CN112161998B - Soil water content measuring method and device, electronic equipment and storage medium - Google Patents
Soil water content measuring method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112161998B CN112161998B CN202010911955.8A CN202010911955A CN112161998B CN 112161998 B CN112161998 B CN 112161998B CN 202010911955 A CN202010911955 A CN 202010911955A CN 112161998 B CN112161998 B CN 112161998B
- Authority
- CN
- China
- Prior art keywords
- soil
- remote sensing
- moisture content
- input
- sensing image
- 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 261
- 238000000034 method Methods 0.000 title claims abstract description 54
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 50
- 230000006870 function Effects 0.000 claims description 34
- 230000010287 polarization Effects 0.000 claims description 32
- 230000004913 activation Effects 0.000 claims description 27
- 238000000691 measurement method Methods 0.000 claims description 13
- NEBFIUZIGRTIFY-BJDJZHNGSA-N Ala-Met-Ser-Arg Chemical compound CSCC[C@H](NC(=O)[C@H](C)N)C(=O)N[C@@H](CO)C(=O)N[C@H](C(O)=O)CCCNC(N)=N NEBFIUZIGRTIFY-BJDJZHNGSA-N 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000003908 quality control method Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 2
- 230000000875 corresponding effect Effects 0.000 description 40
- 238000012549 training Methods 0.000 description 18
- 238000010606 normalization Methods 0.000 description 12
- 238000004891 communication Methods 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000002474 experimental method Methods 0.000 description 6
- 239000000443 aerosol Substances 0.000 description 5
- 230000015556 catabolic process Effects 0.000 description 5
- 238000006731 degradation reaction Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000008034 disappearance Effects 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000013507 mapping Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 238000011144 upstream manufacturing Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000012886 linear function Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 230000037361 pathway Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000002759 z-score normalization Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000004439 roughness measurement Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N22/00—Investigating or analysing materials by the use of microwaves or radio waves, i.e. electromagnetic waves with a wavelength of one millimetre or more
- G01N22/04—Investigating moisture content
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N2021/1793—Remote sensing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Chemical & Material Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biochemistry (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Electromagnetism (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
One or more embodiments of the present specification provide a soil moisture content measuring method including: for an area, acquiring a passive microwave remote sensing image and soil roughness data corresponding to the area; extracting brightness temperature data from the passive microwave remote sensing image; taking the bright temperature data and the soil roughness data as input features of a soil water content prediction model; determining the soil moisture content of the area according to the input characteristics based on the soil moisture content prediction model; wherein the soil moisture content prediction model is a residual network model for determining soil moisture content based on input features. The specification also provides a soil moisture content measuring device, electronic equipment and a computer readable medium corresponding to the soil moisture content measuring method.
Description
Technical Field
One or more embodiments of the present disclosure relate to the field of remote sensing image inversion technology, and in particular, to a soil moisture content measurement method, a soil moisture content measurement device, an electronic apparatus, and a computer readable storage medium.
Background
Soil moisture is the most important component in the land ecological system, is an important parameter in researches of hydrology, weather, agriculture and the like, and has a particularly important significance in a crop estimation model and agricultural drought monitoring research. Therefore, how to effectively measure the soil moisture content with high precision is the most concern of current research.
The traditional soil moisture content measuring method is obtained by taking soil samples from the ground and measuring and weighing the soil samples in a laboratory. Although the method has higher precision, the method is time-consuming and labor-consuming, and is difficult to meet the large-scale soil water content measurement task.
With the development of remote sensing technology, a large-scale soil water content measurement is made possible. At present, optical remote sensing and microwave remote sensing are two main flow directions for inversion of soil water content based on remote sensing means. The soil moisture content inversion method based on optical remote sensing mainly builds a relation between a vegetation index, surface temperature and soil moisture content to invert the soil moisture content. However, the optical remote sensing is easily affected by cloud, overcast and rainy, aerosol and the like, so that the soil water content obtained by inversion has a certain influence on the precision. The soil water content inversion method based on microwave remote sensing mainly inverts the soil water content through multi-frequency bright temperature combination. Because the microwave remote sensing is less affected by cloud, overcast and rainy, aerosol and the like, the inversion of the soil water content by utilizing the microwave remote sensing is one of the relatively effective soil water content inversion methods at present.
Disclosure of Invention
In view of this, one or more embodiments of the present disclosure provide a soil moisture content measurement method that can determine a soil moisture content of an area according to passive microwave remote sensing images and soil roughness data.
The soil moisture content measuring method according to one or more embodiments of the present specification may include: for an area, acquiring a passive microwave remote sensing image and soil roughness data corresponding to the area; extracting brightness temperature data from the passive microwave remote sensing image; taking the bright temperature data and the soil roughness data as input features of a soil water content prediction model; and determining a soil moisture content of the area based on the soil moisture content prediction model according to the input features; wherein the soil moisture content prediction model is a residual network model for determining soil moisture content based on input features.
In some embodiments of the present disclosure, acquiring the passive microwave remote sensing image corresponding to the region includes: and acquiring a passive microwave remote sensing image shot by an advanced microwave scanning radiometer 2 satellite (AMSR 2) and/or a passive microwave remote sensing image shot by a soil moisture and ocean salinity satellite (SMOS) corresponding to the area.
In some embodiments of the present disclosure, extracting the bright temperature data from the passive microwave remote sensing image includes: extracting C-band bright temperature data from the passive microwave remote sensing image corresponding to the region shot by the AMSR 2; and/or extracting the brightness temperature data of the L wave band from the passive microwave remote sensing image corresponding to the region shot by the SMOS.
In some embodiments of the present description, the above method may further include: eliminating radio interference of C wave Duan Liangwen data by using bright temperature data of other wave bands except the C wave band in the passive microwave remote sensing image shot by the AMSR 2; and/or removing the pixel data when the electromagnetic interference index of a certain pixel is greater than a preset threshold value by utilizing an L3-level electromagnetic interference quality control mechanism in the SMOS.
In some embodiments of the present description, the above method may further include: extracting backscattering coefficients of C wave band H polarization and V polarization from the passive microwave remote sensing image shot by the AMSR2 as one of input features of the soil water content prediction model; and/or extracting the backscattering coefficients of L wave band H polarization and V polarization from the passive microwave remote sensing image shot by the SMOS as one of the input features of the soil water content prediction model.
In some embodiments of the present description, the above method may further include: and acquiring soil texture data corresponding to the region as one of input features of the soil water content prediction model.
In some embodiments of the present description, the residual network model comprises: the input layer, N residual units connected in a cascading manner and the output layer; wherein N is a natural number greater than 1;
the output of the input layer is connected to the input of the 1 st residual unit; the output of the nth residual unit is connected to the input of the (n+1) th residual unit, wherein N < N is 1-N; the output of the nth residual unit is connected to the input of the output layer; wherein,
the residual unit includes: at least two hidden layers, a characteristic superposition layer and an activation function layer which are connected in a cascading manner; the hidden layer is used for expanding the characteristics input by the input layer according to different dimensions and extracting high-dimensional characteristics; the characteristic overlapping layer is used for overlapping the output characteristics of the input layer or the previous residual error unit with the output characteristics of the last hidden layer connected in a cascading mode; the activation function of the activation function layer is ReLU (x) =max (0, x).
Corresponding to the above soil moisture content measuring method, one or more embodiments of the present disclosure further disclose a soil moisture content measuring device, including:
the characteristic acquisition module is used for acquiring a passive microwave remote sensing image and soil roughness data corresponding to an area;
the characteristic extraction module is used for extracting bright temperature data from the passive microwave remote sensing image, and taking the bright temperature data and the soil roughness data as input characteristics of a soil water content prediction model;
the prediction module is used for determining the soil moisture content of the area according to the input characteristics based on the soil moisture content prediction model; wherein the soil moisture content prediction model is a residual network model for determining soil moisture content based on input features.
One or more embodiments of the present specification also provide an electronic device, which may include: the soil moisture content measuring device comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the soil moisture content measuring method when executing the program.
One or more embodiments of the present specification also provide a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions for causing the computer to perform the above-described soil moisture content measurement method.
It can be seen that the soil moisture content measurement method adopts the residual network model obtained by the supervised training mode as the soil moisture content prediction model, can fully utilize the characteristic that the residual network is added with the input value again before the output value is activated, not only solves the problems of weak generalization capability, low precision and low reliability when other neural networks are adopted for soil moisture content prediction, but also can effectively solve the problems of gradient disappearance, network degradation and the like generated when the number of layers of the network is increased. Experiments prove that the residual network model is very suitable for predicting the water content of soil and has the characteristics of strong generalization capability, high precision and strong reliability.
In addition, in the soil moisture content measuring method, a passive microwave remote sensing image is used as one of input features of a soil moisture content prediction model. Because the microwave remote sensing image is less affected by cloud, overcast and rainy, aerosol and the like, the soil water content accuracy obtained by predicting the soil water content according to the passive microwave remote sensing image is higher.
Furthermore, besides the passive microwave remote sensing image, the soil moisture content measuring method also utilizes the soil roughness as the input of the soil moisture content prediction model, fully considers that the roughness of the soil has great influence on the radiation brightness temperature detected by the passive microwave remote sensing satellite, and can further improve the accuracy of the soil moisture content prediction.
Drawings
For a clearer description of one or more embodiments of the present description or of the solutions of the prior art, the drawings that are necessary for the description of the embodiments or of the prior art will be briefly described, it being apparent that the drawings in the description below are only one or more embodiments of the present description, from which other drawings can be obtained, without inventive effort, for a person skilled in the art.
FIG. 1 is a schematic flow chart of a method for measuring soil moisture content according to some embodiments of the present disclosure;
fig. 2 shows an internal structure of a residual unit according to one or more embodiments of the present disclosure.
FIG. 3 is a schematic flow chart of a method for measuring soil moisture content according to other embodiments of the present disclosure;
FIG. 4 is a graph of the predicted soil moisture content versus measured soil moisture content using the soil moisture content prediction method described in the examples herein;
FIG. 5 is a schematic diagram of a training process for a soil moisture content prediction model according to one or more embodiments of the present disclosure;
FIG. 6 is a schematic view illustrating an internal structure of a soil moisture content measuring device according to one or more embodiments of the present disclosure;
Fig. 7 is a schematic diagram of a hardware structure of an electronic device according to one or more embodiments of the present disclosure.
Detailed Description
For the purposes of promoting an understanding of the principles and advantages of the disclosure, reference will now be made to the embodiments illustrated in the drawings and specific language will be used to describe the same.
It is noted that unless otherwise defined, technical or scientific terms used in one or more embodiments of the present disclosure should be taken in a general sense as understood by one of ordinary skill in the art to which the present disclosure pertains. The use of the terms "first," "second," and the like in one or more embodiments of the present description does not denote any order, quantity, or importance, but rather the terms "first," "second," and the like are used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that elements or items preceding the word are included in the element or item listed after the word and equivalents thereof, but does not exclude other elements or items. The terms "connected" or "connected," and the like, are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", etc. are used merely to indicate relative positional relationships, which may also be changed when the absolute position of the object to be described is changed.
As mentioned above, soil moisture is the most important component of the land ecosystem, and is an important parameter in hydrologic, meteorological, agricultural and other studies, and is of particular importance in crop estimation models and agricultural drought monitoring studies. Therefore, how to effectively measure the soil moisture content with high precision is the most concern of current research.
One or more embodiments of the present disclosure provide a soil moisture content measurement method that can determine a soil moisture content of an area based on passive microwave remote sensing images and soil roughness data.
Fig. 1 shows a flow chart of an implementation of a soil moisture content measurement method according to one or more embodiments of the present disclosure. As shown in fig. 1, the method may include:
in step 102, a passive microwave remote sensing image and soil roughness data corresponding to an area are acquired for the area.
In the embodiments of the present description, the above-mentioned region may be any specified region of the earth's surface, for example, a region upstream of a black river basin, or the like.
In an embodiment of the present disclosure, the passive microwave remote sensing image may be obtained by at least one of the following two approaches.
Pathway 1: obtained by advanced microwave scanning radiometer 2 satellite (The Advanced Microwave Scanning Radiometer-2, amsr2).
The AMSR2 sensor was installed on GCOM-W1 satellite in japan and launched into orbit in 2012. The transit time per day is approximately 1:30 a.m. local (derailment) and 13:30 a.m. afternoon (derailment). Wherein its track-up and track-down data can cover a large part of the world in two days, except for the polar region. The observation incident angle is 55 degrees, the working frequency comprises 14 dual polarized channels (6.925 GHz, 7.3GHz, 10.65GHz, 18.7GHz, 23.8GHz, 36.5GHz and 89 GHz), and the multi-band earth surface observation bright temperature can be provided. In the embodiment of the present disclosure, the passive microwave remote sensing image corresponding to the region captured by the AMSR2 may be obtained as the passive microwave remote sensing image described in the step 102.
Pathway 2: successful transmission by the european union in month 11 of 2009 was achieved by soil moisture and marine salinity satellites (Soil Moisture and Ocean Salinity, SMOS). The SMOS sensor is firstly carried with an L-band synthetic aperture radiometer, the working frequency of the SMOS sensor is 1.4GHz, and the transit time of the satellite is 6:00am (up-orbit) and 6:00pm (down-orbit) of the local satellite. Thus, in the embodiment of the present disclosure, the passive microwave remote sensing image corresponding to the region captured by SMOS may be obtained as the passive microwave remote sensing image described in the step 102.
In the embodiments of the present specification, the soil roughness data may be indexed by the root mean square height of the surface relief. The relation between the microwave index and the soil roughness measurement value is found to show strong correlation, so in the embodiment of the present specification, in order to improve the accuracy of soil moisture content measurement, besides the passive microwave remote sensing image, soil roughness data corresponding to the area is further selected as one of the bases of soil moisture content measurement.
In some embodiments of the present disclosure, soil roughness data corresponding to the above-described areas may be calculated using C-band H-polarization and V-polarization bright temperatures of an advanced microwave scanning radiometer AMSR-E. In order to simplify the calculation, in the embodiment of the present specification, the influence of frequency on the soil roughness may be ignored, and therefore, the soil roughness data calculated by C-band H-polarization and V-polarization bright temperatures may be used as the above-described soil roughness data.
In other embodiments of the present disclosure, soil roughness data corresponding to the above-described areas may also be obtained from soil information collection sites. The soil information acquisition site can measure the soil roughness in the district by various methods so as to determine and store the soil roughness data in the district.
In step 104, bright temperature data is extracted from the passive microwave remote sensing image.
In an embodiment of the present disclosure, the extracting the bright temperature data from the passive microwave remote sensing image may include: extracting brightness temperature data of a C wave band from a passive microwave remote sensing image shot by an AMSR 2; and/or extracting the L-band bright temperature data from the passive microwave remote sensing image shot by the SMOS.
Compared with bright temperature data of other frequencies in AMSR2, the passive microwave remote sensing image shot by AMSR2 has relatively longer wavelength corresponding to the C wave band (working frequency is 6.9 GHz), stronger penetrating power, least influence by vegetation and atmosphere and is more suitable for soil moisture inversion. Therefore, in the embodiment of the present specification, the bright temperature data of the C-band may be extracted from the passive microwave remote sensing image captured by the AMSR2 as one of the input features of the soil moisture content prediction model.
For passive microwave remote sensing images shot by SMOS, the response degree to soil moisture is higher due to the smaller wavelength of the L wave band. Therefore, in the embodiment of the present specification, the bright temperature data of the L-band can be extracted from the passive microwave remote sensing image photographed by SMOS as one of the input features of the soil moisture content prediction model.
In addition, in the embodiment of the present disclosure, in order to further improve the prediction accuracy of the soil moisture content, the influence of radio interference (RFI) is removed, and after the bright temperature data of the C band is extracted, according to the AMSR2 radio removal method, it may be further determined that the bright temperature data of the C band is subjected to stronger or moderate radio interference by using the pearson correlation coefficient value of the bright temperature data with the operating frequencies of 7.3GHz and 10.65GHz in the passive microwave remote sensing image captured by the AMSR2, and then removed.
For the L-band bright temperature data, after the L-band bright temperature data is extracted, a L3-level RFI quality control mechanism (for example, rfi_prob) in SMOS may be further used to determine, when an electromagnetic interference index rfi_prob of a certain pixel is greater than a preset threshold, for example, 30%, the pixel data may be considered to be severely affected by electromagnetic interference, and the pixel data may be removed.
In step 106, the bright temperature data and the soil roughness data are used as input features of a soil moisture content prediction model.
In the embodiment of the present disclosure, the bright temperature data of the C-band is extracted from the passive microwave remote sensing image captured by AMSR2 as the input feature of the soil moisture content prediction model, and/or the bright temperature data of the L-band is extracted from the passive microwave remote sensing image captured by SMOS as the input feature of the soil moisture content prediction model, taking into consideration that the surface moisture content and the surface temperature are strongly correlated, where the bright temperature data of the C-band in the passive microwave remote sensing image captured by AMSR2 and the bright temperature data of the L-band in the passive microwave remote sensing image captured by SMOS most accurately reflect the surface temperature, so that the bright temperature data is selected as the input feature of the soil moisture content prediction model to obtain higher prediction accuracy.
On the other hand, in the embodiment of the present disclosure, since only the bright temperature data of a partial wave band is extracted from the passive microwave remote sensing image as the input feature of the soil moisture content prediction model, but not the bright temperature data of all wave bands, the number of the input features of the soil moisture content prediction model can be greatly reduced, thereby reducing the complexity of the soil moisture content prediction model and improving the training and prediction efficiency of the soil moisture content prediction model.
In addition, in order to ensure the scientificity and feasibility of the input features of the soil moisture content prediction model, in some embodiments of the present specification, feature fusion can be further performed on the bright temperature data from the multi-source satellite and the soil roughness data.
In embodiments of the present description, the feature fusion described above may be specifically a normalization of the input features. Specifically, the plurality of input features of the soil moisture content prediction model described above may be normalized in one of two normalization manners.
First normalization: and (5) normalizing the linear function. I.e., linear transformation of the original data, mapping the original data into the [0,1] range. The normalization can be achieved specifically by the following formula (1):
Wherein X represents the original data; x is X max And X min Representing the maximum and minimum values of the original data.
Second normalization: standard deviation normalization, also called Z-Score normalization. The normalization can be achieved specifically by the following formula (2):
wherein μ represents the mean of the raw data; sigma represents the variance of the raw data; x represents the original data. In general, Z-Score normalization works better when distance is needed in the algorithm to measure similarity or when dimensionality reduction is performed using covariance analysis (PCA) techniques.
Determining the soil moisture content of the area according to the input characteristics based on the soil moisture content prediction model in step 108; the soil moisture content prediction model is a residual network model for determining the soil moisture content based on input features.
The aim of the embodiments of the present description is to quantitatively invert the soil moisture content by multi-feature input. Considering that inversion is a complex model, the inversion model combined with the neural network is simple at present, so that the problems of weak generalization capability, low precision and low reliability of the model are caused. Increasing the number of layers of the network causes problems of gradient disappearance and network degradation of the network model. Thus, a residual network model is introduced in embodiments of the present description to improve inversion capabilities of the network.
In an embodiment of the present specification, the residual network model includes: the input layer, at least two residual units that are connected through the cascade mode, and output layer. Assume that there are N residual units in total, where N is a natural number greater than 1. Wherein, the output of the input layer is connected with the input of the 1 st residual unit connected in a cascading mode; the output of the nth residual unit connected in a cascading manner is connected to the input of the (n+1) th residual unit, wherein N is more than or equal to 1; the output of the Nth residual unit connected in a cascade mode is connected to the input of the output layer.
Fig. 2 shows an internal structure of a residual unit according to one or more embodiments of the present disclosure. As shown in fig. 2, the residual unit may include: at least two hidden layers 202, a feature overlay layer 204, and an activation function layer 206 connected in a cascaded manner. The hidden layer 202 is configured to expand the features input by the input layer according to different dimensions, and extract high-dimensional features; the feature superimposing layer 204 is configured to superimpose the feature output by the input layer or the previous residual unit of the current residual unit with the feature output by the last hidden layer connected in cascade; the activation function of the activation function layer 206 is ReLU (x) =max (0, x).
It can be seen that in the embodiment of the present description, the residual network model is formed by connecting individual residual units as shown in fig. 2 to form a complete residual network. It should be noted that only two hidden layers are shown in fig. 2, and in practice, the number of hidden layers may be more than two, for example, 3 or even more. As can be seen from fig. 2, X is the input of this one residual unit and F (X) is the output that varies linearly through the concealment layer 202 and before activation. In the residual units represented in fig. 2, the input X of the one residual unit is added on an F (X) basis by the feature superimposing layer 204 before the activation function layer 206 is linearly changed and activated, and then is outputted after being activated by the activation function layer 206. Input X is added before the output value is activated, and this path is called a Shortcut (Shortcut) connection. The residual unit adds an identity mapping layer to the residual network model through shortcut connection, and mainly solves the gradient disappearance problem and the network degradation problem caused by the increase of the layer number of the traditional neural network.
It can be seen that the soil moisture content measurement method adopts the residual network model obtained by the supervised training mode as the soil moisture content prediction model, can fully utilize the characteristic that the residual network is added with the input value again before the output value is activated, solves the problems of weak generalization capability, low precision and low reliability when other neural networks are adopted for soil moisture content prediction, and can also effectively solve the problems of gradient disappearance, network degradation and the like generated when the number of layers of the neural network is increased. Experiments prove that the residual network model is very suitable for predicting the water content of soil and has the characteristics of strong generalization capability, high precision and strong reliability.
In addition, in the soil moisture content measuring method, a passive microwave remote sensing image is used as one of input features of a soil moisture content prediction model. Because the microwave remote sensing image is less affected by cloud, overcast and rainy, aerosol and the like, the accuracy of the soil water content obtained by predicting the soil water content according to the passive microwave remote sensing image is higher.
Furthermore, besides the passive microwave remote sensing image, the soil moisture content measuring method also utilizes the soil roughness as the input of a soil moisture content prediction model. The method fully utilizes the fact that the rough soil is fully considered to have great influence on the radiation brightness temperature detected by the passive microwave remote sensing satellite, so that the accuracy of predicting the water content of the soil can be further improved.
In other embodiments of the present disclosure, to further improve the prediction accuracy of the soil moisture content, the method may further include, in addition to the above C-band and L-band bright temperature data: extracting backscattering coefficients of H polarization and V polarization of a C wave band from the passive microwave remote sensing image shot by the AMSR2 to serve as one of input features of the soil water content prediction model; and/or extracting the backscattering coefficients of the L wave band H polarization and the V polarization from the passive microwave remote sensing image shot by the SMOS as one of the input features of the soil water content prediction model.
In these embodiments, the backscattering coefficients of the C-band H polarization and the V polarization in the passive microwave remote sensing image captured by the AMSR2 and/or the backscattering coefficients of the L-band H polarization and the V polarization in the passive microwave remote sensing image captured by the SMOS are further selected as one of the input features of the soil moisture content prediction model, and the capability of the convolutional neural network itself based on the residual structure to extract the parameter features is considered. By inputting the backscattering coefficients of the H polarization and the V polarization of the wave bands and combining the characteristic extraction capability of the network, the high-dimensional characteristics in the original data can be automatically extracted to invert the soil water content, so that the prediction precision of the soil water content is further improved.
Furthermore, in still other embodiments of the present description, in order to further increase the soilThe method may further include, in addition to the backscattering coefficients of H-polarization and V-polarization or based on the bright temperature data and the soil roughness data, predicting accuracy of the water content: and acquiring soil texture data corresponding to the region as one of input features of a soil water content prediction model. This is because the change of the soil quality plays a critical role in the water content of the soil, and the different soil qualities cause different water retention properties of the soil. Therefore, in the embodiment of the present specification, it is possible to further acquire the soil texture classification data corresponding to the above-mentioned region, that is, acquire the volume weight (g/cm) of the obtained soil 3 ) The soil quality data are used as one of the input characteristics of a soil water content prediction model. In the embodiment of the present specification, soil texture data corresponding to the above-described region may be obtained from the soil texture classification data. Those skilled in the art will appreciate that soil texture classification data is typically provided by the international agriculture organization (Food and Agriculture Organization, FAO).
As can be seen from the above description, the input features of the soil moisture content prediction model at least include: and the bright temperature data and the soil roughness data. In some embodiments of the present disclosure, the input features of the soil moisture content prediction model may further include: the backscattering coefficients of the C-band H polarization and the V polarization in the passive microwave remote sensing image shot by the AMSR2 and/or the backscattering coefficients of the L-band H polarization and the V polarization in the passive microwave remote sensing image shot by the SMOS. In other embodiments of the present disclosure, the input features of the soil moisture content prediction model may further include: soil texture data corresponding to the above-mentioned region.
Based on the above input features, some embodiments of the present description provide a soil moisture content measurement method. Fig. 3 shows a flow chart of a soil moisture content measuring method according to other embodiments of the present disclosure. As shown in fig. 3, the method may include:
In step 302, for an area, bright temperature data of a C-band and backscattering coefficients of H polarization and V polarization thereof in a passive microwave remote sensing image corresponding to the area captured by AMSR2, bright temperature data of an L-band and backscattering coefficients of H polarization and V polarization thereof in a passive microwave remote sensing image corresponding to the area captured by SMOS, soil roughness data corresponding to the area, and soil texture data corresponding to the area are obtained.
In step 304, normalization processing is performed on the bright temperature data of the C-band and the backscattering coefficients of the H-polarization and the V-polarization thereof in the passive microwave remote sensing image corresponding to the region captured by AMSR2, the bright temperature data of the L-band and the backscattering coefficients of the H-polarization and the V-polarization thereof in the passive microwave remote sensing image corresponding to the region captured by SMOS, the soil roughness data corresponding to the region, and the soil texture data corresponding to the region.
In an embodiment of the present specification, the normalization process described above may include: linear function normalization or standard deviation normalization.
In step 306, the data obtained after the normalization is used as the input feature of the soil moisture content prediction model.
Determining the soil moisture content of the area according to the input characteristics based on the soil moisture content prediction model in step 308; the soil moisture content prediction model is a residual network model for determining the soil moisture content based on input features.
It should be noted that, the specific implementation method of the steps 302 to 308 may refer to the steps 102 to 108, and the description is not repeated here.
It can be seen that the soil moisture content measurement method adopts the residual network model obtained by the supervised training mode as the soil moisture content prediction model, can fully utilize the characteristic that the residual network is added with the input value again before the output value is activated, solves the problems of weak generalization capability, low precision and low reliability when other neural networks are adopted for soil moisture content prediction, and can also effectively solve the problems of gradient disappearance, network degradation and the like generated when the number of layers of the neural network is increased. Experiments prove that the residual network model is very suitable for predicting the water content of soil and has the characteristics of strong generalization capability, high precision and strong reliability.
In addition, in the soil moisture content measurement method, the passive microwave remote sensing image is used as one of input features of the soil moisture content prediction model, and the microwave remote sensing image is less affected by cloud, overcast and rains, aerosol and the like, so that the accuracy of the soil moisture content obtained by predicting the soil moisture content according to the passive microwave remote sensing image is high.
In addition, the soil water content measuring method selects the bright temperature data collected by the radiometers carried by the two satellites and selects the bright temperature data of the wave band which can reflect the earth surface temperature condition most accurately as the input characteristic of the soil water content prediction model, so that the complexity of the soil water content prediction model can be reduced, and higher prediction precision can be obtained.
In addition, the backscattering coefficient of the wave band is further input into the soil water content prediction model, so that the high-dimensional characteristics in the original data can be automatically extracted by combining the characteristic extraction capability of the network to perform soil water content inversion, and the prediction accuracy of the soil water content can be further improved.
Furthermore, besides the passive microwave remote sensing image, the soil moisture content measuring method also utilizes the soil roughness and the soil texture data as the input of a soil moisture content prediction model, fully considers the great influence of the roughness of the soil on the radiation brightness temperature detected by the passive microwave remote sensing satellite, and utilizes the characteristics of the soil roughness and the correlation between the soil texture data and the soil moisture content, thereby further improving the precision of the soil moisture content prediction.
Those skilled in the art will understand that the construction of the network model structure is an experimental process from simple to complex, and the number of layers and parameters of the network will affect the inversion accuracy of the network. For example, as the number of layers of the network increases and the number of parameters increases, the feature extraction capability of the network and the learning capability of the training sample are greatly improved. However, if the number of network layers or parameters is too large, the network will not converge and the network will be too fit. Therefore, the residual network model used in the examples of the present specification is obtained by repeating experiments based on the degree of response to the soil moisture content of the image characteristics, the soil roughness, the soil characteristics, and the like used in the soil moisture content test method described above.
Specifically, in the embodiment of the present disclosure, through multiple simulation experiments, the structure of the residual network model is optimized according to the prediction result of the soil moisture content and the actual moisture content of the soil, so as to obtain the residual network model structure shown in the following table 1. Experiments prove that the residual error network model with the network structure can quickly converge, and meanwhile, the accurate prediction of the water content of the soil is ensured.
TABLE 1
As can be seen from table 1, the total of the residual network models used in the embodiments of the present specification total 101489 super parameters, wherein the residual network models include 10 residual units. The 10 residual single units each contain 2 to 3 concealment layers, and a total of 28 concealment layers. Specifically, the remaining 8 residual units each include 3 concealment layers, except that the second residual unit and the sixth residual unit include only 2 concealment layers.
Further, in the embodiment of the present specification, the activation function of the activation function layer described above may be set to ReLU (x) =max (0, x). The network training can be faster with the activation function described above. This is because the derivatives of the activation functions are better solved than those of Sigmoid, tanh, etc., thus making back propagation simpler and training faster. In addition, the activation function is a nonlinear function, and the network can be fitted with nonlinear mapping by adding the activation function to the neural network, so that the nonlinearity of the network can be increased by adopting the activation function. Furthermore, the activation function is an unsaturated activation function, so that the problem that the gradient disappears due to the fact that the reciprocal of the activation function is close to 0 when the value is too large or too small can be prevented. Finally, the activation function is smaller than 0 and has a value when the activation function is larger than 0, so that the overfitting of a residual network model can be reduced, and the grid has sparsity.
Fig. 4 is a graph showing the relationship between the predicted soil moisture content and the measured soil moisture content, using the soil moisture content prediction method described in the examples of the present specification. Wherein the Mean Square Error (MSE) between the predicted soil moisture content and the measured soil moisture content is about 0.002, and determining the coefficient R 2 About 0.645, and the prediction accuracy is about 86.7%. As can be seen from fig. 4, the soil moisture content prediction using the above-described soil moisture content prediction model can obtain higher prediction accuracy.
The following describes the training method of the residual network model in detail with reference to the accompanying drawings. Fig. 5 shows a method of training a residual network model according to one or more embodiments of the present disclosure.
In step 502, input features of a residual network model corresponding to a plurality of regions are obtained as a plurality of training samples of the residual network model.
In an embodiment of the present specification, the residual network model input features may include: bright temperature data and soil roughness data extracted from the passive microwave remote sensing image corresponding to the area. The bright temperature data may be bright temperature data of a C-band in the passive microwave remote sensing image corresponding to the region captured by AMSR2 and/or bright temperature data of an L-band in the passive microwave remote sensing image corresponding to the region captured by SMOS.
In an embodiment of the present disclosure, the input features may further include: soil texture data and/or backscattering coefficients of H polarization and V polarization corresponding to bright temperature data extracted from passive microwave remote sensing images corresponding to the region.
At step 504, the actual soil moisture content for each region of the plurality of training samples is obtained.
In the practice of this specification, the actual soil moisture content may be obtained from a soil information acquisition site. For example, the actual soil moisture content upstream of a black river basin may be obtained by 40 sites of an eight-treasure river basin upstream of the black river.
In step 506, the input features are input into a residual network model to be trained, so as to obtain predicted values of soil moisture content corresponding to the plurality of training samples output by the residual network model.
In step 508, a gap between the output of the residual network model and the actual soil moisture content of the region is determined using a predefined loss function, and parameters of the residual network model are back-propagated and adjusted according to the gap, thereby completing training of the soil moisture content prediction model.
In some examples of the present specification, the predefined Loss function may be a Regression Loss (Regression Loss) function, and specifically, an average absolute error (MAE) may be used as the Loss function, as shown in the following expression (3).
Wherein,an output of the residual network model corresponding to the ith training sample; y is i Representing the actual soil moisture content corresponding to the ith training sample; n represents the number of training samples.
The average absolute error is a loss function for regression model, and represents the sum of absolute values of differences between the target variable and the predicted variable. Thus, it scales the average size of the error in a set of predictions, regardless of the direction of the error, and the loss range is also 0 to ≡infinity. Therefore, using the absolute error as a loss function can improve the robustness of the data.
In practical application, for the training of the residual network model, 3000 iteration steps can be set, each iteration step is monitored, and finally the super parameter with the highest precision is stored. Experiments prove that the residual error network model with the network structure is adopted to predict the soil water content, and the prediction accuracy can reach 86.7%.
Based on the above soil moisture content measuring method, one or more embodiments of the present disclosure further provide a soil moisture content measuring device, an internal structure of which is shown in fig. 6, and mainly includes:
the feature acquisition module 602 is configured to acquire, for an area, a passive microwave remote sensing image and soil roughness data corresponding to the area;
The feature extraction module 604 is configured to extract bright temperature data from the passive microwave remote sensing image, and use the bright temperature data and the soil roughness data as input features of a soil moisture content prediction model; and
a prediction module 606, configured to determine a soil moisture content of the area according to the input feature based on the soil moisture content prediction model; the soil moisture content prediction model is a residual network model for determining the soil moisture content based on input features.
It should be noted that, the specific implementation method of each module of the soil moisture content measuring device may refer to each foregoing embodiment, and the description is not repeated here.
It should be noted that the methods of one or more embodiments of the present description may be performed by a single device, such as a computer or server. The method of the embodiment can also be applied to a distributed scene, and is completed by mutually matching a plurality of devices. In the case of such a distributed scenario, one of the devices may perform only one or more steps of the methods of one or more embodiments of the present description, which interact with each other to accomplish the methods described above.
The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
For convenience of description, the above devices are described as being functionally divided into various modules, respectively. Of course, the functions of each module may be implemented in one or more pieces of software and/or hardware when implementing one or more embodiments of the present description.
The device of the foregoing embodiment is configured to implement the corresponding method in the foregoing embodiment, and has the beneficial effects of the corresponding method embodiment, which is not described herein.
Fig. 7 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present disclosure, where the device may include: processor 710, memory 720, input/output interface 730, communication interface 740, and bus 750. Wherein processor 710, memory 720, input/output interface 730, and communication interface 740 implement a communication connection among each other within the device via bus 750.
Processor 710 may be implemented in a general purpose CPU (Central Processing Unit ), microprocessor, application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc. for executing associated routines to implement the soil moisture content measurement methods provided in the embodiments of the present disclosure.
The Memory 720 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory ), static storage device, dynamic storage device, or the like. Memory 720 may store an operating system and other application programs, and when the soil moisture content measuring method provided by the embodiments of the present specification is implemented in software or firmware, the relevant program code is stored in memory 720 and invoked by processor 710 for execution.
The input/output interface 730 is used to connect with an input/output module to realize information input and output. The input/output module may be configured as a component in a device (not shown) or may be external to the device to provide corresponding functionality. Wherein the input devices may include a keyboard, mouse, touch screen, microphone, various types of sensors, etc., and the output devices may include a display, speaker, vibrator, indicator lights, etc.
The communication interface 740 is used to connect with a communication module (not shown) to enable communication interactions between the device and other devices. The communication module may implement communication through a wired manner (such as USB, network cable, etc.), or may implement communication through a wireless manner (such as mobile network, WIFI, bluetooth, etc.).
Bus 750 includes a path to transfer information between elements of the device (e.g., processor 710, memory 720, input/output interface 730, and communication interface 740).
It should be noted that although the above-described device only shows processor 710, memory 720, input/output interface 730, communication interface 740, and bus 750, in particular implementations, the device may include other components necessary to achieve proper operation. Furthermore, it will be understood by those skilled in the art that the above-described apparatus may include only the components necessary to implement the embodiments of the present description, and not all the components shown in the drawings.
The computer readable media of the present embodiments, including both permanent and non-permanent, removable and non-removable media, may be used to implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device.
Those of ordinary skill in the art will appreciate that: the discussion of any of the embodiments above is merely exemplary and is not intended to suggest that the scope of the disclosure, including the claims, is limited to these examples; combinations of features of the above embodiments or in different embodiments are also possible within the spirit of the present disclosure, steps may be implemented in any order, and there are many other variations of the different aspects of one or more embodiments described above which are not provided in detail for the sake of brevity.
Additionally, well-known power/ground connections to Integrated Circuit (IC) chips and other components may or may not be shown within the provided figures, in order to simplify the illustration and discussion, and so as not to obscure one or more embodiments of the present description. Furthermore, the apparatus may be shown in block diagram form in order to avoid obscuring the one or more embodiments of the present description, and also in view of the fact that specifics with respect to implementation of such block diagram apparatus are highly dependent upon the platform within which the one or more embodiments of the present description are to be implemented (i.e., such specifics should be well within purview of one skilled in the art). Where specific details (e.g., circuits) are set forth in order to describe example embodiments of the disclosure, it should be apparent to one skilled in the art that one or more embodiments of the disclosure can be practiced without, or with variation of, these specific details. Accordingly, the description is to be regarded as illustrative in nature and not as restrictive.
While the present disclosure has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of those embodiments will be apparent to those skilled in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.
The present disclosure is intended to embrace all such alternatives, modifications and variances which fall within the broad scope of the appended claims. Any omissions, modifications, equivalents, improvements, and the like, which are within the spirit and principles of the one or more embodiments of the disclosure, are therefore intended to be included within the scope of the disclosure.
Claims (4)
1. A method for measuring soil moisture content, comprising:
for an area, acquiring a passive microwave remote sensing image and soil roughness data corresponding to the area;
extracting brightness temperature data from the passive microwave remote sensing image;
taking the bright temperature data and the soil roughness data as input features of a soil water content prediction model;
determining the soil moisture content of the area according to the input characteristics based on the soil moisture content prediction model; the soil water content prediction model is a residual network model for determining the soil water content based on input features;
The obtaining the passive microwave remote sensing image corresponding to the region comprises the following steps: acquiring a passive microwave remote sensing image shot by an advanced microwave scanning radiometer 2 satellite AMSR2 and/or a passive microwave remote sensing image shot by a soil moisture and ocean salinity satellite SMOS corresponding to the region;
wherein, extracting bright temperature data from the passive microwave remote sensing image comprises: extracting C-band bright temperature data from the passive microwave remote sensing image corresponding to the region shot by the AMSR 2; and/or extracting brightness temperature data of an L wave band from the passive microwave remote sensing image corresponding to the region shot by the SMOS;
wherein the method further comprises: eliminating radio interference of C wave Duan Liangwen data by using bright temperature data of other wave bands except the C wave band in the passive microwave remote sensing image shot by the AMSR 2; and/or removing the pixel data when the electromagnetic interference index of a certain pixel is greater than a preset threshold value by utilizing an L3-level electromagnetic interference quality control mechanism in the SMOS;
extracting backscattering coefficients of C wave band H polarization and V polarization from the passive microwave remote sensing image shot by the AMSR2 as one of input features of the soil water content prediction model; and/or extracting backscattering coefficients of L wave band H polarization and V polarization from the passive microwave remote sensing image shot by the SMOS as one of input features of the soil water content prediction model;
Acquiring soil texture data corresponding to the region as one of input features of the soil water content prediction model;
wherein the residual network model comprises: the input layer, N residual units connected in a cascading manner and the output layer; wherein N is a natural number greater than 1;
the output of the input layer is connected to the input of the 1 st residual unit; the output of the nth residual unit is connected to the input of the (n+1) th residual unit, wherein N < N is 1-N; the output of the nth residual unit is connected to the input of the output layer; wherein,
the residual unit includes: at least two hidden layers, a characteristic superposition layer and an activation function layer which are connected in a cascading manner; the hidden layer is used for expanding the characteristics input by the input layer according to different dimensions and extracting high-dimensional characteristics; the characteristic overlapping layer is used for overlapping the output characteristics of the input layer or the previous residual error unit with the output characteristics of the last hidden layer connected in a cascading mode; the activation function of the activation function layer is ReLU (x) =max (0, x).
2. A soil moisture content measuring device comprising:
the characteristic acquisition module is used for acquiring a passive microwave remote sensing image and soil roughness data corresponding to an area; the obtaining the passive microwave remote sensing image corresponding to the region comprises the following steps: acquiring a passive microwave remote sensing image shot by an advanced microwave scanning radiometer 2 satellite AMSR2 and/or a passive microwave remote sensing image shot by a soil moisture and ocean salinity satellite SMOS corresponding to the region;
The characteristic extraction module is used for extracting bright temperature data from the passive microwave remote sensing image, and taking the bright temperature data and the soil roughness data as input characteristics of a soil water content prediction model; wherein, extracting bright temperature data from the passive microwave remote sensing image comprises: extracting C-band bright temperature data from the passive microwave remote sensing image corresponding to the region shot by the AMSR 2; and/or extracting brightness temperature data of an L wave band from the passive microwave remote sensing image corresponding to the region shot by the SMOS;
the prediction module is used for determining the soil moisture content of the area according to the input characteristics based on the soil moisture content prediction model; the soil water content prediction model is a residual network model for determining the soil water content based on input features; the residual network model comprises: the input layer, N residual units connected in a cascading manner and the output layer; wherein N is a natural number greater than 1; the output of the input layer is connected to the input of the 1 st residual unit; the output of the nth residual unit is connected to the input of the (n+1) th residual unit, wherein N < N is 1-N; the output of the nth residual unit is connected to the input of the output layer; wherein the residual unit comprises: at least two hidden layers, a characteristic superposition layer and an activation function layer which are connected in a cascading manner; the hidden layer is used for expanding the characteristics input by the input layer according to different dimensions and extracting high-dimensional characteristics; the characteristic overlapping layer is used for overlapping the output characteristics of the input layer or the previous residual error unit with the output characteristics of the last hidden layer connected in a cascading mode; the activation function of the activation function layer is ReLU (x) =max (0, x); wherein,
The soil moisture content measuring device further includes:
the module is used for eliminating radio interference of C wave Duan Liangwen data by using bright temperature data of other wave bands except the C wave band in the passive microwave remote sensing image shot by the AMSR 2; and/or a module for removing the pixel data when the electromagnetic interference index of a certain pixel is greater than a preset threshold value by utilizing an L3-level electromagnetic interference quality control mechanism in the SMOS;
a module for extracting backscattering coefficients of C wave band H polarization and V polarization from the passive microwave remote sensing image shot by the AMSR2 as one of input features of the soil water content prediction model; and/or extracting backscattering coefficients of L-band H polarization and V polarization from the passive microwave remote sensing image photographed by the SMOS as one of input features of the soil moisture content prediction model; and
and means for acquiring soil texture data corresponding to the region as one of the input features of the soil moisture content prediction model.
3. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the soil moisture content measurement method of claim 1 when the program is executed by the processor.
4. A non-transitory computer readable storage medium, wherein the non-transitory computer readable storage medium stores computer instructions for causing the computer to perform the soil moisture content measurement method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010911955.8A CN112161998B (en) | 2020-09-02 | 2020-09-02 | Soil water content measuring method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010911955.8A CN112161998B (en) | 2020-09-02 | 2020-09-02 | Soil water content measuring method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112161998A CN112161998A (en) | 2021-01-01 |
| CN112161998B true CN112161998B (en) | 2023-12-05 |
Family
ID=73858627
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010911955.8A Active CN112161998B (en) | 2020-09-02 | 2020-09-02 | Soil water content measuring method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112161998B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105606631A (en) * | 2016-02-01 | 2016-05-25 | 中国科学院遥感与数字地球研究所 | Method for jointly inversing soil moisture through salinity satellite dual-waveband brightness temperature data |
| CN106569210A (en) * | 2016-10-18 | 2017-04-19 | 长安大学 | Soil moisture inversion method and soil moisture inversion device based on remote sensing |
| CN106779062A (en) * | 2016-11-23 | 2017-05-31 | 苏州科技大学 | A kind of multi-layer perception (MLP) artificial neural network based on residual error network |
| CN108268735A (en) * | 2018-01-29 | 2018-07-10 | 浙江大学 | Soil moisture NO emissions reduction method based on multi-source remote sensing satellite fused data |
| CN108982548A (en) * | 2018-07-20 | 2018-12-11 | 浙江大学 | A kind of soil moisture inversion method based on passive microwave remote sensing data |
| CN109884279A (en) * | 2019-04-04 | 2019-06-14 | 内蒙古工业大学 | The inversion method and inverting device of soil moisture content |
| CN110221301A (en) * | 2019-07-24 | 2019-09-10 | 中国科学院遥感与数字地球研究所 | Soil Moisture Inversion method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015195988A1 (en) * | 2014-06-18 | 2015-12-23 | Texas Tech University System | Portable apparatus for soil chemical characterization |
-
2020
- 2020-09-02 CN CN202010911955.8A patent/CN112161998B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105606631A (en) * | 2016-02-01 | 2016-05-25 | 中国科学院遥感与数字地球研究所 | Method for jointly inversing soil moisture through salinity satellite dual-waveband brightness temperature data |
| CN106569210A (en) * | 2016-10-18 | 2017-04-19 | 长安大学 | Soil moisture inversion method and soil moisture inversion device based on remote sensing |
| CN106779062A (en) * | 2016-11-23 | 2017-05-31 | 苏州科技大学 | A kind of multi-layer perception (MLP) artificial neural network based on residual error network |
| CN108268735A (en) * | 2018-01-29 | 2018-07-10 | 浙江大学 | Soil moisture NO emissions reduction method based on multi-source remote sensing satellite fused data |
| CN108982548A (en) * | 2018-07-20 | 2018-12-11 | 浙江大学 | A kind of soil moisture inversion method based on passive microwave remote sensing data |
| CN109884279A (en) * | 2019-04-04 | 2019-06-14 | 内蒙古工业大学 | The inversion method and inverting device of soil moisture content |
| CN110221301A (en) * | 2019-07-24 | 2019-09-10 | 中国科学院遥感与数字地球研究所 | Soil Moisture Inversion method and device |
Non-Patent Citations (5)
| Title |
|---|
| ESTIMATE OF SOIL MOISTURE USING REFINED MICROWAVE VEGETATION INDEX BASED ON AMSR-E;Wang Shu;2013 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS);3758-3761 * |
| 基于卷积神经网络和AMSR2微波遥感的土壤水分反演研究;谭建灿 等;高技术通讯(第05期);21-30 * |
| 基于多频被动微波遥感的土壤水分反演-以黑河上游为例;王舒;遥感技术与应用;第35卷(第6期);1414-1425 * |
| 基于被动微波遥感的地表粗糙度及土壤水分反演研究;王舒;测绘学报;第50卷(第10期);1419 * |
| 被动微波反演土壤水分的L波段新发展及未来展望;赵天杰;地理科学进展(第02期);28-43 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112161998A (en) | 2021-01-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jiang et al. | Quantitative assessment of soil salinity using multi-source remote sensing data based on the support vector machine and artificial neural network | |
| Kaba et al. | Estimation of daily global solar radiation using deep learning model | |
| Singh et al. | Deep learning and data fusion to estimate surface soil moisture from multi-sensor satellite images | |
| Pasolli et al. | Polarimetric RADARSAT-2 imagery for soil moisture retrieval in alpine areas | |
| Senthilnath et al. | Integration of speckle de-noising and image segmentation using Synthetic Aperture Radar image for flood extent extraction | |
| Pasolli et al. | Estimation of soil moisture in an alpine catchment with RADARSAT2 images | |
| Jia et al. | Rice biomass retrieval from multitemporal ground-based scatterometer data and RADARSAT-2 images using neural networks | |
| Rocha de Souza Pereira et al. | Mapping of mangrove forests on the southern coast of São Paulo, Brazil, using synthetic aperture radar data from ALOS/PALSAR | |
| Zarei et al. | A comparison of machine learning models for soil salinity estimation using multi-spectral earth observation data | |
| Emrullah et al. | Machine learning based regression model for prediction of soil surface humidity over moderately vegetated fields | |
| Chen et al. | ARU-net: Reduction of atmospheric phase screen in SAR interferometry using attention-based deep residual U-net | |
| Wan et al. | Modeling wetland aboveground biomass in the Poyang Lake National Nature Reserve using machine learning algorithms and Landsat-8 imagery | |
| Novelli et al. | Improving the ANN classification accuracy of landsat data through spectral indices and linear transformations (PCA and TCT) aimed at LU/LC monitoring of a river basin | |
| Vafadar et al. | Hyperspectral anomaly detection using outlier removal from collaborative representation | |
| Camacho Velasco et al. | A comparative study of target detection algorithms in hyperspectral imagery applied to agricultural crops in Colombia | |
| Barkataki et al. | Estimation of soil moisture from GPR data using artificial neural networks | |
| Byun et al. | Deep learning-based rainfall prediction using cloud image analysis | |
| CN116879192B (en) | Water bloom prediction method, device, equipment and medium based on satellite remote sensing data | |
| Geng et al. | Wishart distance‐based joint collaborative representation for polarimetric SAR image classification | |
| CN112161998B (en) | Soil water content measuring method and device, electronic equipment and storage medium | |
| Stamenković et al. | Joining a discrete radiative transfer model and a kernel retrieval algorithm for soil moisture estimation from SAR data | |
| Kononov et al. | Model-associated forest parameter retrieval using VHF SAR data at the individual tree level | |
| CN102663420A (en) | Hyperspectral image classification method based on wavelet packet transformation and grey prediction model | |
| Géant et al. | Mapping small inland wetlands in the South-Kivu province by integrating optical and SAR data with statistical models for accurate distribution assessment | |
| Hu et al. | RFI localization in synthetic aperture interferometric radiometers based on sparse Bayesian inference |
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 |