CN106779067B - Soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data - Google Patents
Soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data Download PDFInfo
- Publication number
- CN106779067B CN106779067B CN201611101527.9A CN201611101527A CN106779067B CN 106779067 B CN106779067 B CN 106779067B CN 201611101527 A CN201611101527 A CN 201611101527A CN 106779067 B CN106779067 B CN 106779067B
- Authority
- CN
- China
- Prior art keywords
- data
- remote sensing
- predeterminable area
- soil moisture
- training
- 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.)
- Active
Links
- 239000002689 soil Substances 0.000 title claims abstract description 210
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 141
- 230000003287 optical effect Effects 0.000 claims abstract description 99
- 238000013528 artificial neural network Methods 0.000 claims abstract description 93
- 238000005259 measurement Methods 0.000 claims abstract description 62
- 238000012952 Resampling Methods 0.000 claims description 15
- 230000001537 neural effect Effects 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 7
- 238000013481 data capture Methods 0.000 claims description 4
- 210000004218 nerve net Anatomy 0.000 claims description 4
- 210000005036 nerve Anatomy 0.000 claims description 2
- 241001269238 Data Species 0.000 claims 2
- 210000004027 cell Anatomy 0.000 description 13
- 238000010586 diagram Methods 0.000 description 12
- 230000005855 radiation Effects 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 238000012216 screening Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000002310 reflectometry Methods 0.000 description 2
- VAYOSLLFUXYJDT-RDTXWAMCSA-N Lysergic acid diethylamide Chemical compound C1=CC(C=2[C@H](N(C)C[C@@H](C=2)C(=O)N(CC)CC)C2)=C3C2=CNC3=C1 VAYOSLLFUXYJDT-RDTXWAMCSA-N 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000002352 surface water Substances 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The present invention relates to a kind of soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data, which comprises obtain space and time continuous optical remote sensing data and microwave remote sensing soil moisture measurements training data；Determine the optical remote sensing training data of predeterminable area；Optical remote sensing training data, three-dimensional geographic information and timeline information are determined as to input information, soil moisture is determined as output information, constructs backpropagation neural network to be trained；Using microwave remote sensing soil moisture measurements training data as training sample, the training backpropagation neural network to be trained obtains soil moisture backpropagation neural network；By space and time continuous optical remote sensing data, three-dimensional geographic information and timeline information, the soil moisture backpropagation neural network is inputted, obtains soil moisture data.The present invention obtains the microwave remote sensing soil data of space and time continuous by constructing and training neural network using the optical remote sensing data of the space and time continuous after rebuilding.
Description
Technical field
The present invention relates to surface water resources technical fields, rebuild more particularly to the soil moisture based on multi- source Remote Sensing Data data
Method and system.
Background technique
As water circulation in a key element, soil humidity information agricultural, the hydrology, meteorology, ecology etc. science and
Production field is all extremely important, but soil moisture spatial distribution is uneven, and dynamic change at any time, therefore when accurate acquisition
Empty continuous soil moisture has difficulties, but is of great significance.Draught monitor based on soil moisture is vegetation especially agriculture
Crops Drought situation is most directly expressed, and the soil moisture product of space and time continuous has emphatically crop growth monitoring and the yield by estimation
The indicative function wanted shows the significant economic value of soil moisture product.
The main means that satellite remote sensing is observed as a wide range of soil moisture are being rapidly developed in the latest 20 years, but
It is reset period, underlying surface type and model algorithm by satellite to be limited, all satellite remote sensing soil moisture products exist not
With the missing of degree.Remote Sensing of Soil Moisture product such as AMSR-E (the Advanced Microwave Scanning of mainstream at present
Radiometer-Earth Observing System) US National Aeronautics and Space Administration's version soil moisture and Japanese universe
Aeronautical research Development institution version soil moisture, SMOS (the Soil Moisture and Ocean Salinity) soil of European Space Agency
Earth humidity product, FY (Fengyun) soil moisture product of China national weather bureau, Dutch ASCAT (Advanced
Scatterometer) soil moisture product of the soil moisture product etc. about more than 10 based on satellite remote sensing has one seriously
Missing problem is extremely difficult to 50% or more time coverage, drastically influences Related product to earth's surface in addition to high latitude area
Continuous monitoring, restrict the excavation of soil moisture product timeliness application value.
Summary of the invention
Based on this, it is necessary to calculate the discontinuous problem of soil moisture data for satellite data, provide a kind of based on more
The soil moisture method for reconstructing and system of source remotely-sensed data, wherein the described method includes:
Obtain the space and time continuous optical remote sensing data of predeterminable area and the microwave remote sensing soil moisture measurements instruction of the predeterminable area
Practice data；
According to the microwave remote sensing soil moisture measurements training data of the predeterminable area, the optical remote sensing training of predeterminable area is determined
Data；
By the optical remote sensing training data of the predeterminable area, the three-dimensional geographic information of the predeterminable area and described default
The timeline information in region is determined as inputting information, and soil moisture is determined as output information, constructs back-propagating mind to be trained
Through network；
Using the microwave remote sensing soil moisture measurements training data of the predeterminable area as training sample, training is described backward to training
Propagation Neural Network obtains soil moisture backpropagation neural network；
By the space and time continuous optical remote sensing data of the predeterminable area, the three-dimensional geographic information of the predeterminable area and described
The timeline information of predeterminable area inputs the soil moisture backpropagation neural network, obtains the soil moisture of predeterminable area
Data.
The microwave remote sensing soil moisture measurements training data for obtaining the predeterminable area in one of the embodiments, packet
It includes: obtaining the microwave remote sensing soil moisture measurements initial data of predeterminable area；In the microwave remote sensing soil moisture measurements initial data, sieve
It selects not by snow cover and surface temperature is greater than the pel data of zero degrees celsius；The pel data filtered out is determined
For microwave remote sensing soil moisture measurements training data.
The space and time continuous optical remote sensing data for obtaining predeterminable area in one of the embodiments, comprising: obtain pre-
If the optical remote sensing initial data in region, the optical remote sensing initial data includes vegetation index initial data, surface temperature original
Beginning data and surface albedo initial data；The vegetation index initial data of the predeterminable area is passed through humorous based on time series
Wave analysis algorithm for reconstructing calculates the space and time continuous vegetation index of predeterminable area；By the surface temperature original number of the predeterminable area
According to by the algorithm for reconstructing based on reference sequences, the space and time continuous surface temperature of predeterminable area is calculated；By the predeterminable area
Surface albedo initial data calculates the space and time continuous earth's surface reflection of light of predeterminable area by the algorithm for reconstructing based on spatio-temporal filtering
Rate；By the space and time continuous vegetation index of the predeterminable area, the space and time continuous surface temperature of the predeterminable area and described default
The surface albedo in region is determined as the space and time continuous optical remote sensing data of predeterminable area.
In one of the embodiments, it is described by the space and time continuous optical remote sensing training data of the predeterminable area, it is described
The timeline information of the three-dimensional geographic information of predeterminable area and the predeterminable area is determined as inputting information, comprising: according to described
The grid cell size of the microwave remote sensing soil moisture measurements training data of predeterminable area, the vegetation index, surface temperature and earth's surface is anti-
Resampling is carried out according to rate, obtains input vegetation index, input surface temperature and input surface albedo；The input vegetation is referred to
Number, input surface temperature, input surface albedo, the three-dimensional geographic information of the predeterminable area and the time of the predeterminable area
Axis information is determined as inputting information.
Backpropagation neural network to be trained described in the training in one of the embodiments, comprising: in M*N institute
Within the scope of the grid cell size for stating microwave remote sensing soil moisture measurements training data, the training backpropagation neural network to be trained,
Middle M and N is positive integer.
Soil moisture method for reconstructing provided by the present invention based on multi- source Remote Sensing Data data, utilizes optical remote sensing product and three
Geography information and timeline information are tieed up, backpropagation neural network is constructed, is with preset microwave soil moisture training data
Sample set after being trained to the backpropagation neural network, obtains soil moisture backpropagation neural network, then by space-time
The continuous optical remote sensing data input soil moisture backpropagation neural network, obtains continuous soil moisture data.
By the input of neural network, the optical remote sensing data of space and time continuous and three-dimensional geographic information, timeline information, predict that microwave is distant
The missing values for feeling soil moisture data, obtain the microwave remote sensing soil data of space and time continuous.
In one of the embodiments, by the screening to microwave remote sensing soil moisture measurements training data, make the backward of building
Propagation Neural Network has accurate training sample, to improve the accurate of the backpropagation neural network output valve after training
Rate.
In one of the embodiments, by carrying out the unification of spatial resolution, Yi Jigen to optical remote sensing training data
Resampling is carried out according to microwave remote sensing training data, makes the backpropagation neural network of building that there is the input of equal resolution scale
Output data, the accuracy rate of the backpropagation neural network output valve after improving training.
The backpropagation neural network is instructed within the scope of certain grid cell size in one of the embodiments,
Practice, it can the continuity for improving output data, having, which can guarantee, to lead to the accuracy rate of output data because range is excessive
It reduces.
After the discontinuous optical remote sensing data of space-time are carried out space-time reconstruction in one of the embodiments, when getting
After empty continuous optical remotely-sensed data, trained backpropagation neural network is inputted, the soil that can obtain space and time continuous is wet
Degree evidence, the space-time algorithm for reconstructing of each optical remote sensing data, i.e., so that the soil moisture data that neural computing obtains
Space-time expending, and can guarantee the accuracy for the soil moisture data being calculated.
The present invention also provides a kind of soil moisture reconstructing system based on multi- source Remote Sensing Data data, comprising:
Optical remote sensing data and microwave training data obtain module, for obtaining the space and time continuous optical remote sensing of predeterminable area
The microwave remote sensing soil moisture measurements training data of data and the predeterminable area；
Optical remote sensing training data obtains module, for according to the microwave remote sensing soil moisture measurements of predeterminable area training number
According to determining the optical remote sensing training data of predeterminable area；
Neural network constructs module, for by the optical remote sensing training data of the predeterminable area, the predeterminable area
The timeline information of three-dimensional geographic information and the predeterminable area is determined as inputting information, and soil moisture is determined as output letter
Breath, constructs backpropagation neural network to be trained；
Neural metwork training module, for being training sample with the microwave remote sensing soil moisture measurements training data of the predeterminable area
This, the training backpropagation neural network to be trained obtains soil moisture backpropagation neural network；
Soil moisture data obtains module, for by space and time continuous optical remote sensing data of the predeterminable area, described pre-
If the timeline information of the three-dimensional geographic information in region and the predeterminable area, the soil moisture back-propagating nerve net is inputted
Network obtains the soil moisture data of predeterminable area.
The optical remote sensing data and microwave training data obtain module in one of the embodiments, comprising: microwave instruction
Practice data capture unit, for obtaining the microwave remote sensing soil moisture measurements initial data of predeterminable area；In the microwave remote sensing soil
In humidity initial data, filters out not by snow cover and surface temperature is greater than the pel data of zero degrees celsius；By the sieve
The pel data selected is determined as microwave remote sensing soil moisture measurements training data.
The optical remote sensing data and microwave training data obtain module in one of the embodiments, comprising: optics is distant
Feel data capture unit, for obtaining the optical remote sensing initial data of predeterminable area, the optical remote sensing initial data includes planting
By index initial data, surface temperature initial data and surface albedo initial data；By the vegetation index of the predeterminable area
Initial data calculates the space and time continuous vegetation index of predeterminable area by being based on time series frequency analysis algorithm for reconstructing；By institute
The surface temperature initial data of predeterminable area is stated by the algorithm for reconstructing based on reference sequences, calculates the space and time continuous of predeterminable area
Surface temperature；By the surface albedo initial data of the predeterminable area by the algorithm for reconstructing based on spatio-temporal filtering, calculate pre-
If the space and time continuous surface albedo in region；By the space and time continuous vegetation index of the predeterminable area, the predeterminable area when
The surface albedo of empty continuous surface temperature and the predeterminable area, is determined as the space and time continuous optical remote sensing number of predeterminable area
According to.
The neural network constructs module in one of the embodiments, comprising:
Resampling computing unit, for the pixel ruler according to the microwave remote sensing soil moisture measurements training data of the predeterminable area
The vegetation index, surface temperature and surface albedo are carried out resampling by degree, obtain input vegetation index, input earth's surface temperature
Degree and input surface albedo；
Input information determination unit, for by the input vegetation index, input surface temperature, input surface albedo,
The timeline information of the three-dimensional geographic information of the predeterminable area and the predeterminable area is determined as inputting information.
The neural metwork training module in one of the embodiments, is also used in the M*N microwave remote sensing soil
Within the scope of the grid cell size of humidity training data, the training backpropagation neural network to be trained, wherein M and N is positive integer.
Soil moisture reconstructing system provided by the present invention based on multi- source Remote Sensing Data data, utilizes optical remote sensing product and three
Geography information and timeline information are tieed up, backpropagation neural network is constructed, is with preset microwave soil moisture training data
Sample set after being trained to the backpropagation neural network, obtains soil moisture backpropagation neural network, then by space-time
The continuous optical remote sensing data input soil moisture backpropagation neural network, obtains continuous soil moisture data.
By the input of neural network, the optical remote sensing data of space and time continuous and three-dimensional geographic information, timeline information, predict that microwave is distant
The missing values for feeling soil moisture data, obtain the microwave remote sensing soil data of space and time continuous.
In one of the embodiments, by the screening to microwave remote sensing soil moisture measurements training data, make the backward of building
Propagation Neural Network has accurate training sample, to improve the accurate of the backpropagation neural network output valve after training
Rate.
In one of the embodiments, by carrying out the unification of spatial resolution, Yi Jigen to optical remote sensing training data
Resampling is carried out according to microwave remote sensing training data, makes the backpropagation neural network of building that there is the input of equal resolution scale
Output data, the accuracy rate of the backpropagation neural network output valve after improving training.
The backpropagation neural network is instructed within the scope of certain grid cell size in one of the embodiments,
Practice, it can the continuity for improving output data, having, which can guarantee, to lead to the accuracy rate of output data because range is excessive
It reduces.
After the discontinuous optical remote sensing data of space-time are carried out space-time reconstruction in one of the embodiments, when getting
After empty continuous optical remotely-sensed data, trained backpropagation neural network is inputted, the soil that can obtain space and time continuous is wet
Degree evidence, the space-time algorithm for reconstructing of each optical remote sensing data, i.e., so that the soil moisture data that neural computing obtains
Space-time expending, and can guarantee the accuracy for the soil moisture data being calculated.
Detailed description of the invention
Fig. 1 is the flow diagram of the soil moisture method for reconstructing based on multi- source Remote Sensing Data data in one embodiment；
Fig. 2 is the flow diagram of the soil moisture method for reconstructing based on multi- source Remote Sensing Data data in another embodiment；
Fig. 3 is the flow diagram of the soil moisture method for reconstructing based on multi- source Remote Sensing Data data in further embodiment；
Fig. 4 is the structural schematic diagram of the soil moisture reconstructing system based on multi- source Remote Sensing Data data in one embodiment；
Fig. 5 is the structural schematic diagram of the soil moisture reconstructing system based on multi- source Remote Sensing Data data in another embodiment；
Fig. 6 is the structural schematic diagram of the soil moisture reconstructing system based on multi- source Remote Sensing Data data in further embodiment.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, right with reference to the accompanying drawings and embodiments
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention.
Fig. 1 is the flow diagram of the soil moisture method for reconstructing based on multi- source Remote Sensing Data data in one embodiment, such as
Soil moisture method for reconstructing shown in FIG. 1 based on multi- source Remote Sensing Data data includes:
Step S100 obtains the space and time continuous optical remote sensing data of predeterminable area and the microwave remote sensing soil of the predeterminable area
Earth humidity training data.
Specifically, the predeterminable area is the region for needing to carry out soil moisture reconstruction.The optical remote sensing data of the satellite,
Refer to that satellite utilizes the optical instruments such as infrared ray, the remotely-sensed data got.The optical remote sensing data of the satellite include:
NDVI, LST and Albedo.
NDVI (Normalized Difference Vegetation Index, NDVI) i.e. normalized differential vegetation index, can
It is obtained with directly being calculated by Satellite Observations, formula are as follows: NDVI=(NIR-R)/(NIR+R), wherein NIR is remote sensing satellite
Near infrared band reflectivity, R are red band reflectivity, and the index value is between -1 and 1: 0 represents the region does not have substantially
Vegetation growth；Negative value represents the region of non-vegetative coverage；Between value 0-1, the bigger area coverage for representing vegetation of number is more
Greatly, the amount of vegetation is more, therefore NDVI can reflect influence of the backgrounds such as soil moisture to plant canopy, and NDVI is also microwave simultaneously
It is to influence soil moisture product remote-sensing inversion with one of the key parameter of estimation vegetation water content in remote sensing soil moisture retrieval
Key variables.
LST (Land Surface Temperature) i.e. surface temperature makes ground after earth's surface absorbs sunlamp radiation energy
The gain of heat, the temperature, that is, surface temperature on ground.In identical solar radiation, the variation of LST generally depends on specific heat capacity, and
Soil moisture is the key element for influencing specific heat capacity, and LST increases with soil moisture and reduced, and therefore, LST is soil dry-wet shape
The important indication index of condition.
The total radiation energy and incident total radiation energy that all directions are emitted on Albedo, that is, unit time, unit area it
Than, earth's surface is reacted externally come the albedo radiated, moisture is significantly higher than other soil elements to the absorbability of external radiation,
Therefore, Albedo is also one of the index for reacting soil dry-wet situation.
Since the optical remote sensing data of satellite are easy to be influenced missing data by weather etc., lead to the optical remote sensing of satellite
The space-time of data is discontinuous.The present embodiment needs the light by the discontinuous optical remote sensing data prediction of space-time for space and time continuous
Learn remotely-sensed data.
The microwave remote sensing soil moisture measurements data refer to that satellite carries out the soil moisture data of remote sensing using microwave, by
The influence of the transmission characteristic of remote sensing microwave is easy to be caused to count by the various impression for hindering microwave transmission such as geographic factor
According to missing.The microwave remote sensing soil moisture measurements training data refers to for instructing to the neural network constructed in subsequent step
Experienced sample set needs in microwave remote sensing soil moisture measurements data, obtains the data for meeting sample conditions, as microwave remote sensing soil
Earth humidity training data.
Step S200 determines the light of predeterminable area according to the microwave remote sensing soil moisture measurements training data of the predeterminable area
Learn remote sensing training data.
Specifically, according to the microwave remote sensing soil moisture measurements training data of predeterminable area, in the optical remote sensing number of space and time continuous
In, the data with identical geography information and temporal information are chosen, that is, selected space and time continuous optical remote sensing training data
With the microwave remote sensing soil moisture measurements training data of the predeterminable area, there is space-time consistency.
Step S300, by the optical remote sensing training data of the predeterminable area, the three-dimensional geographic information of the predeterminable area
It is determined as inputting information with the timeline information of the predeterminable area, soil moisture is determined as output information, is constructed wait train
Backpropagation neural network.
Specifically, the three-dimensional geographic information is three-dimensional geographic information, including elevation (Digital Elevation
Model, DEM), longitude (Latitude, Lat), latitude (Longitude, Lon))；Timeline information (the Day of
Year,DOY).The space and time continuous optical remote sensing training data of the predeterminable area includes: NDVI, LST and Albedo.
Artificial neural network (Artificial Neural Networks, abbreviation ANN) is substantially single based on human brain
The modeling of member-neuron be coupled, simulate human brain nervous system, being formed a kind of has study, association, memory and pattern-recognition
The manual system of equal Intelligent Information Processing.Including an input layer, multiple hidden layers and an output layer.Back-propagating nerve net
Network (Back Propagation Neural Network, BPNN) is a kind of artificial neural network of classics, is had from group
It knits, self study and nonlinear feature, is widely used in nonlinear system, to analyze the influence of more elements.After BPNN passes through
Keep the mean square error (RMSE) between predicted value and target value minimum to propagation algorithm.Theoretically, enough in neuron
In the case of, approaching for any nonlinear function may be implemented in three-layer neural network.
Using above-mentioned seven parameters as the input information of backpropagation neural network to be trained, soil moisture is determined as defeated
Information out constructs backpropagation neural network to be trained, for example, the backpropagation neural network to be trained can be middle layer
10 layers of neural network and the neural network of 7-10-1, the number of plies of middle layer may be other numbers of plies for meeting calculating demand.
Step S400, using the microwave remote sensing soil moisture measurements training data of the predeterminable area as training sample, described in training
Backpropagation neural network to be trained obtains soil moisture backpropagation neural network.
Specifically, to the backpropagation neural network to be trained so that the soil moisture data that it is exported after training with
It is identical between the microwave remote sensing soil moisture measurements training data of the predeterminable area, or meeting within certain demand error range,
It can terminate to train, obtain soil moisture backpropagation neural network.
Step S500, the space and time continuous optical remote sensing data of the predeterminable area, the three-dimensional of the predeterminable area is geographical
The timeline information of information and the predeterminable area inputs the soil moisture backpropagation neural network, obtains predeterminable area
Soil moisture data.
Specifically, the soil moisture backpropagation neural network after training, can input above-mentioned seven inputs information
On the basis of, provide the actual soil moisture data of fitting, therefore, by the space and time continuous optical remote sensing data of the predeterminable area,
After the timeline information of the three-dimensional geographic information of the predeterminable area and the predeterminable area re-enters, so that it may get pre-
If the soil moisture data in region, and the soil moisture data of the predeterminable area is the soil moisture data of space and time continuous.
Soil moisture method for reconstructing based on multi- source Remote Sensing Data data provided by the present embodiment, using optical remote sensing product and
Three-dimensional geographic information and timeline information construct backpropagation neural network, with preset microwave soil moisture training data
For sample set, after being trained to the backpropagation neural network, obtain soil moisture backpropagation neural network, then by when
The empty continuous optical remote sensing data input soil moisture backpropagation neural network, obtains continuous soil moisture number
According to.Pass through the input of neural network, the optical remote sensing data of space and time continuous and three-dimensional geographic information, timeline information, pre- micrometer
The missing values of wave remote sensing soil moisture data obtain the microwave remote sensing soil data of space and time continuous.
The microwave remote sensing soil moisture measurements training data for obtaining the predeterminable area in one of the embodiments, packet
It includes: obtaining the microwave remote sensing soil moisture measurements initial data of predeterminable area；In the microwave remote sensing soil moisture measurements initial data, sieve
It selects not by snow cover and surface temperature is greater than the pel data of zero degrees celsius；The pel data filtered out is determined
For microwave remote sensing soil moisture measurements training data.
Specifically, need to measure the accuracy of soil moisture when choosing microwave remote sensing soil moisture measurements training data,
If thering is snow cover or surface temperature to be less than zero degrees celsius, then it is assumed that the water in soil is solid-state, needs to be measured with water content, and
It cannot be expressed with the humidity for measuring liquid water.
In the present embodiment, by the screening to microwave remote sensing soil moisture measurements training data, make the back-propagating mind of building
There is accurate training sample through network, to improve the accuracy rate of the backpropagation neural network output valve after training.
Backpropagation neural network to be trained described in the training in one of the embodiments, is included in described in M*N
Within the scope of the grid cell size of microwave remote sensing soil moisture measurements training data, the backpropagation neural network to be trained is trained, wherein M
It is positive integer with N.
Specifically, for example, the range scale of the pixel of 4*4 microwave remote sensing soil moisture measurements training data can be chosen,
The range scale of the pixel of 8*8 or 4*8 microwave remote sensing soil moisture measurements training data can be selected according to the actual situation, it is right
The backpropagation neural network to be trained is trained.Due to soil moisture data, there is certain continuity in practice,
It will receive weather again in a big way or discontinuity occur in the influence of geographic factor.The M*N grid cell size range
It chooses, is chosen according to practical situation, training effectiveness can be improved and will not be because of the grid cell size range of selection
It is excessive, so that the error after training increases.
In the present embodiment, the backpropagation neural network is trained within the scope of certain grid cell size, i.e.,
The continuity that output data can be improved, having, which can guarantee, to cause the accuracy rate of output data to reduce because range is excessive.
Fig. 2 is the flow diagram of the soil moisture method for reconstructing based on multi- source Remote Sensing Data data in another embodiment,
Soil moisture method for reconstructing based on multi- source Remote Sensing Data data as shown in Figure 2 includes:
Step S110 obtains the optical remote sensing initial data of predeterminable area, and the optical remote sensing initial data includes vegetation
Index initial data, surface temperature initial data and surface albedo initial data.
Specifically, the optical remote sensing initial data includes vegetation index initial data, surface temperature initial data and ground
Table albedo initial data has space-time discontinuity.
Step S120, by the vegetation index initial data of the predeterminable area by being rebuild based on time series frequency analysis
Algorithm calculates the space and time continuous vegetation index of predeterminable area；By the surface temperature initial data of the predeterminable area by being based on
The algorithm for reconstructing of reference sequences calculates the space and time continuous surface temperature of predeterminable area；By the surface albedo of the predeterminable area
Initial data calculates the space and time continuous surface albedo of predeterminable area by the algorithm for reconstructing based on spatio-temporal filtering.
Specifically, the space-time that data prediction includes optical articles (NDVI, LST, Albedo) is rebuild and spatial clustering.
NDVI, that is, normalized differential vegetation index space-time is rebuild, and the method based on time series frequency analysis is used
(Harmonic Analysis of Time Series, Hants):
Wherein, NDVI is original NDVI sequence,For the NDVI sequence after reconstruction, ε is error sequence, tj(j=1,
2 ... N) it is the time that NDVI takes, N is maximum observation number, and nf is that frequency is fiComponent quantity, a0, ai, biIt is coefficient.
The space-time of LST is rebuild, based on an assumption that in similar vegetation growth, time closer two scapes LST shadow
It, then can be by establishing with reference to the recurrence between LST image and LST image to be filled as the surface temperature of same place changes linearly
Relational expression predicts the LST of image to be filled by the known LST with reference to image.It mainly include three steps: first to image
Classify, and 0,0.1,0.2 is split to image using NDVI ..., 1；In the same section NDVI, to different images
The LST of (with reference to image and image to be filled) carries out logistic regression (quadratic polynomial recurrence), obtains using LST pairs of reference image
The expression formula of image LST prediction to be filled, and go to predict shadow to be filled using the LST for returning obtained expression formula and reference image
As the LST of missing；LST after prediction is post-processed, inconsistency spatially is eliminated.
The space-time of Albedo is rebuild, using spatio-temporal filtering method, it is assumed that there is correlation between the albedo of temporally adjacent pixel
Property, the albedo of kth day can be by the Albedo linear expression of kth+△ k days:
αk=a△kαk+△k+b△k+e△k
For any pixel, regression coefficient a△k、b△kAnd e△kIt can be obtained by many years observation by maximal possibility estimation.
Step S130, by the space and time continuous earth's surface of the space and time continuous vegetation index of the predeterminable area, the predeterminable area
The surface albedo of temperature and the predeterminable area is determined as the space and time continuous optical remote sensing data of predeterminable area.
In the present embodiment, after the discontinuous optical remote sensing data of space-time being carried out space-time reconstruction, space and time continuous is got
After optical remote sensing data, trained backpropagation neural network is inputted, the soil moisture data of space and time continuous can be obtained,
The space-time algorithm for reconstructing of each optical remote sensing data, i.e., so that the space-time for the soil moisture data that neural computing obtains connects
Continuous property, and can guarantee the accuracy for the soil moisture data being calculated.
Fig. 3 is the flow diagram of the soil moisture method for reconstructing based on multi- source Remote Sensing Data data in further embodiment,
Soil moisture method for reconstructing based on multi- source Remote Sensing Data data as shown in Figure 3 includes:
Step S310 will be described according to the grid cell size of the microwave remote sensing soil moisture measurements training data of the predeterminable area
Vegetation index, surface temperature and surface albedo carry out resampling, obtain input vegetation index, input surface temperature and input ground
Table albedo.
Specifically, being needed between the input data and output data to constructed backpropagation neural network to be trained
Linear module having the same is wanted, just can be carried out calculating.Therefore, it is necessary to which the input data of resolution ratio will have been unified, according to output
The grid cell sizes of data carries out resampling, for example, resampling is to microwave remote sensing product grid cell size (1km to 25km).
Step S320, by the input vegetation index, input surface temperature, input surface albedo, the predeterminable area
Three-dimensional geographic information and the predeterminable area timeline information, be determined as input information.
In the present embodiment, by the unification to optical remote sensing training data progress spatial resolution, and according to microwave
Remote sensing training data carries out resampling, and the backpropagation neural network of building is made to have the input and output number of equal resolution scale
According to the accuracy rate of the backpropagation neural network output valve after improving training.
Since between different satellites or between the different optical remote sensing products of same satellite, all there is different resolutions
Rate needs to calculate the unification of all data to identical resolution ratio latitude.It in one of the embodiments, will be described default
The space and time continuous vegetation index training data in region, the space and time continuous surface temperature training data of predeterminable area and predeterminable area
Space and time continuous surface albedo training data calculates separately the vegetation with same spatial resolution by space average algorithm
Index, surface temperature and surface albedo, to guarantee the accuracy of subsequent calculated result,
Fig. 4 is the structural schematic diagram of the soil moisture reconstructing system based on multi- source Remote Sensing Data data in one embodiment, such as
Soil moisture reconstructing system shown in Fig. 4 based on multi- source Remote Sensing Data data includes:
Optical remote sensing data and microwave training data obtain module 100, for obtaining the space and time continuous optics of predeterminable area
The microwave remote sensing soil moisture measurements training data of remotely-sensed data and the predeterminable area.
Optical remote sensing training data obtains module 200, for being instructed according to the microwave remote sensing soil moisture measurements of the predeterminable area
Practice data, determines the optical remote sensing training data of predeterminable area.
Neural network constructs module 300, for by the optical remote sensing training data of the predeterminable area, the predeterminable area
Three-dimensional geographic information and the predeterminable area timeline information be determined as input information, by soil moisture be determined as output letter
Breath, constructs backpropagation neural network to be trained.
Neural metwork training module 400, for being instruction with the microwave remote sensing soil moisture measurements training data of the predeterminable area
Practice sample, the training backpropagation neural network to be trained obtains soil moisture backpropagation neural network；It is also used in M*
Within the scope of the grid cell size of N number of microwave remote sensing soil moisture measurements training data, the training back-propagating nerve net to be trained
Network, wherein M and N is positive integer.
Soil moisture data obtains module 500, for by space and time continuous optical remote sensing data of the predeterminable area, described
The timeline information of the three-dimensional geographic information of predeterminable area and the predeterminable area inputs the soil moisture back-propagating nerve
Network obtains the soil moisture data of predeterminable area.
Soil moisture reconstructing system based on multi- source Remote Sensing Data data provided by the present embodiment, using optical remote sensing product and
Three-dimensional geographic information and timeline information construct backpropagation neural network, with preset microwave soil moisture training data
For sample set, after being trained to the backpropagation neural network, obtain soil moisture backpropagation neural network, then by when
The empty continuous optical remote sensing data input soil moisture backpropagation neural network, obtains continuous soil moisture number
According to.Pass through the input of neural network, the optical remote sensing data of space and time continuous and three-dimensional geographic information, timeline information, pre- micrometer
The missing values of wave remote sensing soil moisture data obtain the microwave remote sensing soil data of space and time continuous.
Fig. 5 is the structural schematic diagram of the soil moisture reconstructing system based on multi- source Remote Sensing Data data in another embodiment, such as
Soil moisture reconstructing system shown in fig. 5 based on multi- source Remote Sensing Data data includes:
Microwave training data acquiring unit 110, for obtaining the microwave remote sensing soil moisture measurements initial data of predeterminable area；?
In the microwave remote sensing soil moisture measurements initial data, filters out not by snow cover and surface temperature is greater than the picture of zero degrees celsius
Metadata；The pel data filtered out is determined as microwave remote sensing soil moisture measurements training data.
Optical remote sensing data capture unit 120, for obtaining the optical remote sensing initial data of predeterminable area, the optics is distant
Feeling initial data includes vegetation index initial data, surface temperature initial data and surface albedo initial data；It will be described pre-
If the vegetation index initial data in region is by being based on time series frequency analysis algorithm for reconstructing, the space-time for calculating predeterminable area connects
Continuous vegetation index；By the surface temperature initial data of the predeterminable area by the algorithm for reconstructing based on reference sequences, calculate pre-
If the space and time continuous surface temperature in region；The surface albedo initial data of the predeterminable area is passed through based on spatio-temporal filtering
Algorithm for reconstructing calculates the space and time continuous surface albedo of predeterminable area；By the space and time continuous vegetation index of the predeterminable area, institute
The space and time continuous surface temperature of predeterminable area and the surface albedo of the predeterminable area are stated, the space-time for being determined as predeterminable area connects
Continuous optical remote sensing data.
In the present embodiment, by the screening to microwave remote sensing soil moisture measurements training data, make the back-propagating mind of building
There is accurate training sample through network, to improve the accuracy rate of the backpropagation neural network output valve after training.Pass through
The unification of spatial resolution is carried out to optical remote sensing training data, and resampling is carried out according to microwave remote sensing training data, is made
The backpropagation neural network of building has the inputoutput data of equal resolution scale, the back-propagating mind after improving training
Accuracy rate through network output valve.
Fig. 6 is the structural schematic diagram of the soil moisture reconstructing system based on multi- source Remote Sensing Data data in further embodiment,
Soil moisture reconstructing system as shown in FIG. 6 based on multi- source Remote Sensing Data data includes:
Resampling computing unit 310, for the picture according to the microwave remote sensing soil moisture measurements training data of the predeterminable area
First vegetation index, the first surface temperature and the first surface albedo are carried out resampling by first scale, obtain input vegetation
Index, input surface temperature and input surface albedo；
Information determination unit 320 is inputted, is used for the input vegetation index, input surface temperature, the input earth's surface reflection of light
The timeline information of rate, the three-dimensional geographic information of the predeterminable area and the predeterminable area is determined as inputting information.
In the present embodiment, after the discontinuous optical remote sensing data of space-time being carried out space-time reconstruction, space and time continuous is got
After optical remote sensing data, trained backpropagation neural network is inputted, the soil moisture data of space and time continuous can be obtained,
The space-time algorithm for reconstructing of each optical remote sensing data, i.e., so that the space-time for the soil moisture data that neural computing obtains connects
Continuous property, and can guarantee the accuracy for the soil moisture data being calculated.
Each technical characteristic of embodiment described above can be combined arbitrarily, for simplicity of description, not to above-mentioned reality
It applies all possible combination of each technical characteristic in example to be all described, as long as however, the combination of these technical characteristics is not deposited
In contradiction, all should be considered as described in this specification.
The embodiments described above only express several embodiments of the present invention, and the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to protection of the invention
Range.Therefore, the scope of protection of the patent of the invention shall be subject to the appended claims.
Claims (8)
1. a kind of soil moisture method for reconstructing based on multi- source Remote Sensing Data data, which is characterized in that the described method includes:
Obtain the space and time continuous optical remote sensing data of predeterminable area and the microwave remote sensing soil moisture measurements training number of the predeterminable area
According to；
According to the microwave remote sensing soil moisture measurements training data of the predeterminable area, the optical remote sensing training number of predeterminable area is determined
According to；
By the optical remote sensing training data of the predeterminable area, the three-dimensional geographic information and the predeterminable area of the predeterminable area
Timeline information be determined as input information, soil moisture is determined as output information, constructs back-propagating nerve net to be trained
Network；
Using the microwave remote sensing soil moisture measurements training data of the predeterminable area as training sample, the training back-propagating to be trained
Neural network obtains soil moisture backpropagation neural network；
By the space and time continuous optical remote sensing data of the predeterminable area, the three-dimensional geographic information of the predeterminable area and described default
The timeline information in region inputs the soil moisture backpropagation neural network, obtains the soil moisture data of predeterminable area；
The microwave remote sensing soil moisture measurements training data for obtaining the predeterminable area, comprising:
Obtain the microwave remote sensing soil moisture measurements initial data of predeterminable area；
In the microwave remote sensing soil moisture measurements initial data, filter out not by snow cover and surface temperature to be greater than zero Celsius
The pel data of degree；
The pel data filtered out is determined as microwave remote sensing soil moisture measurements training data.
2. the soil moisture method for reconstructing based on multi- source Remote Sensing Data data according to claim 1, which is characterized in that the acquisition
The space and time continuous optical remote sensing data of predeterminable area, comprising:
Obtain predeterminable area optical remote sensing initial data, the optical remote sensing initial data include vegetation index initial data,
Surface temperature initial data and surface albedo initial data；
By the vegetation index initial data of the predeterminable area by being based on time series frequency analysis algorithm for reconstructing, calculate default
The space and time continuous vegetation index in region；The surface temperature initial data of the predeterminable area is passed through into the reconstruction based on reference sequences
Algorithm calculates the space and time continuous surface temperature of predeterminable area；The surface albedo initial data of the predeterminable area is passed through into base
In the algorithm for reconstructing of spatio-temporal filtering, the space and time continuous surface albedo of predeterminable area is calculated；
By the space and time continuous vegetation index of the predeterminable area, the space and time continuous surface temperature of the predeterminable area and described default
The surface albedo in region is determined as the space and time continuous optical remote sensing data of predeterminable area.
3. the soil moisture method for reconstructing based on multi- source Remote Sensing Data data according to claim 2, which is characterized in that described by institute
State the space and time continuous optical remote sensing training data of predeterminable area, the three-dimensional geographic information and the predeterminable area of the predeterminable area
Timeline information be determined as input information, comprising:
According to the grid cell size of the microwave remote sensing soil moisture measurements training data of the predeterminable area, by the vegetation index, earth's surface
Temperature and surface albedo carry out resampling, obtain input vegetation index, input surface temperature and input surface albedo；
By the input vegetation index, input surface temperature, the three-dimensional geographic information for inputting surface albedo, the predeterminable area
With the timeline information of the predeterminable area, it is determined as inputting information.
4. the soil moisture method for reconstructing based on multi- source Remote Sensing Data data according to claim 3, which is characterized in that the training
The backpropagation neural network to be trained, comprising:
Within the scope of the grid cell size of the M*N microwave remote sensing soil moisture measurements training datas, to biography after training described in training
Neural network is broadcast, wherein M and N is positive integer.
5. a kind of soil moisture reconstructing system based on multi- source Remote Sensing Data data characterized by comprising
Optical remote sensing data and microwave training data obtain module, for obtaining the space and time continuous optical remote sensing data of predeterminable area
With the microwave remote sensing soil moisture measurements training data of the predeterminable area；
Optical remote sensing training data obtains module, for the microwave remote sensing soil moisture measurements training data according to the predeterminable area,
Determine the optical remote sensing training data of predeterminable area；
Neural network constructs module, for by the three-dimensional of the optical remote sensing training data of the predeterminable area, the predeterminable area
The timeline information of geography information and the predeterminable area is determined as inputting information, and soil moisture is determined as output information, structure
Build backpropagation neural network to be trained；
Neural metwork training module, for the microwave remote sensing soil moisture measurements training data using the predeterminable area as training sample,
The training backpropagation neural network to be trained, obtains soil moisture backpropagation neural network；
Soil moisture data obtains module, for by the space and time continuous optical remote sensing data of the predeterminable area, the preset areas
The timeline information of the three-dimensional geographic information in domain and the predeterminable area inputs the soil moisture backpropagation neural network,
Obtain the soil moisture data of predeterminable area；
The optical remote sensing data and microwave training data obtain module, comprising:
Microwave training data acquiring unit, for obtaining the microwave remote sensing soil moisture measurements initial data of predeterminable area；Described micro-
In wave remote sensing soil moisture initial data, filters out not by snow cover and surface temperature is greater than the pixel number of zero degrees celsius
According to；The pel data filtered out is determined as microwave remote sensing soil moisture measurements training data.
6. the soil moisture reconstructing system based on multi- source Remote Sensing Data data according to claim 5, which is characterized in that the optics
Remotely-sensed data and microwave training data obtain module, comprising:
Optical remote sensing data capture unit, for obtaining the optical remote sensing initial data of predeterminable area, the optical remote sensing is original
Data include vegetation index initial data, surface temperature initial data and surface albedo initial data；By the predeterminable area
Vegetation index initial data by be based on time series frequency analysis algorithm for reconstructing, calculate the space and time continuous vegetation of predeterminable area
Index；By the surface temperature initial data of the predeterminable area by the algorithm for reconstructing based on reference sequences, predeterminable area is calculated
Space and time continuous surface temperature；The surface albedo initial data of the predeterminable area is calculated by the reconstruction based on spatio-temporal filtering
Method calculates the space and time continuous surface albedo of predeterminable area；By the space and time continuous vegetation index of the predeterminable area, described default
The space and time continuous surface temperature in region and the surface albedo of the predeterminable area, are determined as the space and time continuous optics of predeterminable area
Remotely-sensed data.
7. the soil moisture reconstructing system based on multi- source Remote Sensing Data data according to claim 6, which is characterized in that the nerve
Network struction module, comprising:
Resampling computing unit, for the grid cell size according to the microwave remote sensing soil moisture measurements training data of the predeterminable area,
By the vegetation index, surface temperature and surface albedo carry out resampling, obtain input vegetation index, input surface temperature and
Input surface albedo；
Information determination unit is inputted, is used for the input vegetation index, input surface temperature, inputs surface albedo, described
The timeline information of the three-dimensional geographic information of predeterminable area and the predeterminable area is determined as inputting information.
8. the soil moisture reconstructing system based on multi- source Remote Sensing Data data according to claim 7, it is characterised in that:
The neural metwork training module is also used to the grid cell size in the M*N microwave remote sensing soil moisture measurements training datas
In range, the training backpropagation neural network to be trained, wherein M and N is positive integer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611101527.9A CN106779067B (en) | 2016-12-02 | 2016-12-02 | Soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611101527.9A CN106779067B (en) | 2016-12-02 | 2016-12-02 | Soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106779067A CN106779067A (en) | 2017-05-31 |
| CN106779067B true CN106779067B (en) | 2019-04-05 |
Family
ID=58884354
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611101527.9A Active CN106779067B (en) | 2016-12-02 | 2016-12-02 | Soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106779067B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107389895B (en) * | 2017-06-08 | 2019-08-30 | 环境保护部卫星环境应用中心 | Soil moisture mixed type remote sensing inversion method and system |
| CN107862255B (en) * | 2017-10-23 | 2020-12-22 | 交通运输部科学研究院 | Wetland information extraction and ecological sensitivity evaluation method based on microwave remote sensing and optical remote sensing technology |
| CN108157147A (en) * | 2018-01-22 | 2018-06-15 | 五邑大学 | A kind of intelligent irrigation system based on 3 D stereo soil moisture figure |
| CN108764461B (en) * | 2018-05-24 | 2020-04-14 | 西安电子科技大学 | Earthen site temperature prediction method based on solar term characteristics |
| CN108717044B (en) * | 2018-05-24 | 2021-07-30 | 青海师范大学 | Surface soil water content satellite remote sensing estimation method for removing vegetation coverage influence |
| CN109522516B (en) * | 2018-09-30 | 2020-06-09 | 广州地理研究所 | Soil humidity detection method and device based on random forest regression algorithm and electronic equipment |
| CN111652404B (en) * | 2020-02-06 | 2023-04-28 | 北京师范大学 | All-weather earth surface temperature inversion method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105116399A (en) * | 2015-08-27 | 2015-12-02 | 电子科技大学 | Soil humidity inversion method aiming for ultra wide band radar echo |
| CN106169058A (en) * | 2015-12-25 | 2016-11-30 | 中国石油大学(华东) | Pixel LST evaluation method under a kind of cloud based on microwave remote sensing with space time information |
-
2016
- 2016-12-02 CN CN201611101527.9A patent/CN106779067B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105116399A (en) * | 2015-08-27 | 2015-12-02 | 电子科技大学 | Soil humidity inversion method aiming for ultra wide band radar echo |
| CN106169058A (en) * | 2015-12-25 | 2016-11-30 | 中国石油大学(华东) | Pixel LST evaluation method under a kind of cloud based on microwave remote sensing with space time information |
Non-Patent Citations (1)
| Title |
|---|
| "基于云参数干旱遥感监测模型与集合卡尔曼滤波的土壤湿度同化研究";杨娜;《中国博士学位论文全文数据库 农业科技辑》;20110815;第1-7章 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106779067A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106779067B (en) | Soil moisture method for reconstructing and system based on multi- source Remote Sensing Data data | |
| Xu et al. | A new land surface temperature fusion strategy based on cumulative distribution function matching and multiresolution Kalman filtering | |
| Huang et al. | Jointly assimilating MODIS LAI and ET products into the SWAP model for winter wheat yield estimation | |
| CN111678866B (en) | Soil water content inversion method for multi-model ensemble learning | |
| Hank et al. | Using a remote sensing-supported hydro-agroecological model for field-scale simulation of heterogeneous crop growth and yield: Application for wheat in central Europe | |
| Zhan et al. | Sharpening thermal imageries: A generalized theoretical framework from an assimilation perspective | |
| Zhuo et al. | Crop yield prediction using MODIS LAI, TIGGE weather forecasts and WOFOST model: A case study for winter wheat in Hebei, China during 2009–2013 | |
| Wang et al. | Monitoring maize growth conditions by training a BP neural network with remotely sensed vegetation temperature condition index and leaf area index | |
| CN105004320B (en) | A kind of high score satellite data land table vegetation coverage inversion method and system | |
| Xu et al. | Monitoring crop water content for corn and soybean fields through data fusion of MODIS and Landsat measurements in Iowa | |
| CN103345707A (en) | Crop maturation stage remote sensing prediction method based on multi-source remote sensing data | |
| Zhang et al. | Estimating wheat yield by integrating the WheatGrow and PROSAIL models | |
| WO2023088366A1 (en) | Method for jointly estimating soil profile salinity by using time-series remote sensing image | |
| Feng et al. | Estimating surface heat and water vapor fluxes by combining two-source energy balance model and back-propagation neural network | |
| Wu et al. | Estimation of cotton canopy parameters based on unmanned aerial vehicle (UAV) oblique photography | |
| Wang et al. | Estimation of surface soil moisture by combining a structural equation model and an artificial neural network (SEM-ANN) | |
| Su et al. | An integrative information aqueduct to close the gaps between satellite observation of water cycle and local sustainable management of water resources | |
| Liu et al. | Comparison of two satellite-based soil moisture reconstruction algorithms: A case study in the state of Oklahoma, USA | |
| Lou et al. | An effective method for canopy chlorophyll content estimation of marsh vegetation based on multiscale remote sensing data | |
| Wang et al. | Research on cropping intensity mapping of the Huai River Basin (China) based on multi-source remote sensing data fusion | |
| Han et al. | Integrating an attention-based deep learning framework and the SAFY-V model for winter wheat yield estimation using time series SAR and optical data | |
| Zhou et al. | Estimation of aboveground biomass of senescence grassland in China's arid region using multi-source data | |
| Zhang et al. | Estimation of quasi-full spatial coverage soil moisture with fine resolution in China from the combined use of ERA5-Land reanalysis and TRIMS land surface temperature product | |
| Bao et al. | Forecasting spring maize yield using vegetation indices and crop phenology metrics from UAV observations | |
| Li et al. | Synergetic use of DEM derivatives, Sentinel-1 and Sentinel-2 data for mapping soil properties of a sloped cropland based on a two-step ensemble learning method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |