阅读说明：本技术 一种基于多尺度峰值监测的雷达高度计波形重跟踪方法 (Radar altimeter waveform retracing method based on multi-scale peak monitoring ) 是由 廖静娟 陈嘉明 于 2020-07-22 设计创作，主要内容包括：本发明公开了一种基于多尺度峰值监测的雷达高度计波形重跟踪方法,包括以下步骤：获取雷达高度计对水体的过境观测数据；从所述雷达高度计对水体的过境观测数据提取多尺度自适应子波形；对于每个所述多尺度自适应子波形,采取阈值子波形重跟踪算法计算对应的多尺度重跟踪高程值；从所述多尺度重跟踪高程值中去除异常高程值；采用最短路径算法确定每个观测点的最优重跟踪点的高程值。本发明的方法通过一种普适的统计方法来达到既提高多波峰波形的处理能力,又可在多种高度计数据中进行应用的效果,同时,解决现有波形重跟踪方法受限于周边区域地势影响的情况,使波形重跟踪方法既可用于山地湖泊亦可用于平原湖泊,这对于在大空间尺度进行湖泊水位变化监测具有重要意义。(The invention discloses a radar altimeter waveform retracing method based on multi-scale peak value monitoring, which comprises the following steps of: acquiring transit observation data of the radar altimeter on a water body; extracting a multi-scale self-adaptive sub-waveform from the transit observation data of the radar altimeter on the water body; for each multi-scale self-adaptive sub-waveform, calculating a corresponding multi-scale re-tracking elevation value by adopting a threshold sub-waveform re-tracking algorithm; removing abnormal elevation values from the multi-scale re-tracking elevation values; and determining the elevation value of the optimal retracing point of each observation point by adopting a shortest path algorithm. The method of the invention achieves the effects of improving the processing capacity of multi-peak waveforms and being applied to various altimeter data through a universal statistical method, and simultaneously solves the problem that the existing waveform retracing method is limited by the influence of the topography of the peripheral area, so that the waveform retracing method can be used for mountainous lakes and plain lakes, and has important significance for monitoring the water level change of lakes in a large space scale.)

1. A radar altimeter waveform retracing method based on multi-scale peak monitoring comprises the following steps:

s1: acquiring transit observation data of the radar altimeter on a water body;

s2: extracting a multi-scale self-adaptive sub-waveform from the transit observation data of the radar altimeter on the water body;

s3: for each multi-scale self-adaptive sub-waveform, calculating a corresponding multi-scale re-tracking elevation value by adopting a threshold sub-waveform re-tracking algorithm;

s4: removing abnormal elevation values from the multi-scale re-tracking elevation values;

s5: and determining the elevation value of the optimal retracing point of each observation point by adopting a shortest path algorithm.

2. The method of claim 1, wherein: and step S6, acquiring the mean water level of the whole-track data according to the elevation value of the optimal retracing point of each observation point.

3. The method of claim 1, wherein: the abnormal elevation values in the step S4 include obvious abnormal value water level data and invalid far-point water level data.

4. The method of claim 1, wherein: the radar altimeter is a radar altimeter of a Cryosat-2 or Sentine1-3 satellite.

5. The method of claim 1, wherein: the step S2 is to extract a multi-scale adaptive sub-waveform from the transit observation data of the radar altimeter to the water body based on a local extremum matrix method; the local extremum matrix method includes:

constructing an L multiplied by N local extremum matrix (m) for each echo waveform data_k，i)_L×NCalculating local extrema of the waveform data by a window moving average, wherein the window length w_kThe variation range satisfies { w_k2k +1| k ═ 1,2,.., L }, with a window length of w_kIs determined according to equation (1):

wherein r is in [0,1 ]]Random number of range change, N is length of signal, L is maximum window length of moving average, when L is 5, it is optimum value of waveform peak value detection, m_k，iIn addition to the assignment of 0 to the range of i-k + 2., N-k +1, the remaining portions are assigned r +1, on the basis of which the local extremum matrix M (M) can be constructed_k，i)_L×N；

Calculating local extreme momentsThe sum γ of the rows of the array M ═ γ₁，γ₂，...，γ_L]And calculating the line lambda corresponding to the minimum value in gamma as argmin (gamma)_k) Representing the dimension containing the most local maxima. Then, local extreme value matrixes are recombined, and all elements m meeting the condition that k is larger than lambda are removed_k，iTo obtain a new matrix M_r＝(m_k，i)_λ×N；

Peak detection of waveform by calculating M using equation (2)_rStandard deviation σ of each column of the matrix_iIs determined, wherein σ is satisfied_iThe index i which is 0 is an index gate of the detected peak, and this value is set as the end gate stopgatei of the sub-waveform, whereby one end gate vector can be obtained for each echo waveform

Dividing the waveform data by the maximum power for normalization processing; stop pgate for each end-gate position_iSearching the first derivative of the normalized waveform from back to front, wherein the first wave gate with the power value lower than 0.001 is the corresponding start wave gate position start gate_iBased on the method, all the ending wave gates in stoprate are searched in the same way, and finally, the position vector of the starting wave gate of the sub-waveform can be obtained through calculation

6. The method of claim 5, wherein: step S3 adopts threshold sub-waveform retracing algorithm to the

wherein P is_iFor the echo waveform power value, M is the amplitude of the rectangle surrounding the multi-scale adaptive sub-waveform.

7. The method of claim 3, wherein: in step S4, the significant outlier level data is an outlier of the multi-scale re-tracking elevation value that exceeds three times the standard deviation of the mean of the point cloud data.

8. The method of claim 7, wherein: the apparent outlier water level data removal loop is performed three times for outliers exceeding three times the standard deviation of the mean of the point cloud data.

9. The method of claim 3, wherein: in step S4, rounding all elevation values in the point cloud data, and establishing a Cumulative Distribution Function (CDF) of the rounded elevation values, so as to obtain an elevation value corresponding to a minimum value of a second-order difference quotient of the CDF, and setting the elevation value as an elevation threshold; and combining the difference between the average value of the plurality of height values contained in each sub-satellite point and the height threshold value, and if the height threshold value is greater than the set threshold value of the far sub-satellite point, considering the corresponding sub-satellite point as the far sub-satellite point, and removing the height value of the far sub-satellite point as invalid water level data of the far sub-satellite point.

10. The method of claim 1, wherein: the shortest path algorithm in the step S5 is Dijkstra algorithm.

Technical Field

The invention relates to the field of remote sensing water level inversion, in particular to a radar altimeter waveform retracing method based on multi-scale peak value monitoring.

Background

Satellite radar altimetry is an important tool for monitoring the water level of ocean and inland water bodies, especially in areas with no or few measured data. Altimeters often measure the water level of inland bodies of water subject to the influence of the land echo signals, which can usually be improved by waveform retracing (wave retracking), i.e. correcting the initial distance calculated from the pulse echo two-way time. Wave shape retracing of inland water has achieved a lot of research results, and the following are some classical methods: NPPR (the nano rectangular Peak tracker) for monitoring echo main peaks based on the maximum reflection energy of the water surface is used for retracing algorithm (Jain M, Andersen O B, Dall J, et al. Sea surface height estimation in the array using Cryosa-2 SAR data from Primary peak detectors [ J ]. Advances in Space Research,2015,55(1):40-50.), MWaPP (the multi wave form periodic Peak) algorithm (villasten H, Deng X, Andersen O B, et al. modified floor waves water level estimation from echo height estimation in water surface) algorithm (horizontal H, depletion X, Andersen O B, SAR. real. modified floor waves from echo peaks) for extracting complex tracking algorithm [ 10J ]. J. (attachment J.: 85-50. for tracking algorithm) for tracking peak height estimation in water surface area, and for tracking algorithm [ 10J., (J.: crack J.),234, and for tracking algorithm (load J.: load J.: 10. for tracking algorithm J, J.: load J.: 10. for tracking algorithm J. (gravity, for tracking algorithm J.), 2018: 112-.

The existing waveform re-tracking method has some limitations, which are mainly reflected in the following aspects: 1) the influence of complex terrain and non-intersatellite point observation on the waveform of the altimeter cannot be fully considered, and the processing capacity of complex multi-wave-crest waveforms is weak; 2) the method has the advantages that the problem of selection of the re-tracking algorithm of different altimeter data exists during the multi-source altimeter water level inversion, corresponding system deviation exists among different re-tracking algorithms, and the accuracy after water level data fusion is influenced to a certain degree; 3) the method is limited by the influence of the topography of the peripheral area, and the existing algorithm cannot give consideration to both mountainous lakes and plain lakes.

Disclosure of Invention

The invention aims to provide a high-precision inland water body elevation extraction algorithm which can be suitable for the influence of various altimeter data and various peripheral topographic factors so as to make up for the defects of the prior art.

The invention provides a radar altimeter waveform retracing method based on multi-scale peak value monitoring, which comprises the following steps of:

s1, acquiring transit observation data of the radar altimeter on the water body;

s2, extracting multi-scale adaptive sub-waveforms from the transit observation data of the radar altimeter to the water body;

s3, for each multi-scale self-adaptive sub-waveform, calculating a corresponding multi-scale re-tracking elevation value by adopting a threshold sub-waveform re-tracking algorithm;

s4, removing abnormal elevation values from the multi-scale re-tracking elevation values;

and S5, determining the elevation value of the optimal retracing point of each observation point by adopting a shortest path algorithm.

Optionally, the method of the present invention further includes step S6, obtaining a mean water level of the whole track data according to the elevation value of the optimal re-tracking point of each observation point.

Optionally, the abnormal elevation values in step S4 include significant abnormal value water level data and invalid far-point water level data.

Optionally, the radar altimeter is a radar altimeter for a Cryosat-2 or Sentinil-3 satellite.

Optionally, step S2 is to extract a multi-scale adaptive sub-waveform from transit observation data of the radar altimeter on the water body based on a local extremum matrix method; the local extremum matrix method includes:

constructing an L multiplied by N local extremum matrix (m) for each echo waveform data_k,i)_L×NCalculating local extrema of the waveform data by a window moving average, wherein the window length w_kVariations inThe range satisfies { w_k2k +1| k ═ 1,2, …, L }, window length w_kIs determined according to equation (1):

wherein r is in [0,1 ]]Random number of range change, N is length of signal, L is maximum window length of moving average, when L is 5, it is optimum value of waveform peak value detection, m_k,iThe local extremum matrix M (M) can be constructed on the basis of the assignment of 0 to the remaining part, except for the assignment of 0 to i (k +2, …) and N-k +1_k,i)_L×N；

Calculating the sum gamma of each row of the local extremum matrix M ═ gamma₁,γ₂,…,γ_L]And calculating the line lambda corresponding to the minimum value in gamma as argmin (gamma)_k) Representing the dimension containing the most local maxima. Then, local extreme value matrixes are recombined, and all elements m meeting the condition that k is larger than lambda are removed_k,iTo obtain a new matrix M_r＝(m_k,i)_λ×N；

Peak detection of waveform by calculating M using equation (2)_rStandard deviation σ of each column of the matrix_iIs determined, wherein σ is satisfied_iThe index i of 0 is the index gate of the detected peak, and this value is set as the end gate stop of the sub-waveform_iWhereby for each echo waveform a final gate vector is obtained

Dividing the waveform data by the maximum power for normalization processing; stop pgate for each end-gate position_iSearching the first derivative of the normalized waveform from back to front, the first power valueThe wave gate below 0.001 is the starting gate position corresponding to the starting wave gate_iBased on the method, all the ending wave gates in stoprate are searched in the same way, and finally, the position vector of the starting wave gate of the sub-waveform can be obtained through calculation

Together with stomate constitute

A plurality of multi-scale adaptive sub-waveforms.

Optionally, step S3 uses a threshold sub-waveform re-tracking algorithm to the data

Carrying out waveform retracing calculation on a plurality of multi-scale self-adaptive sub-waveforms, wherein a given Threshold in a Threshold sub-waveform retracing algorithm is 0.5M;

wherein P is_iFor the echo waveform power value, M is the amplitude of the rectangle surrounding the multi-scale adaptive sub-waveform.

Optionally, in step S4, the significant outlier water level data is an outlier of the multi-scale re-tracking elevation value that exceeds three times a standard deviation of the mean of the point cloud data.

Optionally, in step S4, the significant outlier water level data removal loop is performed three times for outliers exceeding three times the standard deviation of the mean of the point cloud data.

Optionally, in step S4, rounding all elevation values in the point cloud data, and establishing a Cumulative Distribution Function (CDF) of the rounded elevation values, so as to obtain an elevation value corresponding to a minimum value of a second-order difference quotient of the CDF, and setting the elevation value as an elevation threshold; and combining the difference between the average value of the plurality of height values contained in each sub-satellite point and the height threshold value, and if the height threshold value is greater than the set threshold value of the far sub-satellite point, considering the corresponding sub-satellite point as the far sub-satellite point, and removing the height value of the far sub-satellite point as invalid water level data of the far sub-satellite point.

Optionally, the shortest path algorithm in step S5 is Dijkstra algorithm.

The technical scheme of the invention has the beneficial technical effects that: the radar altimeter waveform retracing method based on multi-scale peak monitoring provided by the invention achieves the effects of improving the processing capacity of multi-peak waveforms and being applied to various altimeter data through a universal statistical method, and meanwhile, solves the problem that the existing waveform retracing method is limited by the influence of the terrain of peripheral areas, so that the waveform retracing method can be used for mountainous lakes and plain lakes, and has important significance for monitoring the water level change of lakes in a large spatial scale.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a flow chart diagram of the method of the present invention.

FIG. 2 is a schematic flow diagram of one embodiment of the method of the present invention.

Fig. 3 shows data of a part of special waveforms of a Cryosat-2 satellite transit Qinghai lake on 16 th month 4 in 2016, and a dotted line shows a sampling gate position corresponding to an actually measured water level. Wherein (a) is a far-off intersatellite point waveform influenced by other water bodies, and (b) is a multi-peak waveform influenced by land echo.

Fig. 4 shows transit orbit data of a Cryosat-2 satellite on 16 days 4/4 in the year of the Qinghai lake 2016, a single point is multi-scale re-tracking height point cloud data, a connecting line is a shortest path obtained by calculation by using a Dijkstra algorithm, and a gray filling part is far off-satellite point observation which does not participate in calculation.

FIG. 5 is a schematic diagram of estimating optimal re-tracking points using Dijkstra's algorithm. Dijkstra finds the optimal path by the ordered water level heights (shown in grey).

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of systems consistent with certain aspects of the invention, as detailed in the appended claims.

In the prior art waveform re-tracking process, only one elevation can be acquired per echo waveform. If a plurality of obvious peaks exist in the waveform, the re-tracking of the waveform to an error peak value can be caused, and the measured value is invalid. Because the water level of the lake keeps stable along the track in the same observation period and the surrounding terrain is changed randomly, the problem of solving the optimal re-tracking point can be converted into the shortest path search problem.

Based on the basic thought, the invention develops the waveform sub-waveform extraction, sub-waveform retracing calculation and precise extraction research of lake water level aiming at the waveform data of the radar altimeter. The method of the invention is illustrated primarily using data from Cryosat-2 and Sentinel-3 satellites.

Referring to fig. 1, the radar altimeter waveform retracing method AMPDR (automatic multiscale-based peak detection tracker) based on multi-scale peak monitoring of the invention comprises the following steps:

s1, acquiring transit observation data of the radar altimeter on the water body;

s2, extracting multi-scale adaptive sub-waveforms from the transit observation data of the radar altimeter to the water body;

s3, for each multi-scale self-adaptive sub-waveform, calculating a corresponding multi-scale re-tracking elevation value by adopting a threshold sub-waveform re-tracking algorithm;

s4, removing abnormal elevation values from the multi-scale re-tracking elevation values;

and S5, determining the elevation value of the optimal retracing point of each observation point by adopting a shortest path algorithm.

According to one embodiment of the method of the present invention, the abnormal elevation values in step S4 include significant abnormal water level data and invalid far sub-satellite water level data.

According to an embodiment of the method of the invention, the method further comprises the steps of: and S6, acquiring the average water level of the whole-track data according to the elevation value of the optimal retracing point of each observation point.

The radar altimeter waveform retracing method based on multi-scale peak monitoring provided by the invention achieves the effects of improving the processing capacity of multi-peak waveforms and being applied to various altimeter data through a universal statistical method, and meanwhile, solves the problem that the existing waveform retracing method is limited by the influence of the terrain of peripheral areas, so that the waveform retracing method can be used for mountainous lakes and plain lakes, and has important significance for monitoring the water level change of lakes in a large spatial scale.

Referring to fig. 2, an embodiment of a radar altimeter waveform retracing method based on multi-scale peak monitoring according to the present invention includes the following steps:

A. and acquiring transit observation data of the radar altimeter of the Cryosat-2 or Sentinel-3 satellite on the water body through a lake mask.

B. Method for extracting multi-scale self-adaptive sub-waveforms from transit observation data of radar altimeter on water body based on local extremum matrix

For the transit observation data of the radar altimeter on the water body, the multi-scale self-adaptive sub-waveform is extracted by detecting the waveform peak value by adopting a method based on a local extremum matrix, and the waveform peak value can be detected as much as possible under the condition of not occupying too much memory. As two special waveforms in fig. 3, the horizontal signal reflection point exists among the third wavelet peaks in (a) of fig. 3; wave in (b) of FIG. 3In the form of multi-peak waveform data, the point reflecting the water surface signal appears in the last peak rather than the maximum peak range. The core idea is to construct an LXN local extremum matrix (m) for each echo waveform data_k,i)_L×NThe local pole calculation of the waveform data is calculated by a window moving average, wherein the window length w_kThe variation range satisfies { w_k2k +1| k ═ 1,2, …, L }, window length w_kIs determined according to the following equation:

wherein r is in [0,1 ]]Random number of range change, N is length of signal, L is maximum window length of moving average, when L is 5, it is optimum value of waveform peak value detection, m_k,iThe local extremum matrix M (M) can be constructed on the basis of the assignment of 0 to the remaining part, except for the assignment of 0 to i (k +2, …) and N-k +1_k,i)_L×N。

Calculating the sum gamma of each row of the local extremum matrix M ═ gamma₁,γ₂,…,γ_L]And calculating the line lambda corresponding to the minimum value in gamma as argmin (gamma)_k) Representing the dimension containing the most local maxima. Then, local extreme value matrixes are recombined, and all elements m meeting the condition that k is larger than lambda are removed_k,iTo obtain a new matrix M_r＝(m_k,i)_λ×N。

Peak detection of waveform by calculating M using equation (2)_rStandard deviation σ of each column of the matrix_iIs determined, wherein σ is satisfied_iThe index i of 0 is the index gate of the detected peak, and this value is set as the end gate stop of the sub-waveform_iWhereby for each echo waveform a final gate vector is obtained

And (4) dividing the waveform data by the maximum power for normalization processing. Stop pgate for each end-gate position_iSearching the first derivative of the normalized waveform from back to front, wherein the first wave gate with the power value lower than 0.001 is the corresponding start wave gate position start gate_iBased on the method, all the ending wave gates in stoprate are searched in the same way, and finally, the position vector of the starting wave gate of the sub-waveform can be obtained through calculationTogether with stomate constitute

Considering that the number of wavegates of the sub-waveform is less than 5 in part of data, 1-2 wavegates are required to be expanded back and forth to form a new sub-waveform.

C. For each multi-scale self-adaptive sub-waveform, calculating a corresponding multi-scale re-tracking elevation value by adopting a threshold sub-waveform re-tracking algorithm

Adopting threshold sub-waveform re-tracking algorithm (Jain M, Andersen O B, Dall J, et al. sea surface height determination in the alignment using Cryosat-2 SAR data from primary peak detectors [ J]Advances in Space Research,2015,55(1):40-50.) for step B, respectively

And carrying out waveform retracing calculation on the multiple multi-scale adaptive sub-waveforms. The threshold sub-waveform retracing algorithm is an extension of the COG sub-waveform retracing device, and uses the amplitude parameter M in the OCOG, as shown in equation (3). The purpose of this is to identify the first gate position ithres at which the power exceeds a given Threshold.

Retracing pointC_{retrack_Thres}Obtained by means of interpolation, see formula (4).

Wherein P is_iFor the echo waveform power value, M is the amplitude of the rectangle surrounding the sub-waveform; w is the width of the rectangle surrounding the sub-waveform; threshold is 0.5M, which is the given Threshold in the Threshold re-tracking algorithm, and ithres is the first gate position in the sub-waveform that exceeds Threshold; c_{retrack_Thres}The points are re-tracked for the threshold sub-waveform. And obtaining a corresponding multi-scale re-tracking elevation value for each multi-scale adaptive sub-waveform.

D. Removing obvious abnormal value water level data and invalid far-intersatellite point water level data from multi-scale re-tracking elevation values

Firstly, outliers exceeding three times of standard deviation of the mean value of the point cloud data in the multi-scale re-tracking elevation values are regarded as obvious outlier water level data to be removed. According to one example of embodiment of the present invention, a significant outlier water level data removal loop is performed three times for outliers exceeding three times the standard deviation of the mean of the point cloud data. This is the most efficient in the experimental process and does not delete too many elevations.

And rounding all elevation values in the point cloud data, establishing a Cumulative Distribution Function (CDF) of the rounded elevation values, thereby obtaining an elevation value corresponding to the minimum value of the second-order difference quotient of the CDF, and setting the elevation value as an elevation threshold DistanceThres. Combining the difference between the average value of the plurality of height values contained in each sub-satellite point and the height threshold DistanceThres, if the threshold is greater than 1/2 × 128 × 0.4684 (for example, Sentinel-3), the sub-satellite points are regarded as far sub-satellite points, and the height values of the far sub-satellite points are regarded as invalid sub-satellite water level data to be removed.

E. Determining an elevation value of an optimal re-tracking point by using Dijkstra shortest path algorithm

After removing obvious abnormal value water level data and invalid far-off-satellite water level data, remaining all 'elevation point clouds' are used for finding a shortest path and estimating an elevation value of an optimal re-tracking point at each observation point, as shown in fig. 4, taking Cryosat-2 satellite transit orbit data of 2016, 4, 16, months, of Qinghai lake as an example, a single point is multi-scale re-tracking height point cloud data, a connecting line is the shortest path obtained by resolving through a Dijkstra algorithm, and a gray filling part is far-off-satellite under-point observation which does not participate in calculation.

As shown in fig. 5, a Dijkstra graph (Dijkstra, e.w.e.w.,1959.a node on twoply in connectionis with graph, nut.math.1, 269-271) is constructed in the form of fully connected layers, i.e., each node is connected to all nodes in the previous layer. As shown in fig. 5, all elevation values of each observation point are regarded as one layer, and nodes of each two layers need to be connected with each other; arranging the sequence of each layer outside the first layer and the last layer from small to large according to the latitude of the observation point, wherein the elevation values of DistanceThres are adopted by the starting node and the final node; the Dijkstra method requires selecting edge weights between each connected node, in which method the height difference between the connected nodes is selected as the edge weight.

F. Calculating the mean water level of the whole track data

Calculating the elevation value H of the optimal retracing point of each observation point:

H＝Alt-(R_range+ΔR_m+ΔR_dry+ΔR_wet+ΔR_iono

+ΔR_tide+ΔR_retrack)-N_geoid(5)

wherein Alt is the height from the center of mass of the satellite to the reference ellipsoid, R_rangeIs the distance from the satellite to the sub-satellite point, Δ R_mIs the correction of the center of mass, Δ R, of the satellite_dry、ΔR_wetIs dry and wet tropospheric delay correction, Δ R_ionoIs ionospheric delay correction, Δ R_tideIncluding correction for solid tides, extreme tides and sea tides, Δ R_retrackIs a re-tracking correction. N is a radical of_geoidIs the fluctuation of the geodetic surface relative to the ellipsoid, and the geodetic surface model adopted herein is the earth gravity model 2008(EGM 2008). Except for Δ R_retrackIn addition, all of the above corrections can be read directly in the altimeter data.

After the elevation values of the optimal retracing points of all the observation points along the track are obtained, the elevation values of all the optimal retracing points are averaged to obtain the final average water level of the whole track data.

And (4) analyzing results:

five different re-tracking algorithms were selected for comparison, (1) the threshold levels were 50% and 80% NPPTR algorithms (NPPTR [0.5], NPPTR [0.8) (Jain M, Andersen O B, Dall J, et al. Sea surface height determination in the apparatus using Cryosat-2 SAR data from primary peak detectors [ J ]. Advances in Space Research,2015,55(1): 40-50.); (2) narrow Primary Peak OCOG Reccker (NPPOR); (3) MWaPP algorithm (Villadsen H, Deng X, Andersen O B, et al. improved in-and-water levels from SAR optimization using experimental algorithms [ J ]. Journal of hydrology.2016,537: 234-; (4) Wingham/Wallis fitting Algorithm (ESA L2) (Wingham D, Francis C, Baker S, effective. Cryosat: A mission to determination the structures in Earth' S land and maritime fields [ J ]. Advances in Space research.2006,37(4): 841-4) 871.). (5) SAMOSA algorithm (Ray, C., Martin-Puig, C., Clarizia, M.P., Ruffini, G., Dinardo, S., Gommenginger, C., Benveniste, J.,2015.SAR estimator backscattered waveform model. IEEETrans. Geosci. remote Sens.53, 911-919.), and the precision of different algorithms is compared and analyzed by carrying out precision verification on the measured water level data.

The statistical results are shown in Table 1, and it can be seen that the AMPDR algorithm obtains good results under different altimeter data, and the average Root Mean Square Error (RMSE) of the lakes transiting through Cryosat-2 and Sentinel-3 is the minimum, which is 0.149m and 0.139m respectively. In the Cryosat-2 transit lakes, the best results were obtained except for the Ewing lake and Bambu mistakes, especially the Qinghai lake, RMSE (0.079m) obtained by the AMPDR algorithm; in the lakes crossed by Sentinel-3, the optimal results are obtained in east Ping lake and Zaina Namo, and the optimal results are not obtained in other lakes but are similar to the results of other algorithms.

TABLE 1 comparison of different algorithm precisions

The features and benefits of the present invention are illustrated by reference to the examples. Accordingly, the invention is expressly not limited to these exemplary embodiments illustrating some possible non-limiting combination of features which may be present alone or in other combinations of features.

The above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.