CN114265127A - GNSS-R snow detection method based on support vector machine - Google Patents
GNSS-R snow detection method based on support vector machine Download PDFInfo
- Publication number
- CN114265127A CN114265127A CN202111591910.8A CN202111591910A CN114265127A CN 114265127 A CN114265127 A CN 114265127A CN 202111591910 A CN202111591910 A CN 202111591910A CN 114265127 A CN114265127 A CN 114265127A
- Authority
- CN
- China
- Prior art keywords
- signal
- gnss
- noise ratio
- support vector
- ratio data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Position Fixing By Use Of Radio Waves (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides a GNSS-R snow detection method based on a support vector machine, which comprises the following steps: selecting an experimental area; respectively collecting satellite navigation signal-to-noise ratio data of the accumulated snow on the ground and the satellite navigation signal-to-noise ratio data of the snow on the ground when the ground is exposed; processing the acquired signal-to-noise ratio data according to the GNSS orbit parameters; and selecting the signal-to-noise ratio data according to the quality control requirement and using the signal-to-noise ratio data as an input sample of a support vector classifier to classify and predict whether the accumulated snow exists on the ground. The GNSS-R snow detection method based on the support vector machine can realize high-precision ground snow detection.
Description
Technical Field
The invention relates to the technical field of snow detection, in particular to a GNSS-R snow detection method based on a support vector machine.
Background
Snow is an important fresh water resource and an important component of the earth ecological environment. It is closely related to human survival and social development. With global warming, snowfall will also be affected as part of the freezing circle and climate cycle, and severe snowfall can cause traffic congestion, crop growth difficulties and ecological imbalance. The research on the depth of accumulated snow is beneficial to the research on surface climate and hydrology and is also an important index for snow melting runoff forecasting and snow disaster prevention. Therefore, the snow depth data can be acquired quickly with high precision, so that the snow depth data is not only beneficial to the survival and development of human beings, but also beneficial to monitoring the change of the nature, and has great significance.
The Global Navigation Satellite System (GNSS) has the characteristics of all weather, high resolution, high precision and near real-time performance, is widely applied to services such as time service, communication, positioning and the like, and can also be used for inverting the depth of accumulated snow according to the relation between the modulation frequency of a satellite navigation signal-to-noise ratio signal and the vertical reflection height. The existing research results obtain better inversion accuracy when the snow depth is larger, but the inversion error is larger when the snow depth is smaller, because the monitoring area has rugged terrain and covers such as vegetation, stones and the like possibly existing around generate interference on signal-to-noise ratio data. The existing manual means can not observe abnormal data from the collected signal-to-noise ratio data, so that the accuracy of the inversion result of the snow depth is limited.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a GNSS-R snow detection method based on a support vector machine, which can realize high-precision ground snow detection.
In order to solve the problems, the technical scheme of the invention is as follows:
a GNSS-R snow detection method based on a Support Vector Machine (SVM), the method comprising the steps of:
selecting an experimental area;
respectively collecting satellite navigation signal-to-noise ratio data of the accumulated snow on the ground and the satellite navigation signal-to-noise ratio data of the snow on the ground when the ground is exposed;
processing the acquired signal-to-noise ratio data according to the GNSS orbit parameters;
and selecting signal-to-noise ratio data as an input sample of the support vector classifier to perform classification prediction according to the quality control requirement.
Optionally, the separately collecting of the snow cover and the bare groundIn the step of satellite navigation signal-to-noise ratio data in the ground, the signal-to-noise ratio data is expressed as:wherein the content of the first and second substances,andrespectively representing the direct signal power and the reflected signal power phiγIs the phase delay between the direct signal and the reflected signal.
Optionally, the phase delay between the direct signal and the reflected signal is phiγCan be derived from the path delay delta between the direct signal and the reflected signal, said phase delay phiγThe expression of (a) is:wherein λ is the wavelength of the navigation system signal, h is the vertical height from the phase center of the GNSS antenna to the reflecting surface, and e is the satellite elevation angle, i.e., the included angle between the direct satellite signal and the reflecting surface.
Optionally, in the step of processing the acquired signal-to-noise ratio data according to the GNSS orbit parameters, the signal-to-noise ratio may be represented as:i.e., the signal-to-noise ratio data varies with the variation of the satellite elevation sinusoid.
Optionally, the step of selecting appropriate signal-to-noise ratio data as an input sample of the support vector classifier for high-dimensional feature space mapping according to the quality control requirement, and constructing an optimal hyperplane which can be used as a decision boundary for identification and classification specifically includes: the support vector classifier constructs a free hyperplane in a selected feature space to maximize a classification interval, and an optimization model is expressed as:
wherein x represents a sample, phi (x) represents a new vector after x is mapped to a new feature space, xi is a relaxation variable, and C is a penalty parameter.
Optionally, by solving for the optimal separation hyperplane wTPhi (x) + b is 0, the classification decision function, i.e. the classification model, is obtained: (x) sign (w)Tφ (x) + b), wherein sign is a step function.
Compared with the prior art, the GNSS-R snow detection method based on the support vector machine is used for researching a ground snow detection method in a global navigation satellite system (GNSS-R) reflection signal.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
FIG. 1 is a block diagram of a GNSS-R snow detection method based on a support vector machine according to an embodiment of the present invention;
FIG. 2 is a geometric model diagram of a GNSS-R technique according to an embodiment of the present invention;
fig. 3 is a graph of the signal-to-noise ratio of snow and bare ground provided by an embodiment of the present invention.
Detailed Description
The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.
Specifically, fig. 1 is a flowchart of a GNSS-R snow detection method based on a support vector machine according to an embodiment of the present invention, and as shown in fig. 1, the method includes the following steps:
s1: selecting an experimental area;
specifically, the P351 observation site of the plate boundary observation plan in the united states is located in kanehem (north latitude 43.87441 °, west longitude 114.71916 °, elevation 2692.6m) in monta, usa. The P351 station has a longer snow accumulation period and a larger snow accumulation depth change which can reach more than 1.5m at most, and is a suitable place for performing a snow depth inversion experiment. The receiver at the P351 site recorded observations at 15s sampling intervals using a Trimble TRM 29659.00 antenna, with an antenna height of 2 m. The SNOTEL network Galena Summit meteorological station (north latitude 43.87497 °, west longitude 114.71363 °, altitude 2676m) provides actual snow depth data of the P351 station, 1.8km from the P351 station.
S2: respectively collecting satellite navigation signal-to-noise ratio data of the accumulated snow on the ground and the satellite navigation signal-to-noise ratio data of the snow on the ground when the ground is exposed;
the signal-to-noise ratio acquired by the GNSS receiver can be expressed as:
in the formula (I), the compound is shown in the specification,andrespectively representing the direct signal power and the reflected signal power phiγIs the phase delay between the direct signal and the reflected signal. In the embodiment, 38000 segments of signal-to-noise ratio data are collected in 2013-2014 of a P351 observation station, wherein 19000 segments of signal-to-noise ratio data exist when snow exists on the ground.
S3: processing the acquired signal-to-noise ratio data according to the GNSS orbit parameters;
fig. 2 is a geometric model diagram of GNSS-R technology, based on the path delay delta between the direct and reflected signals,the phase delay phi between the direct signal and the reflected signal can be obtainedγExpression (c):
wherein λ is the wavelength of the navigation system signal; h is the vertical height from the phase center of the GNSS antenna to the reflecting surface; e is the satellite elevation angle, i.e. the angle between the direct satellite signal and the reflecting surface. From formula (2):
the signal-to-noise ratio can be considered to vary with the variation of the sine of the satellite elevation.
FIG. 3 is a graph of SNR for different surface features, both snow and bare, taking into account the mapping between SNR and reflection environment, and then selecting as input SNR data at low elevation range (5-25 °) where multipath effect is significant. As the experimental area selected in the embodiment is relatively spacious, the azimuth angle range is selected to be 0-360 degrees.
S4: performing high-dimensional feature space mapping by using the selected signal-to-noise ratio data as an input sample of a support vector classifier, and constructing an optimal hyperplane which can be used as a decision boundary for identification and classification;
specifically, the support vector machine is a machine learning method based on a statistical learning theory, and has the advantages of strong generalization capability, easy training, no local minimum value and the like. In this embodiment, the support vector classifier requires that the feature dimensions of the input samples are consistent, so that the signal-to-noise ratio obtained in step S3 is extended, in this embodiment, the feature dimension degree of the selected input sample (signal-to-noise ratio data) is 512, that is, the input sample is a feature matrix of 38000 × 512. Meanwhile, because the input samples are linear inseparable samples and need to be mapped into a high-dimensional space, the support vector classifier constructs a free hyperplane in a selected feature space to maximize the classification interval, and the optimization model at the moment is as follows:
wherein x represents a sample; phi (x) represents a new vector after x is mapped to a new feature space; ξ is the relaxation variable and C is the penalty parameter. By solving the above optimization problem, the optimal separation hyperplane w is solvedTPhi (x) + b is 0, the classification decision function, i.e. the classification model:
f(x)=sign(wTφ(x)+b) (5)
in the formula, sign is a step function. The classification of the samples can be obtained by introducing new samples. In equation (5), an inner-sum operation [ φ (x) on a high-dimensional feature space is requiredi),φ(xj)]It can be calculated by kernel function, and the present embodiment selects gaussian kernel function K (x)i,xj):
In this embodiment, the leave-out method is used for verifying 38000 inputted samples, that is, 80% of the samples (30400) are used as the training set, and 20% of the samples (7600) are used as the verification set. Through verification, the model classification accuracy can reach 96%.
Compared with the prior art, the GNSS-R snow detection method based on the support vector machine is used for researching a ground snow detection method in a global navigation satellite system (GNSS-R) reflection signal.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.
Claims (6)
1. A GNSS-R snow detection method based on a support vector machine is characterized by comprising the following steps:
selecting an experimental area;
respectively collecting satellite navigation signal-to-noise ratio data of the accumulated snow on the ground and the satellite navigation signal-to-noise ratio data of the snow on the ground when the ground is exposed;
processing the acquired signal-to-noise ratio data according to the GNSS orbit parameters;
and selecting signal-to-noise ratio data as an input sample of the support vector classifier to perform classification prediction according to the quality control requirement.
2. The GNSS-R snowfall detection method based on support vector machine according to claim 1, wherein in the step of collecting the signal-to-noise ratio data of satellite navigation when the ground is snowy and the ground is bare respectively, the signal-to-noise ratio data is expressed as:wherein the content of the first and second substances,andrespectively representing the direct signal power and the reflected signal power phiγIs the phase delay between the direct signal and the reflected signal.
3. The GNSS-R snowfall detection method based on support vector machine according to claim 2, characterized in that the direct-based GNSS-R snowfall detection method isPhase delay phi between signal and reflected signalγCan be derived from the path delay delta between the direct signal and the reflected signal, said phase delay phiγThe expression of (a) is:wherein λ is the wavelength of the navigation system signal, h is the vertical height from the phase center of the GNSS antenna to the reflecting surface, and e is the satellite elevation angle, i.e., the included angle between the direct satellite signal and the reflecting surface.
4. The GNSS-R snowfall detection method based on support vector machine according to claim 3, wherein in the step of processing the collected SNR data according to GNSS orbit parameters, the SNR can be expressed as:i.e., the signal-to-noise ratio data varies with the variation of the satellite elevation sinusoid.
5. The GNSS-R snowfall detection method based on SVM of claim 1, wherein the step of selecting SNR data as input samples of SVM for classification prediction according to quality control requirement specifically comprises: the support vector classifier constructs a free hyperplane in a selected feature space to maximize a classification interval, and an optimization model is expressed as:
6. The GNSS-R snowfall detection method based on support vector machine according to claim 5, characterized in that, by solving the optimal separation hyperplane wTPhi (x) + b is 0, and a classification decision function can be obtainedNamely, classification model: (x) sign (w)Tφ (x) + b), wherein sign is a step function.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111591910.8A CN114265127A (en) | 2021-12-23 | 2021-12-23 | GNSS-R snow detection method based on support vector machine |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111591910.8A CN114265127A (en) | 2021-12-23 | 2021-12-23 | GNSS-R snow detection method based on support vector machine |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114265127A true CN114265127A (en) | 2022-04-01 |
Family
ID=80829330
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111591910.8A Pending CN114265127A (en) | 2021-12-23 | 2021-12-23 | GNSS-R snow detection method based on support vector machine |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114265127A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106767383A (en) * | 2016-11-11 | 2017-05-31 | 清华大学 | The measuring method of the snow depth based on continuous operation GNSS stations signal-to-noise ratio data |
CN106951909A (en) * | 2016-11-16 | 2017-07-14 | 中国科学院遥感与数字地球研究所 | A kind of snow detection method of the satellite remote-sensing images of GF 4 |
CN110927829A (en) * | 2019-11-28 | 2020-03-27 | 桂林理工大学 | Support vector machine assisted GPS signal-to-noise ratio snow depth inversion method |
CN111221011A (en) * | 2018-11-26 | 2020-06-02 | 千寻位置网络有限公司 | GNSS positioning method and device based on machine learning |
CN113075706A (en) * | 2021-03-25 | 2021-07-06 | 上海海洋大学 | GNSS-R based snow depth inversion method and application thereof |
-
2021
- 2021-12-23 CN CN202111591910.8A patent/CN114265127A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106767383A (en) * | 2016-11-11 | 2017-05-31 | 清华大学 | The measuring method of the snow depth based on continuous operation GNSS stations signal-to-noise ratio data |
CN106951909A (en) * | 2016-11-16 | 2017-07-14 | 中国科学院遥感与数字地球研究所 | A kind of snow detection method of the satellite remote-sensing images of GF 4 |
CN111221011A (en) * | 2018-11-26 | 2020-06-02 | 千寻位置网络有限公司 | GNSS positioning method and device based on machine learning |
CN110927829A (en) * | 2019-11-28 | 2020-03-27 | 桂林理工大学 | Support vector machine assisted GPS signal-to-noise ratio snow depth inversion method |
CN113075706A (en) * | 2021-03-25 | 2021-07-06 | 上海海洋大学 | GNSS-R based snow depth inversion method and application thereof |
Non-Patent Citations (1)
Title |
---|
张京江;张双成;郭立峰;: "利用地基GPS反演积雪深度", 气象科技, no. 01 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Gebregiorgis et al. | To what extent is the day 1 GPM IMERG satellite precipitation estimate improved as compared to TRMM TMPA‐RT? | |
Fernandez-Moran et al. | A new calibration of the effective scattering albedo and soil roughness parameters in the SMOS SM retrieval algorithm | |
Koster et al. | Precipitation estimation using L‐band and C‐band soil moisture retrievals | |
Adhikari et al. | Comparative assessment of snowfall retrieval from microwave humidity sounders using machine learning methods | |
Podest et al. | Multisensor microwave sensitivity to freeze/thaw dynamics across a complex boreal landscape | |
Fonseca et al. | High-resolution temperature datasets in Portugal from a geostatistical approach: Variability and extremes | |
Petersen et al. | The GPM ground validation program | |
Tang et al. | Em-earth: the ensemble meteorological dataset for planet earth | |
Kotsuki et al. | Global precipitation forecasts by merging extrapolation-based nowcast and numerical weather prediction with locally optimized weights | |
Mroz et al. | Cross validation of active and passive microwave snowfall products over the continental United States | |
Guilloteau et al. | Beyond the pixel: Using patterns and multiscale spatial information to improve the retrieval of precipitation from spaceborne passive microwave imagers | |
Candy et al. | A Comparison of the impact of QuikScat and WindSat wind vector products on met office analyses and forecasts | |
Bahrami-Pichaghchi et al. | An estimation and multi-step ahead prediction study of monthly snow cover area, based on efficient atmospheric-oceanic dynamics | |
Stolz et al. | Refining the relationship between lightning and convective rainfall over the ocean | |
Durán Moro et al. | Assimilation of satellite swaths versus daily means of sea ice concentration in a regional coupled ocean-sea ice model | |
Kidd et al. | Quantitative precipitation estimation from satellite observations | |
Djibo et al. | High-resolution rainfall maps from commercial microwave links for a data-scarce region in West Africa | |
Hu et al. | GNSS-R snow depth retrieval algorithm based on PSO-LSTM | |
CN114265127A (en) | GNSS-R snow detection method based on support vector machine | |
Mehta et al. | Modeling of Raindrop Size Distribution Observed Using Micro Rain Radar Over Darjeeling (27.05 o N, 88.26 o E): An Eastern Himalayan Region | |
Turso et al. | MicroRadarNet: A network of weather micro radars for the identification of local high resolution precipitation patterns | |
Ortolani et al. | An EnKF-based method to produce rainfall maps from simulated satellite-to-ground MW-link signal attenuation | |
Hallali et al. | Comparison between radar and automatic weather station refractivity variability | |
Turk et al. | Estimating nonraining surface parameters to assist GPM constellation radiometer precipitation algorithms | |
Wijayarathne et al. | Application of weather radar for operational hydrology in Canada–A review |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |