CN113836779A - CNN-based farmland surface soil moisture inversion method for Sentinel multi-source data - Google Patents
CNN-based farmland surface soil moisture inversion method for Sentinel multi-source data Download PDFInfo
- Publication number
- CN113836779A CN113836779A CN202010579494.9A CN202010579494A CN113836779A CN 113836779 A CN113836779 A CN 113836779A CN 202010579494 A CN202010579494 A CN 202010579494A CN 113836779 A CN113836779 A CN 113836779A
- Authority
- CN
- China
- Prior art keywords
- data
- sentinel
- model
- test set
- soil moisture
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Radar Systems Or Details Thereof (AREA)
- Image Processing (AREA)
Abstract
Aiming at the problem that the Sentinel series satellite is influenced by parameters such as vegetation types, density and the like through satellite data moisture inversion, the invention discloses a farmland surface soil moisture inversion method based on CNN (China railway network) Sentinel multi-source data, and high-precision inversion of farmland surface soil moisture can be realized. The method comprises the following steps: step 1) carrying out corresponding pretreatment before using the Sentinel satellite data; step 2) constructing a data set, wherein input characteristic parameters comprise a dual-polarization radar backscattering coefficient (a)
、
) Altitude (1)
) Local Incidence Angle (LIA), polarization decomposition characteristics (H, A, α), and 3 vegetation index (NDVI, MSAVI, DVI); step 3) dividing the prepared 154 samples into a training set and a test set, and dividing the samples according to the proportion of 3:1, wherein the former is used as the training set of the model, and the latter is used as the test set of the model; step 4) inputting the data set into a CNN model for training, verifying the data set by using a test set, and finally passing the Root Mean Square Error (RMSE) and the correlation coefficient of the test set
Description
Technical Field
The invention relates to a farmland surface soil moisture inversion method based on CNN (CNN) -based Sentinel multi-source data, which is a complete soil moisture inversion process for a Sentinel series satellite of the European Bureau and can realize high-precision inversion of soil moisture in a vegetation coverage area.
Background
Soil moisture is a key variable in global water circulation, influences moisture and energy exchange in the fields of agriculture, ecology, climate, hydrology and the like, and has a positive effect on researching global climate change. Soil is composed primarily of a large number of heterogeneous particles, the soil moisture being distributed through the pores between these particles. Although the total amount of water contained in soil accounts for a very small proportion of the global fresh water resources, the soil water is an intermediate medium for interconversion of surface water, underground water and atmospheric water, so the soil water is an important surface parameter in an ecosystem. In ecological field, soil moisture is influencing the growth situation of regional interior green vegetation, and rainfall or irrigation water all need convert into earlier soil water and just can be utilized by the plant, and the plant absorbs soil moisture through the root system and obtains the required moisture of growth, and soil moisture supply is not enough can lead to the vegetation not enough to influence the growth of plant, and then lead to the in-zone ecological environment progressively worsening. In agricultural application, the soil moisture condition can represent whether the soil has drought and waterlogging and the severity of the soil, and can effectively guide the prevention and control of the drought and waterlogging and the management of agricultural water. Therefore, the method can timely and accurately monitor the spatial distribution condition of the soil moisture, has important scientific value and guiding significance for aspects such as crop yield estimation, water resource management, variable irrigation and the like, and is also beneficial to improving the level of agricultural decision scientification and further improving the quality of agricultural production service.
The method for monitoring the soil moisture by utilizing the SAR data is an effective mode and becomes important content of agricultural remote sensing. In recent years, with the rapid development of active microwave remote sensing technology and equipment thereof, functional parameters and performance indexes of radar sensors are continuously improved, the theoretical research of SAR for inverting soil moisture is continuously deepened, the inversion accuracy of soil moisture is greatly improved, and the monitoring of soil moisture by utilizing SAR also obtains wide attention. In the soil moisture inversion process, the inversion result is affected by the problems of model limitation, sensor configuration difference, inversion algorithm accuracy and the like, and the improvement of the soil moisture inversion algorithm and the model to improve the inversion accuracy becomes one of the hot points of agricultural remote sensing. Inversion of soil moisture by SAR data has been well studied at home and abroad and has achieved a lot of achievements, but some problems still remain. The current soil moisture common inversion model is a scattering model which is obtained through actual measurement data and is suitable for a certain area, although the common inversion model can be directly used for inverting the soil moisture, the common inversion model depends on a specific research area and the ground surface condition, such as an Oh model, an AIEM model and the like. When the model is applied, the soil volume water content, the root mean square height of the earth surface and the radar incidence angle need to meet the applicable conditions of the model, so that the application range is narrow. In addition, vegetation cover usually exists on the surface layer of soil, and for example, the surface of farmland often has different crops such as wheat, rape, corn and the like. Vegetation generates complex scattering to reduce the sensitivity of radar signals to soil moisture, so that the inversion needs to separate the scattering contribution of vegetation from the total backward scattering to eliminate the effect of vegetation coverage. Meanwhile, in the existing research, the intensity information of active microwave data is mostly utilized and the phase information is ignored, which affects the inversion accuracy of soil moisture to a certain extent.
Aiming at the defects of soil moisture inversion by microwave remote sensing at present, the invention provides a method for inverting soil moisture of farmland surface based on Sentinel multi-source remote sensing data, uses Sentinel-1 radar data and Sentinel-2 optical data of the European space Bureau as data sources, uses peripheral areas of Yangxi Yangling demonstration areas as research areas, analyzes the applicability of a CNN method in soil moisture inversion of farmland surface, discusses the accuracy of separating covered plants by using different vegetation indexes, then fuses the intensity information and phase information of microwave data in order to extract characteristic information which plays a key role in soil moisture quantitative inversion in satellite data, and constructs a soil moisture quantitative inversion model based on a convolutional neural network by using the characteristic that the CNN can dig deeper characteristic parameters of the satellite data. The method can be used for improving the soil moisture inversion accuracy and has certain practical significance for relevant applications such as drought control, variable irrigation and the like.
Disclosure of Invention
The invention discloses a farmland surface soil moisture inversion method based on Sentel multi-source data of CNN (China railway network), aiming at the problem that the Sentel series satellite of European and vacant Bureau is influenced by parameters such as soil roughness, vegetation type and density through satellite data moisture inversion, and the current commonly used soil moisture model has limitation, wherein the Sentel multi-source remote sensing data comprises Sentel-1 radar data and Sentel-2 optical data, and the high-precision inversion of farmland surface soil moisture can be realized.
In order to achieve the purpose, the main steps of the invention are as follows:
(1) inputting Sentinel-1 and Sentinel-2 image data, preprocessing the data, and representing vegetation information according to 3 optical vegetation indexes (NDVI, MSAVI and DVI).
(2) Constructing a data set, and inputting characteristic parameters including a dual-polarization radar backscattering coefficient (for a non-polarization decomposition characteristic CNN model)
、
) Altitude (1)
) And Local Incidence Angle (LIA) and 3 vegetation index (NDVI, MSAVI, DVI), H, A, alpha polarization decomposition characteristics.
(3) Dividing the prepared samples into a training set and a testing set according to the proportion of 3:1, wherein the training set is used as the training set of the model, and the testing set is used as the testing set of the model.
(4) The data set is input into a CNN model for training and verified by a test setFinally, the root mean square error RMSE and the correlation coefficient of the test set are passed
The predicted effect is evaluated.
Further, the specific method of the step (1) is as follows:
(a) the Sentinel-1 radar data used in the invention comprises two data of SLC (single Look complex) and GRDH (ground Range detected) in an interference width mode, wherein the data in the format of GRDH comprises intensity information of an observation target, and the data in the format of SLC comprises intensity information and phase information of the observation target. The data preprocessing step of the GRDH format comprises radiation correction, terrain radiation correction, filtering and terrain correction, and finally dual-polarization radar backscattering coefficient is output (
、
) Altitude (1)
) Local Angle of Incidence (LIA). The preprocessing steps of the SLC format data comprise radiation correction, splicing and clipping, H/A/alpha polarization decomposition, filtering and denoising, terrain correction, and finally output polarization decomposition characteristics (H, A, alpha).
(b) Carrying out atmospheric correction by installing a special plug-in Sentiel-2 data processing Sen2 Cor;
(c) and after the Sentinel-2 data is preprocessed, calculating vegetation indexes including vegetation indexes NDVI, MSAVI and DVI) by using the red wave band and the near infrared wave band of the Sentinel-2 data so as to represent vegetation coverage influence. And then, taking ER Mapper software as a platform, resampling the Sentinel-1 radar image and the Sentinel-2 optical image, and then carrying out image registration and resampling to provide a data source for soil moisture inversion.
Further, the specific method of the step (2) is as follows:
will be provided with
、
、
LIA, vegetation index, (H, A, alpha) as input characteristic parameters of the CNN model, and the input characteristic parameters comprise a dual-polarization radar backscattering coefficient (a)
、
) Altitude (1)
) Local Angle of Incidence (LIA), 3 vegetation index (NDVI, MSAVI, DVI) and H, A, alpha polarization decomposition characteristics.
Further, the specific method of the step (3) is as follows:
the prepared 154 samples are divided into a training set and a test set according to the proportion of 3:1, wherein the former is used as the training set of the model, and the latter is used as the test set of the model.
Further, the specific method of the step (4) is as follows:
inputting the data set into a CNN model for training, verifying the trained network by using a test set, respectively obtaining soil moisture inversion results of the CNN model with polarization decomposition characteristics under different training sets, and obtaining a Root Mean Square Error (RMSE) and a correlation coefficient of the test set
The predicted effect is evaluated.
Description of the figures (tables)
The disclosure may be better understood by reference to the following description taken in conjunction with the accompanying drawings, which are incorporated in and form a part of this specification, and the following detailed description, in which:
FIG. 1 is a schematic flow diagram of a soil moisture inversion method according to an embodiment not disclosed herein;
FIG. 2 is a CNN network architecture diagram of a soil moisture inversion method according to an embodiment not disclosed herein;
FIG. 3 is a CNN model training process loss function value of a soil moisture inversion method according to an embodiment not disclosed herein;
FIG. 4 is a soil moisture inversion result of a CNN model of a soil moisture inversion method according to an embodiment not disclosed herein;
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to the attached figure 1, the specific implementation steps of the invention are as follows:
(1) The Sentinel-1 data products are divided into three types of Level-0, Level-1 and Level-2, wherein Level-0 Level data refers to unprocessed original data, Level-1 Level data refers to data after preliminary preprocessing, and Level-2 Level data refers to products which can be directly used after processing is completed. The Level-1 data can be divided into two types, namely SLC and GRDH, due to different processing modes, the data in the GRDH format comprises the intensity information of the observation target, and the data in the SLC format comprises the intensity information and the phase information of the observation target. The Sentinel-1 radar image acquired by the method is Level-1 data, comprises data in an SLC format and data in a GRDH format in an IW mode, and comprises two polarization modes of VH and VV. The satellite data needs to be correspondingly preprocessed before being used. The preprocessing process of the Sentinel-1 radar data is mainly performed on corresponding Software (SNAP) specially developed by the European Bureau. For GRDH format data in IW imaging mode, the preprocessing steps include radiation correction, terrain radiation correction, filtering, and terrain correction. For SLC format data, the preprocessing steps include radiation correction, stitching and clipping, polarization decomposition, filtering and denoising, and terrain correction. Wherein, the three steps of radiation correction, filtering denoising and terrain correction are the same as the process of processing GRDH format data. The splicing and cutting process comprises the following steps: in the interference width imaging mode, the sub-strips are spliced into a complete image, and the image in the SLC format has a larger data size than that in the GRDH format, so that a research area can be cut out for subsequent processing.
(2) And (3) carrying out atmospheric correction on Level-1C by installing a special plug-in Sentel-2 data processing Sentnel-2 Senn 2Cor to obtain Level-2A Level data. After the Sentinel-2 data are preprocessed, the ER Mapper software is used as a platform, the Sentinel-1 radar image and the Sentinel-2 optical image are subjected to image registration and resampling after being subjected to resampling, and a data source is provided for soil moisture inversion.
(3) NDVI is one of the most common vegetation indexes, and is commonly used to monitor vegetation growth status and eliminate radiation errors. MSAVI considers both the effects of vegetation and soil background, which increase with increasing vegetation coverage. DVI is sensitive to changes in soil background. The calculation formulas of NDVI, MSAVI and DVI are respectively shown in the specification.
In the formula (I), the compound is shown in the specification,
the red band, the fourth band of Sentinel-2 in this study,
is a near infrared wave, namely the eighth wave band of Sentinel-2.
The structure of a typical CNN network consists of a convolution layer, a pooling layer, an excitation function layer, a full connection layer and the like, and the characteristics and the functions of each layer are different. The convolutional layer is used for learning characteristics, abstracting and extracting characteristics of input data; the pooling layer is used for zooming the learned features to reduce the operation process and mainly comprises a maximum pooling layer and an average pooling layer; the excitation function layer is used for carrying out linear transformation and extracting nonlinear features, and the nonlinear features comprise nonlinear functions such as tanh, Sigmoid, Relu and the like; the full connection layer is the characteristic after the connection convolution pooling processing, and the output value is sent to a classifier for classification. In a CNN structure taking an image as an input, a pooling layer is used for eliminating domain features of data and changing the dimension of input feature data (Naja 2019, etc.), but the difference between data in the text is obvious and the feature dimension is small, so that no pooling layer is arranged, a network only with a convolutional layer and a fully-connected layer is directly created, and a classifier is changed into a regressor. Considering that the Rule function is an unsaturated nonlinear function, and the problems of gradient disappearance and the like do not exist (Nair and Hinton 2010), the excitation function selects the Rule function. For the non-polarized decomposition characteristic CNN model, the input characteristic parameters comprise the backward dispersion of the dual-polarized radarCoefficient of reflection (
、
) Altitude (1)
) And Local Incidence Angle (LIA) and 2 vegetation indexes (NDVI, MSAVI), two polarization decomposition features of H and alpha are added in combination with input data of a polarization decomposition feature CNN model, for comparison, the input data is uniformly expanded into a matrix of 15 × 15 for modeling because of different sizes of data of the polarization decomposition features, after repeated tests are carried out by using the existing data, the sizes of convolution kernels are determined to be 5 × 5 and 3 × 3 respectively, the number of the convolution kernels is 25 and 30 respectively, and the final network structure comprises an input layer, two convolution layers, a full connection layer and a regression layer, wherein the specific structure is shown in fig. 2.
And 3, dividing the prepared sample into a training set and a test set according to the ratio of 3:1, wherein the training set is used as the training set of the model, and the test set is used as the test set of the model.
The network learning rate and the maximum iteration number are respectively set to be 0.001 and 1000 through early training, the loss function value in the network training process when the training set ratio test set is equal to 3:1 in the CNN model training process is shown in figure 3, the abscissa is the iteration number, and the ordinate is the loss function value, as can be seen from figure 3, in the initial stage, the loss function value is rapidly reduced, the network tends to be converged along with the increase of the iteration number, and the network is completely converged when the iteration is finished, which indicates that the network learning state is good and no overfitting state occurs.
Step 4, inputting the data set into a CNN model for training, verifying the data set by using a test set, and finally passing the Root Mean Square Error (RMSE) and the correlation coefficient of the test set
Coming commentAnd (4) predicting the effect.
The trained network is verified by using the test set, fig. 4 shows that the soil moisture inversion results of the CNN model are soil moisture inversion results of the CNN model with the electrodeless decomposition characteristics under different training sets, the abscissa is the measured value of the soil moisture, and the ordinate is the predicted value of the model, and as can be seen from fig. 4, when the training sets: test set =3:1, test set
And RMSE 0.8847 and 0.0221cm3/cm3, respectively. The CNN model has better inversion accuracy and can be used for soil moisture inversion in the region.
Claims (5)
1. The farmland surface soil moisture inversion method of the Sentinel multi-source data based on the CNN comprises the following steps:
(1) inputting Sentinel-1 and Sentinel-2 image data, preprocessing the data, and acquiring input characteristic parameter dual-polarized radar backscattering coefficient (by using the Sentinel-1 data: (
、
) Altitude (1)
) Local Incidence Angle (LIA), polarization decomposition characteristics (H, A, alpha), calculating vegetation indexes including vegetation indexes NDVI, MSAII and DVI by using the red band and the near infrared band of the Sentinel-2 data;
(2) constructing a data set, wherein the model input characteristic parameters comprise the backscattering coefficient (b) of the dual-polarized radar
、
) Altitude (1)
) Local Incidence Angle (LIA), polarization decomposition characteristics (H, A, α), and 3 vegetation index (NDVI, MSAVI, DVI);
(3) dividing the prepared sample into a training set and a test set according to the proportion of 3:1, wherein the training set is used as the training set of the model, and the test set is used as the test set of the model;
(4) inputting the data set into a CNN model for training, verifying the data set by using a test set, and finally passing the Root Mean Square Error (RMSE) and the correlation coefficient of the test set
The predicted effect is evaluated.
2. The method as claimed in claim 1, wherein the specific method of step 1) is as follows:
(a) the Sentinel-1 radar data used in the invention comprises two data of SLC (Single Look Complex) and GRDH (ground Range detected) in an interference width mode, wherein the data in the GRDH format comprises intensity information of an observation target, the data in the SLC format comprises the intensity information and phase information of the observation target, the data in the GRDH format is preprocessed in the steps of radiation correction, splicing cutting, polarization decomposition, filtering denoising and terrain correction, and finally a backscattering coefficient of a dual-polarization radar is output
、
) Altitude (1)
) Local Angle of Incidence (L, L)IA), preprocessing the SLC format data comprises radiation correction, splicing and cutting, H/A/alpha polarization decomposition, filtering and denoising, terrain correction and finally outputting polarization decomposition characteristics (H, A, alpha);
(b) carrying out atmospheric correction by installing a special plug-in Sentiel-2 data processing Sen2 Cor;
(c) after the Sentinel-2 data are preprocessed, the red waveband and the near infrared waveband of the Sentinel-2 data are used for calculating vegetation values including vegetation indexes NDVI, MSAVI and DVI to represent vegetation coverage influence, then ER Mapper software is used as a platform, image registration and resampling are carried out after Sentinel-1 radar images and Sentinel-2 optical images are resampled, and a data source is provided for soil moisture inversion.
3. The method as claimed in claim 1, wherein the specific method of step 2) is:
(a) will be provided with
、
、
LIA, vegetation index, (H, A, alpha) as input characteristic parameters of the CNN model, and the input characteristic parameters comprise a dual-polarization radar backscattering coefficient (a)
、
) Altitude (1)
) Local incident Angle (Local index Angle)LIA), 3 vegetation index (NDVI, MSAVI, DVI) and H, A, alpha polarization decomposition characteristics.
4. The method as claimed in claim 1, wherein the specific method of step 3) is:
(a) the prepared 154 samples are divided into a training set and a test set according to the proportion of 3:1, wherein the former is used as the training set of the model, and the latter is used as the test set of the model.
5. The method as claimed in claim 1, wherein the specific method of step 4) is as follows:
(a) inputting the data set into a CNN model for training, verifying the trained network by using a test set, respectively obtaining soil moisture inversion results of the CNN model with polarization decomposition characteristics under different training sets, and obtaining a Root Mean Square Error (RMSE) and a correlation coefficient of the test set
The predicted effect is evaluated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010579494.9A CN113836779A (en) | 2020-06-23 | 2020-06-23 | CNN-based farmland surface soil moisture inversion method for Sentinel multi-source data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010579494.9A CN113836779A (en) | 2020-06-23 | 2020-06-23 | CNN-based farmland surface soil moisture inversion method for Sentinel multi-source data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113836779A true CN113836779A (en) | 2021-12-24 |
Family
ID=78964000
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010579494.9A Pending CN113836779A (en) | 2020-06-23 | 2020-06-23 | CNN-based farmland surface soil moisture inversion method for Sentinel multi-source data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113836779A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114397276A (en) * | 2022-01-19 | 2022-04-26 | 北华航天工业学院 | Regional soil humidity monitoring method based on equivalent rainfall estimation method |
| CN114441469A (en) * | 2022-02-24 | 2022-05-06 | 龙岩烟草工业有限责任公司 | Calibration method and device of moisture meter and computer equipment |
| CN114660563A (en) * | 2022-03-24 | 2022-06-24 | 自然资源部国土卫星遥感应用中心 | Crop coverage area full-polarization multi-dimensional SAR soil moisture inversion method |
| CN114880883A (en) * | 2022-07-07 | 2022-08-09 | 中国科学院、水利部成都山地灾害与环境研究所 | Mountain land surface soil moisture remote sensing estimation method and device and electronic equipment |
| CN115035413A (en) * | 2022-06-30 | 2022-09-09 | 河南理工大学 | Multi-temporal active and passive remote sensing random forest crop identification method and system |
| CN115410086A (en) * | 2022-08-26 | 2022-11-29 | 南方海洋科学与工程广东省实验室(广州) | Water quality inversion method, device and equipment based on remote sensing image |
| CN116879297A (en) * | 2023-09-07 | 2023-10-13 | 航天宏图信息技术股份有限公司 | Soil moisture collaborative inversion method, device, equipment and medium |
| CN117313563A (en) * | 2023-11-30 | 2023-12-29 | 江汉大学 | Configuration method of soil moisture reconstruction model under physical constraint and deep learning coupling |
-
2020
- 2020-06-23 CN CN202010579494.9A patent/CN113836779A/en active Pending
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114397276A (en) * | 2022-01-19 | 2022-04-26 | 北华航天工业学院 | Regional soil humidity monitoring method based on equivalent rainfall estimation method |
| CN114397276B (en) * | 2022-01-19 | 2023-12-01 | 北华航天工业学院 | Regional soil humidity monitoring method based on equivalent precipitation estimation method |
| CN114441469A (en) * | 2022-02-24 | 2022-05-06 | 龙岩烟草工业有限责任公司 | Calibration method and device of moisture meter and computer equipment |
| CN114441469B (en) * | 2022-02-24 | 2024-06-04 | 龙岩烟草工业有限责任公司 | Calibration method and device of moisture meter and computer equipment |
| CN114660563A (en) * | 2022-03-24 | 2022-06-24 | 自然资源部国土卫星遥感应用中心 | Crop coverage area full-polarization multi-dimensional SAR soil moisture inversion method |
| CN115035413B (en) * | 2022-06-30 | 2023-08-29 | 河南理工大学 | Multi-time-phase active and passive remote sensing random forest crop identification method and system |
| CN115035413A (en) * | 2022-06-30 | 2022-09-09 | 河南理工大学 | Multi-temporal active and passive remote sensing random forest crop identification method and system |
| CN114880883A (en) * | 2022-07-07 | 2022-08-09 | 中国科学院、水利部成都山地灾害与环境研究所 | Mountain land surface soil moisture remote sensing estimation method and device and electronic equipment |
| CN114880883B (en) * | 2022-07-07 | 2022-10-11 | 中国科学院、水利部成都山地灾害与环境研究所 | Mountain land surface soil moisture remote sensing estimation method and device and electronic equipment |
| CN115410086A (en) * | 2022-08-26 | 2022-11-29 | 南方海洋科学与工程广东省实验室(广州) | Water quality inversion method, device and equipment based on remote sensing image |
| CN116879297A (en) * | 2023-09-07 | 2023-10-13 | 航天宏图信息技术股份有限公司 | Soil moisture collaborative inversion method, device, equipment and medium |
| CN116879297B (en) * | 2023-09-07 | 2023-12-12 | 航天宏图信息技术股份有限公司 | Soil moisture collaborative inversion method, device, equipment and medium |
| CN117313563A (en) * | 2023-11-30 | 2023-12-29 | 江汉大学 | Configuration method of soil moisture reconstruction model under physical constraint and deep learning coupling |
| CN117313563B (en) * | 2023-11-30 | 2024-02-27 | 江汉大学 | Configuration method of soil moisture reconstruction model under physical constraint and deep learning coupling |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113836779A (en) | CNN-based farmland surface soil moisture inversion method for Sentinel multi-source data | |
| Jin et al. | Deep neural network algorithm for estimating maize biomass based on simulated Sentinel 2A vegetation indices and leaf area index | |
| CN114076738B (en) | Method for estimating soil salinity of residual straw farmland by using remote sensing construction index | |
| Sun et al. | Retrieval of rapeseed leaf area index using the PROSAIL model with canopy coverage derived from UAV images as a correction parameter | |
| CN105004320B (en) | A kind of high score satellite data land table vegetation coverage inversion method and system | |
| CN112711989B (en) | Corn straw coverage estimation method based on radar remote sensing and optical remote sensing | |
| Zhang et al. | Mapping flood by the object-based method using backscattering coefficient and interference coherence of Sentinel-1 time series | |
| CN103345707A (en) | Crop maturation stage remote sensing prediction method based on multi-source remote sensing data | |
| CN110687053B (en) | Regional organic matter content estimation method and device based on hyperspectral image | |
| CN111832518A (en) | Space-time fusion-based TSA remote sensing image land utilization method | |
| Qiao et al. | Estimating maize LAI by exploring deep features of vegetation index map from UAV multispectral images | |
| Fathololoumi et al. | Surface biophysical features fusion in remote sensing for improving land crop/cover classification accuracy | |
| Shen et al. | Poyang Lake wetland vegetation biomass inversion using polarimetric RADARSAT-2 synthetic aperture radar data | |
| Afrasiabian et al. | Effects of spatial, temporal, and spectral resolutions on the estimation of wheat and barley leaf area index using multi-and hyper-spectral data (case study: Karaj, Iran) | |
| CN111783288A (en) | Inversion method for soil salinity of yellow river delta based on Landsat8 | |
| Lou et al. | An effective method for canopy chlorophyll content estimation of marsh vegetation based on multiscale remote sensing data | |
| Yuan et al. | Research on rice leaf area index estimation based on fusion of texture and spectral information | |
| Qi et al. | Soil salinity inversion in coastal cotton growing areas: An integration method using satellite‐ground spectral fusion and satellite‐UAV collaboration | |
| Liu et al. | An algorithm for early rice area mapping from satellite remote sensing data in southwestern Guangdong in China based on feature optimization and random Forest | |
| Zhu et al. | Research on deep learning method recognition and a classification model of grassland grass species based on unmanned aerial vehicle hyperspectral remote sensing | |
| Yin et al. | Synergistic estimation of soil salinity based on Sentinel-1 image texture and Sentinel-2 salinity spectral indices | |
| CN113534083A (en) | Corn stubble mode identification method, device and medium based on SAR | |
| Yang et al. | Estimation of rice aboveground biomass by UAV imagery with photosynthetic accumulation models | |
| CN112418073A (en) | Wheat plant nitrogen content estimation method based on unmanned aerial vehicle image fusion characteristics | |
| CN117075138A (en) | Remote sensing measurement and calculation method, system and medium for canopy height of 30-meter forest in area |
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 |