background distribution features include the mean deviation of the first-side cold air, the residuals of each field of view of the first-side cold air, and the deviations of each field of view of the second-side reference cold air, wherein:

[0027] The formula for calculating the deviation of the mean value of the first-side cold air is as follows:

[0028] ;

[0029] The formulas for calculating the residuals of each field of view in the first side cold air are as follows:

[0030] ;

[0031] The formulas for calculating the deviation of each field of view of the second-side reference cold air are as follows:

[0032] ;

[0033] In the above formula, Indicates the first The first channel, the first Orbital-like structures in the scan cycle The deviation of the mean value of the first side cold air below; This represents the average value of the cold air on the first side within the corresponding scanning period. Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast mean; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view residual; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view observation value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air One field of view prediction value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Individual field of view deviation; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Observations from each field of view Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side references the average value of cold air.

[0034] As a further improvement of the present invention, in step S5, the anomaly determination threshold includes the first side mean anomaly threshold. First side cold air Anomaly threshold for each field of view Second-side internal dispersion alarm threshold And the second side reference cold air Anomaly threshold for each field of view They are expressed by the following formulas respectively:

[0035] ;

[0036] ;

[0037] ;

[0038] ;

[0039] In the above formula: Indicates the first threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle Deviation of the first side cold air mean The background standard deviation; Indicates the second threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air individual field-of-view residuals The background standard deviation; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side internal dispersion The background mean; Indicates the third threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side internal dispersion The background standard deviation; This represents the fourth threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below individual field deviation The background standard deviation;

[0040] The judgment criteria include the criteria for determining the first-side cold air anomaly and the criteria for determining the second-side reference cold air alarm, wherein:

[0041] The conditions for determining an anomaly in the first-side cold air are: ;

[0042] The conditions for determining a second-side reference cold air alarm are: or .

[0043] As a further improvement of the present invention, step S6, when determining each scanning cycle of the first side cold air observation sample to be processed, specifically includes the following steps:

[0044] S6.1 Based on the observation samples to be processed in the first side cold air and the full scan observation data, for each channel and each orbit type, the observation values ​​of each field of view of the first side cold air and the observation values ​​of each field of view of the second side reference cold air under each scan cycle are obtained one by one, and then the average value of the first side cold air and the average value of the second side reference cold air under the corresponding scan cycle are calculated.

[0045] S6.2. Using the first-side cold air prediction model, the corresponding first-side cold air prediction mean is calculated by using the second-side reference cold air mean under each scanning period.

[0046] S6.3 Calculate the mean deviation of the first side cold air, the residual of each field of view of the first side cold air, the internal dispersion of the second side, and the deviation of each field of view of the second side reference cold air under each current scanning cycle;

[0047] S6.4. Compare the mean deviation of the first side cold air, the residual of each field of view of the first side cold air, the internal dispersion of the second side, and the deviation of each field of view of the second side reference cold air under each current scanning cycle with the anomaly judgment threshold constructed in step S5. Determine whether the first side cold air is abnormal and whether the second side reference cold air is alarmed under the current scanning cycle by the judgment conditions.

[0048] As a further improvement of the present invention, a reconstruction step is included after step S6:

[0049] Only when the determination result indicates that the first-side cold air is abnormal in the current scanning period and the second-side reference cold air does not alarm, it is necessary to reconstruct the observation samples of the corresponding scanning period to obtain the reconstructed first-side multi-field cold air observation sequence, and replace the original first-side cold air field observation sequence with the reconstructed first-side multi-field cold air observation sequence to enter the subsequent calibration process.

[0050] The reconstructed first-side multi-field cold air observation sequence is calculated using the following formula:

[0051] ;

[0052] ;

[0053] ;

[0054] In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air One field of view reconstruction value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view observation value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air mean correction amount; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side cold air after reconstruction; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast average.

[0055] As a further improvement of the present invention, the final output includes: the reconstructed first-side cold air mean, the reconstructed first-side cold air multi-field observation sequence, the first-side anomaly indicator, the second-side reference cold air alarm indicator, and the reconstruction indicator.

[0056] When the first-side cold air is abnormal and the second-side reference cold air does not alarm, the output reconstruction flag is: Reconstruction valid;

[0057] When the cold air on the first side is abnormal and the reference cold air on the second side is alarmed, the output reconstruction identifier is: abnormal;

[0058] When the first side of the cold air is normal, the output reconstruction identifier is: Original cold air is valid.

[0059] Another technical objective of this invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described microwave radiometer cold space reconstruction calibration method based on bilateral cold space observations.

[0060] Based on the above-mentioned technical objectives, the present invention has the following advantages compared with the prior art:

[0061] 1. This invention is aimed at a dual-sided cold space observation structure. When the operational calibration side cold space (i.e., the first side cold space) is affected by anomalies such as lunar contamination, this invention can use the second side reference cold space to reconstruct the first side cold space, thereby improving the recoverability of the cold reference and the retention rate of observation data during the affected period.

[0062] 2. By jointly considering the deviation of the mean value of the first-side cold air (constructing an anomaly threshold for the mean value of the first side), the residual of each field of view of the first-side cold air (constructing an anomaly threshold for each field of view of the first side, reflecting anomalies in the field of view of the first side), the internal dispersion of the second side (constructing an alarm threshold for the internal dispersion of the second side), and the deviation of each field of view of the second-side reference cold air (constructing an anomaly threshold for each field of view of the second side, reflecting anomalies in the field of view of the second side), this invention can better distinguish between different scenarios such as "abnormal first side, stable second-side reference" and "unstable second-side reference", thereby improving the reliability of anomaly identification and reconstruction.

[0063] 3. This invention does not change the original calibration processing framework, but improves the cold reference input entering the original calibration process by reconstructing the cold space. Therefore, it is easy to be compatible with the existing business processing chain and facilitates the retrospective correction of on-orbit historical data and its application in business processing systems.

[0064] 4. This invention is applicable to microwave radiometers with dual- or multi-sided cold space observation capabilities, and has practical application value for improving the retention rate of observation data, calibration continuity and product stability under lunar contamination conditions. Attached Figure Description

[0065] Figure 1 This is the overall flowchart of the microwave radiometer cold space reconstruction and calibration method based on dual-sided cold space observation as described in this invention;

[0066] Figure 2 This is a schematic diagram of the calibration antenna temperature results of channel 1 of the CAMS microwave radiometer, carried on the Tianyan-16 commercial small satellite platform, before correction during the period of lunar pollution impact (September 26, 2025).

[0067] Figure 3 This is a schematic diagram showing the temperature results of the calibration antenna of channel 1, which is carried on the Tianyan-16 commercial small satellite platform during the lunar pollution period (September 26, 2025), after being corrected by the cold space reconstruction calibration method described in this invention, under the orbital ascent conditions. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and should not be regarded as any limitation on the present invention or its application or use. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Unless otherwise specifically stated, the relative arrangement, expressions, and values ​​of components and steps set forth in these embodiments do not limit the scope of the present invention. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0069] like Figure 1 As shown, the microwave radiometer cold space reconstruction and calibration method based on dual-sided cold space observation described in this invention takes the CAMS microwave radiometer, carried on the Tianyan-16 commercial small satellite platform, as the object. The CAMS microwave radiometer adopts a cross-orbit scanning method with full-circle sampling and has dual-sided cold space observation capability within one scanning cycle. The first-sided cold space is used for operational calibration, and the second-sided reference cold space can serve as a redundant reference. Based on the stable correspondence between the dual-sided cold spaces under uncontaminated conditions, when the first-sided cold space is contaminated by the moon, the second-sided reference cold space can be used to detect, constrain, and reconstruct the contaminated cold reference, and further used for calibration correction. This invention processes channels 01 to 18, and the orbit types are divided into ascending and descending orbits, which are modeled and processed separately. Specifically, it includes the following steps:

[0070] Step S1, Obtain data:

[0071] The process reads the observation file to be processed and extracts the observation values ​​of each field of view of the first-side cold sky under each scanning cycle, as well as the full-scan observation data. Based on the observation file to be processed, the observation values ​​of each field of view of the first-side cold sky under each scanning cycle constitute the first-side cold sky observation sample to be processed in the subsequent operational processing stage. In this invention, the selected fields of view for the first-side cold sky are L66, L67, L68, and L69 (L represents the left side). In this invention, the full-scan observation data refers to the observation data acquired by the CAMS microwave radiometer at all scanning positions within one scanning cycle, used to identify the Earth observation area, non-Earth observation area, and to select the subsequent required set of second-side reference cold sky fields of view.

[0072] Based on the first-side cold air observation samples to be processed, observation samples of each field of view of the first-side cold air under a specified modeling period are obtained to construct the first-side cold air modeling observation samples. In this invention, the first-side cold air observation samples to be processed specifically refer to the cold air observations used as the operational calibration side within a scanning cycle. They are directly given in the observation files to be processed and serve as the main cold reference input of the original calibration process (i.e., the existing operational calibration processing process).

[0073] Based on full-scan observation data, a set of continuous scan positions that meet the requirements of internal consistency, temporal stability, and stability of the relationship with the mean of the first-side cold sky are selected from the non-Earth observation area on the other side, opposite to the fields of view of the first-side cold sky, within a specified modeling period. This set serves as the second-side reference cold sky field of view set (including scan positions and observation samples at each scan position). The specified modeling period is mainly an observation period selected based on experience, where there is no lunar contamination or other obvious anomalies, and the correspondence between the two-side cold skies is stable. In this embodiment, the specified modeling period is from July 12, 2025 to August 13, 2025.

[0074] The first-side cold air mean refers to the average value calculated from the observation values ​​of each field of view of the first-side cold air in the first-side cold air unprocessed observation sample / first-side cold air modeling observation sample under the same channel, the same orbit type and the same scanning period. It is used to characterize the overall observation level of all fields of view of the first-side cold air under this scanning period.

[0075] Therefore, the core of this step is to obtain the first-side cold air observation samples to be processed, the first-side cold air modeling observation samples, and the second-side reference cold air field of view set, which is achieved through the following steps:

[0076] Step 1.1, Data Reading:

[0077] The system reads the observation files to be processed and extracts the observation values, full-scan observation data, thermal reference observation values, Earth observation values, and corresponding channel numbers, orbit types, scan frame numbers, time information, and quality identification information for each field of view of the first-side cold air under each scan cycle. Observation samples with missing measurements, saturation, invalid identifiers, or abnormal data formats in each field of view of the first-side cold air are removed or masked to form the observation samples to be processed in the first-side cold air. Observation samples for a specified modeling period are then extracted from the observation samples to construct the first-side cold air modeling observation samples.

[0078] Step 1.2, Sample grouping:

[0079] The observation samples in the full scan observation data are grouped according to channel number and orbit type to obtain independent full scan sample sets for each channel under different orbit types (ascending orbit and descending orbit).

[0080] Similarly, the observation samples in the first side cold air unprocessed observation sample are also grouped by channel number and orbit type to obtain an independent unprocessed sample set for each channel under different orbit types (ascending orbit and descending orbit).

[0081] The subsequent steps, such as the second-side reference cold air field set selection (step 1.3), the first-side cold air prediction model construction (step S3), the construction of anomaly judgment thresholds and judgment conditions (step S5), and cold air reconstruction (step S7), are all carried out separately within the corresponding channel and the corresponding orbit type.

[0082] Step 1.3, Second-side reference cold air field set filtering:

[0083] Since the original observation file to be processed does not directly provide the set of second-side reference cold air fields of view, this invention needs to select the set of second-side reference cold air fields of view from the full-scan observation data, which specifically includes the following steps:

[0084] 1.3.1 Calculate the average observation value at each scanning position:

[0085] During a specified modeling period, the average observation value is calculated for each scan location (or cold air field) of the full-scan observation data, and the Earth observation area and non-Earth observation area are identified by combining the scan location distribution; this is specifically achieved in the following way:

[0086] The first in the full scan observation data Each channel at the scanning position average observed value Calculated using the following formula:

[0087] ;

[0088] In the above formula, Indicates the first in the full scan observation data Each channel at the scanning position , Observations during the scan period; Indicates the first in the full scan observation data Each channel at the scanning position The total number of observation samples is also the total number of scanning cycles for the full scan observation data.

[0089] 1.3.2 Identifying Earth observation areas and non-Earth observation areas:

[0090] Based on the average observations at each scan location within the specified modeling period The distribution characteristics that vary with the scanning position, combined with the spatial continuity of the scanning position, identify the Earth observation area and the non-Earth observation area located on both sides of it.

[0091] 1.3.3 Determine the candidate regions to be used for filtering:

[0092] Within the non-Earth observation area opposite to each field of view of the first side cold air, candidate regions for screening the second side reference cold air field of view set are determined by combining the continuity and availability of candidate scan positions.

[0093] 1.3.4. Evaluate the internal consistency of the candidate continuous scan location set:

[0094] Based on the full-scan observation data, targeting the first [unclear] within the candidate region Set of candidate continuous scan positions for each channel The first number is calculated using the following formula. The first channel, the first Orbital-like structures in the scan cycle mean of the candidate set :

[0095] ;

[0096] Based on the above candidate set mean Calculate the first candidate region Set of candidate continuous scan positions for each channel During the scan cycle Internal consistency index :

[0097] ;

[0098] In the above formula, K is the total number of scan positions in the candidate continuous scan position set; k is the index of the candidate continuous scan position set, k=1,2,3……K; This refers to the k-th scan position in the set of candidate continuous scan positions. Indicates the first The first channel, the first Orbital-like, scanning period Down scan position The observed values.

[0099] 1.3.5. Comprehensive evaluation and selection of the optimal set:

[0100] Based on the full-scan observation data, for each candidate continuous scan position set within the candidate region, combined with the internal consistency index obtained in step 1.3.4, the stability of its temporal variation within the specified modeling period and the degree of residual dispersion in its correspondence with the overall mean of the first-side cold air are further examined for screening. Specifically, among the candidate continuous scan position sets that meet the requirements of internal consistency index and temporal stability, the set with the smallest residual dispersion is selected as the second-side reference cold air field of view set.

[0101] Among them, the stability of time variation The value used to characterize the fluctuation of the candidate continuous scan location set over time within a specified modeling period is defined as: ,in: This represents the standard deviation statistical function.

[0102] To calculate the residual dispersion of the correspondence between each observation value corresponding to the candidate continuous scan position set and the mean value of the first side cold air, this embodiment first uses the same channel Same orbit type Same scan cycle mean of the candidate set Average value of cold air on the first side Establish a linear relationship model, and then calculate the first linear relationship model. The first channel, the first Orbital-like structures in the scan cycle Calculated value of the first side cold air mean :

[0103] ;

[0104] In the above formula, This represents the slope of the candidate linear fit. This represents the intercept of the candidate linear fit.

[0105] In the set of candidate continuous scan positions, the first The first channel, the first Residual dispersion of orbital types Defined as:

[0106] ;

[0107] In the above formula, Indicates the first position in the set of candidate continuous scan positions. The first channel, the first Orbital-like, scanning period The residual is calculated using the following formula:

[0108] .

[0109] 1.3.6 Output the filtering results:

[0110] The final selected set of continuous scan positions is used as the second-side reference cold air field of view set. For example, in this embodiment, the length of the candidate continuous scan position set is 4, and the field of view selection results of the second-side reference cold air are R172, R173, R174, and R175 (R represents the right side).

[0111] The second-side reference cold air field of view set described in this invention consists of the scanning positions and observation values ​​of each field of view of the second-side reference cold air, located on the opposite side of each field of view of the first-side cold air, and selected from the full-scan observation data. Furthermore, in use, the second-side reference cold air field of view set described in this invention is not directly used as input to the original calibration process, but rather as a redundant reference side for anomaly detection and reconstruction of observation values ​​in the first-side cold air observation samples to be processed.

[0112] Step S2, Cold Air Feature Construction:

[0113] Based on the first-side cold air modeling observation samples and the second-side reference cold air field of view set obtained in step S1, cold air features are constructed for each channel and each orbit type within a specified modeling period. The cold air features include bilateral cold air feature quantities and the internal dispersion of the second side, wherein:

[0114] The dual-sided cold air characteristic quantities correspond to the first and second-sided cold air characteristic quantities, each constructed from observations of multiple cold air fields of view on the corresponding side, representing the overall level of the cold air on that side. In this invention, the first-sided and second-sided cold air characteristic quantities correspond to the first-sided cold air mean and the second-sided reference cold air mean, respectively. Both are average values ​​calculated from observations of multiple cold air fields of view on the corresponding side under the same channel, the same orbit type, and the same scanning period, and are used to characterize the overall observation level of the cold air on the corresponding side under that scanning period.

[0115] The internal dispersion of the second side is used to characterize the internal consistency and stability of the second-side reference cold air within the current scan period.

[0116] Based on the first-side cold air modeling observation samples and the second-side reference cold air field of view set, for each scanning cycle, assuming the first... The first channel, the first Orbital-like, scanning period Below, the field of view numbers for the first-side cold air and the second-side reference cold air are both... The mean cold air on the first side and the mean cold air on the second side are expressed as follows:

[0117] ; ;

[0118] In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below; For the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Observations from each field of view For field of view index, ; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air average; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Each field of view observation value.

[0119] Second side internal dispersion Represented as:

[0120] .

[0121] Step S3, Construction of the first-side cold air prediction model:

[0122] To construct the first-side cold air prediction model, this step first uses the average value of the first-side cold air obtained within the specified modeling period. Compared with the second side reference cold air average Linear fitting is used to construct a bilateral cold air baseline relationship between the first-side cold air mean value and the second-side reference cold air mean value in the same channel, same orbit and same scan period in the microwave radiometer. Then, a first-side cold air prediction model is constructed based on the obtained bilateral cold air baseline relationship.

[0123] The specific relationship between the two sides of the cold air reference is as follows:

[0124] ;

[0125] in: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below; Indicates the slope of a linear mapping; Represents the intercept of a linear mapping; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side references the average value of cold air.

[0126] It should be noted that during the fitting process, to reduce the impact of a small number of outliers on the results, an iterative outlier removal method is used to establish the final baseline model. In this example, the maximum number of iterations is 4, and the outlier removal threshold for each round is 4.0 times the standard deviation. The final output baseline model parameters include the linear mapping slope. Linear mapping intercept and the standard deviation of the first-sided mean residual (That is, the background standard deviation of the mean deviation of the first side cold air).

[0127] Based on the aforementioned bilateral cold air baseline relationship, the first-side cold air prediction model of this invention includes two calculation formulas, one for calculating the first-side cold air prediction mean and the other for calculating the prediction values ​​of each channel of the first-side cold air, as follows:

[0128] ;

[0129] ;

[0130] In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast mean; Represents the slope of a linear mapping; Represents the intercept of a linear mapping; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air average; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Predicted values ​​for each field of view; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Observations from each field of view.

[0131] Step S4, statistical background distribution characteristics:

[0132] For each scanning cycle of the first-side cold air modeling observation sample and the second-side reference cold air field set, the following background distribution characteristics are statistically analyzed: the mean deviation of the first-side cold air, the residual of each field of view of the first-side cold air, and the deviation of each field of view of the second-side reference cold air.

[0133] Specifically, the background distribution features mentioned above are defined as follows:

[0134] No. The first channel, the first Orbital-like structures in the scan cycle Deviation of the first side cold air mean Defined as:

[0135] ;

[0136] in, Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below, This indicates that, under the corresponding scanning period, the cold air prediction model on the first side uses the corresponding average cold air value from the second side reference cold air. The calculated mean value of the first-side cold air prediction.

[0137] No. The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air individual field-of-view residuals Defined as:

[0138] ;

[0139] in, Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Observations from each field of view; This indicates that, under the corresponding scanning period, the cold air prediction model on the first side predicts the cold air through the corresponding second-side reference cold air. Observations of each field of view The first side cold air was calculated The predicted value for each field of view.

[0140] No. The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below individual field deviation Defined as:

[0141] ;

[0142] in, Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Observations from each field of view This represents the average value of the second-side reference cold air under the corresponding scanning period.

[0143] Based on the above definitions, the mean deviation of the first-side cold air, the residual of each field of view of the first-side cold air, and the deviation of each field of view of the second-side reference cold air can be obtained under each scanning cycle within the specified modeling period. Combined with the internal dispersion of the second side obtained in step S2, this prepares for the subsequent construction of anomaly judgment thresholds and judgment conditions.

[0144] Step S5: Construct the anomaly detection threshold and detection conditions:

[0145] In this step, based on the internal dispersion of the second side obtained in step S2 and the background distribution characteristics obtained in step S4, anomaly detection thresholds and detection conditions are constructed.

[0146] Step S5.1: Define background statistics:

[0147] To clarify the correspondence between the anomaly detection threshold and background distribution characteristics, the following background statistics are first defined: first-side mean, residual, and standard deviation. First side cold air Background standard deviation of individual field-of-view residuals Background mean of the internal dispersion of the second side Background standard deviation of internal dispersion on the second side and the background standard deviation of the deviation of each field of view of the second-side reference cold air. ,in:

[0148] First side mean residual standard deviation This refers to the deviation of the mean value of the first-side cold air under each scanning cycle within a specified modeling period. Standard deviation; first side cold air Background standard deviation of individual field-of-view residuals To model the first side cold air under each scan cycle within a specified modeling period individual field-of-view residuals Standard deviation; background mean of internal dispersion of the second side To determine the internal dispersion of the second side under each scan cycle within a specified modeling period. The average value; the background standard deviation of the internal dispersion of the second side. To determine the internal dispersion of the second side under each scan cycle within a specified modeling period. Background standard deviation, background standard deviation of deviation of each field of view of the second-side reference cold air. The deviation of the second-side reference cold space from the field of view for each scanning cycle within a specified modeling period. The background standard deviation.

[0149] The formulas for calculating each background statistic are as follows:

[0150] ;

[0151] ;

[0152] ;

[0153] ;

[0154] ;

[0155] In the above formula: This represents the statistical function of standard deviation; This represents a statistical function for the average value.

[0156] Step S5.2: Construct anomaly detection threshold:

[0157] Based on the background statistics calculated in step S5.1, anomaly detection thresholds are constructed. In this invention, the anomaly detection thresholds include a first-side mean anomaly threshold, anomaly thresholds for each field of view of the first-side cold air, a second-side internal dispersion alarm threshold, and anomaly thresholds for each field of view of the second-side reference cold air. Specifically, the first-side mean anomaly threshold is calculated using the standard deviation of the first-side mean residual; the anomaly thresholds for each field of view of the first-side cold air are calculated using the background standard deviation of the residuals of the corresponding fields of view of the first-side cold air; the second-side internal dispersion alarm threshold is calculated by jointly considering the background mean and background standard deviation of the second-side internal dispersion; and the anomaly thresholds for each field of view of the second-side reference cold air are calculated using the background standard deviation of the deviation of the corresponding fields of view of the second-side reference cold air. Each is specifically expressed by the following formula:

[0158] ;

[0159] ;

[0160] ;

[0161] ;

[0162] in, Indicates the first The first channel, the first The threshold for outlier values ​​on the first side of the orbital-like structure; Indicates the first The first channel, the first Standard deviation of the first-side mean residuals of the orbital class; Indicates the first side cold air Anomaly threshold for each field of view, First side cold air Background standard deviation of residuals for each field of view; Indicates the first The first channel, the first The internal dispersion alarm threshold on the second side of the orbital type; Indicates the first The first channel, the first The background mean of the internal dispersion of the second side of the orbital-like structure; Indicates the first The first channel, the first Background standard deviation of the internal dispersion of the second side of the orbital class; Indicates the second side reference cold air. Anomaly threshold for each field of view, Indicates the second side reference cold air. The background standard deviation of the field of view deviation. , , and These represent the first to fourth threshold coefficients, respectively. In this embodiment of the invention, the values ​​of the first to fourth threshold coefficients are all 4.0.

[0163] The determination conditions constructed in this invention include conditions for determining a first-side cold air anomaly, conditions for determining a second-side reference cold air alarm, and reconstruction determination conditions (conditions for determining reconstruction in the current scan cycle), wherein:

[0164] The conditions for determining an anomaly in the first-side cold air are:

[0165] When the absolute value of the deviation of the mean of the first-side cold air exceeds the anomalous threshold of the first-side mean, or when the absolute value of the residual of at least one field of view in each field of view of the first-side cold air exceeds the anomalous threshold of that field of view, that is:

[0166] when If the first-side cold air condition is normal, it is determined that the cold air condition on the first side is abnormal; otherwise, it is determined that the cold air condition on the first side is normal. Therefore, the definition of an abnormal cold air condition on the first side is: an abnormal mean value on the first side, or an abnormality in at least one field of view on the first side.

[0167] The conditions for determining a second-side reference cold air alarm are:

[0168] When the internal dispersion of the second side exceeds the internal dispersion alarm threshold of the second side, or when the absolute value of the deviation of at least two fields of view in the second-side cold air reference field of view from the mean of the second-side reference cold air exceeds the anomaly threshold of the corresponding field of view, that is:

[0169] when or If the current second-side reference cold air is not detected, it is determined that there is no alarm; otherwise, it is determined that there is no alarm. The brackets indicate Iverson. Therefore, the second-side reference cold air warning indicator is defined as: an anomaly inside the second side, or an anomaly in at least two fields of view on the second side.

[0170] The conditions for determining that the current scan cycle meets the reconstruction criteria are: the first side cold air is abnormal and the second side reference cold air does not trigger an alarm.

[0171] Step S6: Determine whether reconstruction is needed in the current scan cycle.

[0172] Based on the anomaly determination threshold and determination conditions constructed in step S5, each scanning cycle of the observation sample to be processed in the first side cold air is determined, and based on the determination results, it is determined whether the observation sample of the corresponding scanning cycle needs to be reconstructed, and the final result is output.

[0173] Specifically, this step of determining each scan cycle of the first-side cold air observation sample to be processed includes the following steps:

[0174] S6.1 Based on the observation samples to be processed in the first side cold air and the full scan observation data, for each channel and each orbit type, the observation values ​​of each field of view of the first side cold air and the observation values ​​of each field of view of the second side reference cold air under each scan cycle are obtained one by one, and then the average value of the first side cold air and the average value of the second side reference cold air under the corresponding scan cycle are calculated.

[0175] S6.2. Using the first-side cold air prediction model, the corresponding first-side cold air prediction mean is calculated by using the second-side reference cold air mean under each scanning period.

[0176] S6.3 Based on the observed values ​​of each field of view of the first side cold air, the observed values ​​of each field of view of the second side reference cold air, the mean value of the first side cold air, the mean value of the second side reference cold air, and the predicted mean value of the first side cold air under the same scanning period, calculate the deviation of the mean value of the first side cold air, the residual of each field of view of the first side cold air, the internal dispersion of the second side, and the deviation of each field of view of the second side reference cold air under the corresponding scanning period.

[0177] S6.4. Compare the calculated mean deviation of the first-side cold air, the residuals of each field of view of the first-side cold air, the internal dispersion of the second-side cold air, and the field of view deviation of the second-side reference cold air under each scanning cycle with the anomaly judgment threshold constructed in step S5. Determine whether the first-side cold air is abnormal and whether the second-side reference cold air is alarmed in the current scanning cycle based on the judgment conditions. Only when the first-side cold air is abnormal and the second-side reference cold air is not alarmed is the current scanning cycle determined to meet the reconstruction conditions, and proceed to step S7; when the first-side cold air is normal, retain the original first-side cold air observation and do not perform reconstruction; when the first-side cold air is abnormal but the second-side reference cold air is alarmed at the same time, the current scanning cycle is determined not to meet the reconstruction conditions, reconstruction is not performed, and the corresponding anomaly identifier is output.

[0178] The final output includes: the reconstructed first-side cold air mean, the reconstructed first-side cold air multi-field observation sequence, the first-side anomaly indicator, the second-side reference cold air alarm indicator, and the reconstruction indicator;

[0179] When the first-side cold air is abnormal and the second-side reference cold air does not alarm, the output reconstruction flag is: Reconstruction valid;

[0180] When the cold air on the first side is abnormal and the reference cold air on the second side is alarmed, the output reconstruction identifier is: abnormal;

[0181] When the first side of the cold air is normal, the output reconstruction identifier is: Original cold air is valid.

[0182] Step S7: Reconstruct the first-side cold air and connect it to the original calibration process:

[0183] When the current scanning cycle meets the reconstruction conditions, the predicted mean of the first-side cold air, calculated from the second-side reference cold air mean using the first-side cold air prediction model, is used to reconstruct the observation values ​​of each field of view of the first-side cold air. Specifically, the first-side cold air mean is first reconstructed using the first-side cold air prediction mean to obtain the reconstructed first-side cold air mean. Then, based on the consistency of the mean correction, the observations of each field of view of the first-side cold air are reconstructed to obtain the reconstructed first-side multi-field cold air observation sequence. The original first-side cold air observation sequence (i.e., the observations of each field of view of the first-side cold air before reconstruction, which can be simply referred to as the original first-side cold air observations) is then replaced by the obtained reconstructed first-side cold air multi-field observation sequence, while maintaining the original calibration process unchanged, and continues to work with the thermal reference observations to complete the subsequent calculation of observation products. In other words, this embodiment does not change the original calibration framework, but improves the cold reference input entering the original calibration process by reconstructing the cold air.

[0184] The reconstructed first-side multi-field cold sky observation sequence was obtained through the following steps:

[0185] First, define the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air mean after reconstruction for:

[0186] ;

[0187] That is, the cold air prediction model on the first side is directly used based on the average value of the cold air reference on the second side. The calculated mean value of the first-side cold air prediction is used as the mean value of the reconstructed first-side cold air.

[0188] Secondly, to maintain the original multi-field structure of the first-side cold air, the mean correction amount of the first-side cold air is defined as follows:

[0189] ;

[0190] Then, this correction amount is added back to the observations of each field of view of the first-side cold air to obtain the reconstructed first-side multi-field cold air observation sequence:

[0191] ;

[0192] In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air One field of view reconstruction value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view observation value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air mean correction amount.

[0193] Through the above processing, the reconstructed first-side multi-field cold-sky observation sequence retains the multi-field structure of the original first-side cold-sky observation sequence, and restores its overall mean to the normal background relationship obtained by the constraint of the second-side reference cold-sky field of view set. For scan periods not judged as abnormal, the original first-side cold-sky observation sequence is retained unchanged.

[0194] In this embodiment, Figure 2 and Figure 3 The calibration antenna temperature results for Channel 1 under the orbital ascent conditions on September 26, 2025. The data processing parameters used in this embodiment are set as follows: the field of view of the first cold sky is selected as L66, L67, L68, and L69, and the field of view of the second reference cold sky is selected as R172, R173, R174, and R175. The field of view of L66 of the first cold sky corresponds to the field of view of R172 of the second reference cold sky, the field of view of L67 of the first cold sky corresponds to the field of view of R173 of the second reference cold sky, the field of view of L68 of the first cold sky corresponds to the field of view of R174 of the second reference cold sky, and the field of view of L69 of the first cold sky corresponds to the field of view of R175 of the second reference cold sky. The first to fourth threshold coefficients are all set to 4.0. The specified modeling period is taken from the observation period when there is no lunar contamination or other obvious anomalies and the correspondence between the two cold skies is stable. Currently, it is selected as July 12, 2025 to August 13, 2025.

[0195] Figure 2The calibration antenna temperature results of CAMS microwave radiometer channel 1, carried on the Tianyan-16 commercial small satellite platform, before correction under ascent conditions, were disclosed during the lunar contamination period (September 26, 2025). Figure 3 The temperature results of the calibration antenna after correction using the cold air reconstruction calibration method described in this invention are disclosed under the same time period, same channel, and same orbit conditions. Figure 2 The results show that, due to the influence of lunar contamination and the cold space reference anomaly on the first side, the calibration antenna temperature results show a localized area of ​​abnormally low temperature, exhibiting obvious cold anomaly characteristics. Figure 3 The results show that after reconstructing the first-side cold space using this invention, the localized areas of abnormally low brightness temperature in the pre-correction calibration antenna temperature results are significantly repaired, the original cold anomaly characteristics are essentially eliminated, and the overall spatial distribution is restored from a discontinuous and uneven state due to contamination to a relatively smooth and consistent distribution pattern, while better preserving the spatial structure characteristics of the original Earth scene. This indicates that this invention can effectively suppress the impact of cold reference anomalies caused by lunar contamination on subsequent calibration results (abnormally low brightness temperature), improving the stability and usability of the calibration results.

[0196] Another technical objective of this invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described microwave radiometer cold space reconstruction calibration method based on bilateral cold space observations.

[0197] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A microwave radiometer cold-space reconstruction calibration method based on bilateral cold-space observations, characterized in that, Includes the following steps: Step S1: Obtain the first side cold air observation sample to be processed and the full scan observation data; Construct first-side cold air modeling observation samples for a specified modeling period based on the first-side cold air observation samples to be processed; The second-side reference cold air field set was selected based on the full-scan observation data; Step S2: Construct cold air features for each scanning cycle based on the first-side cold air modeling observation samples and the second-side reference cold air field of view set; The cold air characteristics include the mean cold air on the first side, the mean cold air reference on the second side, and the internal dispersion of the second side; Step S3: Linearly fit the first-side cold air mean value and the second-side reference cold air mean value under the same scanning period obtained in step S2 to construct the first-side cold air prediction model. Then, under the specified modeling period, calculate the corresponding first-side cold air prediction mean value based on each second-side reference cold air mean value, and calculate the corresponding first-side cold air prediction value based on the observation values ​​of each field of view of the second-side reference cold air in the second-side reference cold air field of view set. Step S4: Based on the mean value of the first side cold air and the mean value of the second side reference cold air obtained in step S2, and the mean value of the first side cold air prediction and the predicted value of each field of view of the first side cold air obtained in step S3, the background distribution characteristics of the first side cold air modeling observation sample and the second side reference cold air field of view set are statistically analyzed in each scanning cycle. Step S5: Construct anomaly detection threshold and detection conditions based on background distribution characteristics and the internal dispersion of the second side; Step S6: Based on the anomaly detection threshold and detection conditions, each scanning cycle of the observation sample to be processed in the first side cold air is detected, and based on the detection results, it is determined whether the observation sample of the corresponding scanning cycle needs to be reconstructed and replaced. After the process is completed, the subsequent calibration process begins.

2. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 1, characterized in that, In step S1, when constructing the first side cold air modeling observation sample, based on the first side cold air observation sample to be processed, the observation samples of each field of view of the first side cold air under the specified modeling period are obtained for each channel and each orbit type, and the first side cold air modeling observation sample can be constructed. When selecting the second-side reference cold air field of view set, based on full-scan observation data, from the non-Earth observation area opposite to each field of view of the first-side cold air, for each channel and each orbit type, a set of continuous scan positions that meet the requirements of internal consistency, temporal stability, and stability of the relationship with the mean of the first-side cold air are selected within a specified modeling period as the second-side reference cold air field of view set. This is obtained through the following steps: During the specified modeling period, the average observation value is calculated for each scan position of the full scan observation data, and the Earth observation area and non-Earth observation area are identified by combining the scan position distribution; in the non-Earth observation area opposite to each field of view of the first side cold air, candidate areas are determined; for the candidate continuous scan position set in the candidate area, the internal consistency, temporal stability and the stability of the relationship with the mean of the first side cold air are evaluated respectively, and among the candidate continuous scan position sets that meet the requirements of internal consistency and temporal stability, the set with the smallest residual dispersion after establishing a benchmark relationship with the mean of the first side cold air is selected as the second side reference cold air field of view set.

3. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 2, characterized in that, In step S2, the mean value of the first side cold air, the mean value of the second side reference cold air, and the internal dispersion of the second side are calculated using the following formulas: ； ； ； In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below; For the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Observations from each field of view This represents the field of view of the first side of the cold air. Indexed by field of view number. ; Indicates the first The first channel, the first Orbital-like structures during scan cycle The second side reference cold air average; Indicates the first The first channel, the first Orbital-like structures during scan cycle The second side reference cold air below Each field of view observation value; For the first The first channel, the first Orbital-like structures in the scan cycle The internal dispersion of the second side below.

4. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 3, characterized in that, In step S3, the first-side cold air prediction model is expressed as: ； ； In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast mean; Indicates the slope of a linear mapping; Represents the intercept of a linear mapping; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air average; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Predicted values ​​for each field of view; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Observations from each field of view.

5. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 4, characterized in that, In step S4, the statistically obtained background distribution features include the mean deviation of the first-side cold air, the residuals of each field of view of the first-side cold air, and the field of view deviations of the second-side reference cold air, wherein: The formula for calculating the deviation of the mean value of the first-side cold air is as follows: ； The formulas for calculating the residuals of each field of view in the first side cold air are as follows: ； The formulas for calculating the deviation of each field of view of the second-side reference cold air are as follows: ； In the above formula, Indicates the first The first channel, the first Orbital-like structures in the scan cycle The deviation of the mean value of the first side cold air below; This represents the average value of the cold air on the first side within the corresponding scanning period. Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast mean; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view residual; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view observation value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air One field of view prediction value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Individual field of view deviation; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below Observations from each field of view Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side references the average value of cold air.

6. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 5, characterized in that, In step S5, the anomaly detection threshold includes the first-side mean anomaly threshold. First side cold air Anomaly threshold for each field of view Second-side internal dispersion alarm threshold And the second side reference cold air Anomaly threshold for each field of view They are expressed by the following formulas respectively: ； ； ； ； In the above formula: Indicates the first threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle Deviation of the first side cold air mean The background standard deviation; Indicates the second threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air individual field-of-view residuals The background standard deviation; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side internal dispersion The background mean; Indicates the third threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side internal dispersion The background standard deviation; This represents the fourth threshold coefficient; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The second side reference cold air below individual field deviation The background standard deviation; The judgment criteria include the criteria for determining the first-side cold air anomaly and the criteria for determining the second-side reference cold air alarm, wherein: The conditions for determining an anomaly in the first-side cold air are: ; The conditions for determining a second-side reference cold air alarm are: or .

7. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 6, characterized in that, In step S6, when determining each scan cycle of the first-side cold air observation sample to be processed, the following steps are specifically included: S6.1 Based on the observation samples to be processed in the first side cold air and the full scan observation data, for each channel and each orbit type, the observation values ​​of each field of view of the first side cold air and the observation values ​​of each field of view of the second side reference cold air under each scan cycle are obtained one by one, and then the average value of the first side cold air and the average value of the second side reference cold air under the corresponding scan cycle are calculated. S6.

2. Using the first-side cold air prediction model, the corresponding first-side cold air prediction mean is calculated by using the second-side reference cold air mean under each scanning period. S6.3 Calculate the mean deviation of the first side cold air, the residual of each field of view of the first side cold air, the internal dispersion of the second side, and the deviation of each field of view of the second side reference cold air under each current scanning cycle; S6.

4. Compare the mean deviation of the first side cold air, the residual of each field of view of the first side cold air, the internal dispersion of the second side, and the deviation of each field of view of the second side reference cold air under each current scanning cycle with the anomaly judgment threshold constructed in step S5. Determine whether the first side cold air is abnormal and whether the second side reference cold air is alarmed under the current scanning cycle by the judgment conditions.

8. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 7, characterized in that, Step S6 is followed by a reconstruction step: Only when the determination result indicates that the first-side cold air is abnormal in the current scanning period and the second-side reference cold air does not alarm, it is necessary to reconstruct the observation samples of the corresponding scanning period to obtain the reconstructed first-side multi-field cold air observation sequence, and replace the original first-side cold air field observation sequence with the reconstructed first-side multi-field cold air observation sequence to enter the subsequent calibration process. The reconstructed first-side multi-field cold air observation sequence is calculated using the following formula: ； ； ； In the above formula: Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air One field of view reconstruction value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side of the cold air Each field of view observation value; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air mean correction amount; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side cold air after reconstruction; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The average value of the first side of the cold air below; Indicates the first The first channel, the first Orbital-like structures in the scan cycle The first side cold air forecast mean.

9. The microwave radiometer cold-space reconstruction calibration method based on dual-sided cold-space observations according to claim 8, characterized in that, The final output includes: the reconstructed first-side cold air mean, the reconstructed first-side cold air multi-field observation sequence, the first-side anomaly indicator, the second-side reference cold air alarm indicator, and the reconstruction indicator; When the first-side cold air is abnormal and the second-side reference cold air does not alarm, the output reconstruction flag is: Reconstruction valid; When the cold air on the first side is abnormal and the reference cold air on the second side is alarmed, the output reconstruction identifier is: abnormal; When the first side of the cold air is normal, the output reconstruction identifier is: Original cold air is valid.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the microwave radiometer cold space reconstruction calibration method based on bilateral cold space observation as described in any one of claims 1-9.

Images