Method and device for detecting spectral reflectivity of lake water body
阅读说明：本技术 湖泊水体光谱反射率检验方法及装置 (Method and device for detecting spectral reflectivity of lake water body ) 是由 陶醉 谢富泰 周翔 吕婷婷 王锦 于 2020-07-15 设计创作，主要内容包括：本发明实施例提供一种湖泊水体光谱反射率检验方法及装置,所述方法包括：利用预设的水体光谱反射率预测模型对实测水体光谱反射率进行筛选；所述水体光谱反射率预测模型是基于水体参量样本数据、环境参量样本数据和对应的水体样本光谱反射率进行训练后得到的；所述实测水体光谱反射率是通过测量设备对待测水体进行抽样检测得到的光谱反射率；根据筛选后的实测水体光谱反射率对卫星反演算法进行检验和修正,使用修正后的卫星反演算法获取待测水体光谱反射率。通过筛选后的实测水体光谱反射率对卫星反演获取的水体光谱反射率进行验证并修正,解决了现有的通过卫星获取的水体光谱反射率可信度低的缺陷,提高了待测水体的光谱反射率可靠性。(The embodiment of the invention provides a method and a device for detecting the spectral reflectivity of a lake water body, wherein the method comprises the following steps: screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment; and checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the water body spectral reflectivity to be measured by using the corrected satellite inversion algorithm. The water body spectral reflectivity obtained by satellite inversion is verified and corrected through the screened actually-measured water body spectral reflectivity, the defect that the existing water body spectral reflectivity obtained through a satellite is low in reliability is overcome, and the reliability of the spectral reflectivity of the water body to be detected is improved.)
1. The method for detecting the spectral reflectivity of the lake water body is characterized by comprising the following steps:
screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment;
and checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the water body spectral reflectivity to be measured by using the corrected satellite inversion algorithm.
2. The method for testing the spectral reflectivity of the lake water body according to claim 1, wherein the step of screening the actually measured spectral reflectivity of the water body by using the preset prediction model of the spectral reflectivity of the water body specifically comprises the following steps:
inputting the target related variable into the preset water body spectral reflectivity prediction model, and outputting the predicted water body spectral reflectivity corresponding to the target related variable;
and screening the actually measured water body spectral reflectivity according to the predicted water body spectral reflectivity.
3. The method for testing the spectral reflectivity of lake water according to claim 2, wherein before inputting the target related variable into the preset water spectral reflectivity prediction model, the method further comprises:
acquiring a correlation value between each water body parameter sample data and the corresponding water body sample spectral reflectivity and a correlation value between each environment sample parameter data and the corresponding water body sample spectral reflectivity based on the Pearson correlation coefficient;
acquiring sample related variables according to the correlation value between the water body sample parameter data and the corresponding water body sample spectral reflectivity and the correlation value between the environment sample parameter data and the corresponding water body sample spectral reflectivity;
and screening out target related variables from the target water body parameter data and the target environment parameter data according to the types of the sample related variables.
4. The method for testing the spectral reflectance of the lake water body according to claim 3, wherein the obtaining of the sample-related variable specifically comprises:
and screening out water body sample parameter data and environment sample parameter data which have correlation values between the water body sample parameter data and the corresponding water body sample spectral reflectivity larger than a preset correlation value threshold value from the water body sample parameter data and the environment sample parameter data as sample correlation variables.
5. The method for testing the spectral reflectivity of the lake water body according to claim 4, wherein the specific steps for obtaining the prediction model of the spectral reflectivity of the water body are as follows:
acquiring the sample related variable and the spectral reflectivity of the water body sample corresponding to the sample related variable;
and generating model parameters of the water body spectral reflectivity prediction model based on the sample related variables and the water body sample spectral reflectivity corresponding to the sample related variables, and obtaining the water body spectral reflectivity prediction model.
6. The method for testing the spectral reflectance of lake water according to claim 2, wherein the step of screening the measured spectral reflectance of water according to the predicted spectral reflectance of water specifically comprises:
calculating a first error value between the predicted water spectral reflectivity and the measured water spectral reflectivity;
and deleting the data of which the first error value exceeds a first preset threshold value from the actually measured water body spectral reflectivity.
7. Lake water spectral reflectivity verifying attachment, its characterized in that includes:
the screening module is used for screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment;
and the inspection module is used for inspecting and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity and acquiring the spectral reflectivity of the water body to be measured by using the corrected satellite inversion algorithm.
8. 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 steps of the method for testing the spectral reflectance of a body of water in a lake according to any one of claims 1 to 6 when executing the program.
9. A non-transitory computer readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, implements the steps of the method for testing the spectral reflectivity of a body of water in a lake according to any one of claims 1 to 6.
Technical Field
The invention relates to the technical field of remote sensing detection, in particular to a method and a device for detecting the spectral reflectivity of a lake water body.
Background
The spectral reflectivity is the basis of most remote sensing products, and has important significance in the aspects of inversion, verification and application of the remote sensing products. The precision of the water body remote sensing information product not only depends on the precision of the corresponding model, but also is limited by the accuracy of the spectral reflectivity to a greater extent. Therefore, authenticity check is carried out on the spectral reflectance of the water body, the precision of the remote sensing model can be effectively improved, and the accuracy and the reliability of the water body remote sensing information product are improved, so that the remote sensing product is popularized and applied in a wider range, and the application barrier of the remote sensing product is broken.
The authenticity inspection method of the remote sensing product comprises direct inspection, indirect inspection and cross inspection, and the authenticity inspection of the spectral reflectivity of the water body usually adopts a direct inspection method, namely, the reflectivity of a satellite image is directly compared with the actually measured reflectivity of the water body. However, due to the change of the mobility of the water body and environmental factors, the spectral reflectance of the water body changes greatly with time, the spectral curves of the water body reflectance separated by 3 hours may have large differences, and the intensity of the change is uncertain at present.
The existing method for evaluating the measured value of the spectral reflectance of the water body does not adjust specific water body to be detected, does not solve the problem of variation of the time scale of the true value of the measured value of the water body reflectance, increases uncertainty in the water body reflectance authenticity detection, and cannot accurately reflect the precision of satellite reflectance, so that the reliability of experimental data of the water body reflectance authenticity detection is low.
Disclosure of Invention
The embodiment of the invention provides a method and a device for testing the spectral reflectivity of a lake water body, which are used for solving the defect of low reliability of the spectral reflectivity of a water body to be tested, which is obtained by a satellite in the prior art, and realizing the function of improving the reliability of the spectral reflectivity of the water body to be tested.
The embodiment of the invention provides a method for detecting the spectral reflectivity of a lake water body, which comprises the following steps:
screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment;
and checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the water body spectral reflectivity to be measured by using the corrected satellite inversion algorithm.
According to the method for testing the spectral reflectivity of the lake water body, the step of screening the actually measured spectral reflectivity of the water body by using the preset water body spectral reflectivity prediction model specifically comprises the following steps:
inputting the target related variable into the preset water body spectral reflectivity prediction model, and outputting the predicted water body spectral reflectivity corresponding to the target related variable;
and screening the actually measured water body spectral reflectivity according to the predicted water body spectral reflectivity.
According to the method for testing the spectral reflectivity of the lake water body, before the target related variable is input into the preset water body spectral reflectivity prediction model, the method further comprises the following steps:
acquiring a correlation value between each water body parameter sample data and the corresponding water body sample spectral reflectivity and a correlation value between each environment sample parameter data and the corresponding water body sample spectral reflectivity based on the Pearson correlation coefficient;
acquiring sample related variables according to the correlation value between the water body sample parameter data and the corresponding water body sample spectral reflectivity and the correlation value between the environment sample parameter data and the corresponding water body sample spectral reflectivity;
and screening out target related variables from the target water body parameter data and the target environment parameter data according to the types of the sample related variables.
According to the method for testing the spectral reflectivity of the lake water body, the obtaining of the relevant variables of the sample specifically comprises the following steps:
and screening out water body sample parameter data and environment sample parameter data which have correlation values between the water body sample parameter data and the corresponding water body sample spectral reflectivity larger than a preset correlation value threshold value from the water body sample parameter data and the environment sample parameter data as sample correlation variables.
According to the method for detecting the spectral reflectivity of the lake water body, the concrete steps of obtaining the water body spectral reflectivity prediction model are as follows:
acquiring the sample related variable and the spectral reflectivity of the water body sample corresponding to the sample related variable;
and generating model parameters of the water body spectral reflectivity prediction model based on the sample related variables and the water body sample spectral reflectivity corresponding to the sample related variables, and obtaining the water body spectral reflectivity prediction model.
According to the method for testing the spectral reflectivity of the lake water body, the screening of the actually measured spectral reflectivity of the water body according to the predicted spectral reflectivity of the water body specifically comprises the following steps:
calculating a first error value between the predicted water spectral reflectivity and the measured water spectral reflectivity;
and deleting the data of which the first error value exceeds a first preset threshold value from the actually measured water body spectral reflectivity.
The embodiment of the invention also provides a device for detecting the spectral reflectivity of the lake water body, which comprises:
the screening module is used for screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment;
and the inspection module is used for inspecting and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity and acquiring the spectral reflectivity of the water body to be measured by using the corrected satellite inversion algorithm.
The embodiment of the invention also provides electronic equipment, which comprises a memory, a processor and a computer program which is stored on the memory and can be run on the processor, wherein when the processor executes the program, the steps of the method for testing the spectral reflectivity of the lake water body are realized.
The embodiment of the invention also provides a non-transitory computer readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the steps of the method for testing the spectral reflectivity of the lake water body.
According to the method and the device for testing the spectral reflectivity of the lake water body, the spectral reflectivity of the water body to be tested, which is obtained through satellite inversion, is tested through the screened actually-measured water body spectral reflectivity, and the satellite inversion model is improved according to the test result, so that the accuracy of the existing method for testing and verifying the authenticity of the satellite water body spectral reflectivity is improved, and the reliability of the spectral reflectivity of the water body to be tested, which is obtained through satellite inversion, is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for testing the spectral reflectivity of a lake water body according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a device for testing the spectral reflectance of a lake water body according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating a test result of a single-band BP neural network model according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating test results of a single-band GRNN neural network model according to an embodiment of the present invention;
FIG. 6 is a diagram illustrating test results of a full-band GRNN neural network model according to an embodiment of the present invention;
fig. 7 is a schematic diagram illustrating a test result of the full-band GRNN neural network model according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of a method for testing spectral reflectance of a lake water body according to an embodiment of the present invention, and as shown in fig. 1, the flow chart may specifically include:
step 101, screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; and the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment.
Specifically, the actually measured water spectral reflectance is the spectral reflectance obtained by sampling and detecting the water to be measured through the measuring equipment, so that the problems of inaccurate measured value and low reliability may exist in the actually measured water spectral reflectance. Screening the actually measured water spectral reflectivity through a preset water spectral reflectivity prediction model, namely performing error judgment on the actually measured water spectral reflectivity and the predicted water spectral reflectivity generated in the preset water spectral reflectivity prediction model, judging that the actually measured water spectral reflectivity meets the requirement if the error is in a preset range, and correcting the actually measured water spectral reflectivity if the error exceeds the preset range.
For example, the actually measured water spectral reflectance is obtained through the measuring equipment, but the actually measured water spectral reflectance may have the problem that part of measured values are inaccurate, and the actually measured water spectral reflectance is screened through the preset water spectral reflectance prediction model, so that the actually measured water spectral reflectance with the error of the predicted water spectral reflectance exceeding the preset threshold value generated in the prediction model can be screened out. The preset water body spectral reflectivity prediction model is obtained after training based on the water body parameter sample data, the environment parameter sample data and the corresponding water body sample spectral reflectivity, and a BP neural network model or a GRNN neural network model and the like can be used.
And 102, checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the spectral reflectivity of the water body to be measured by using the corrected satellite inversion algorithm.
Specifically, the spectral reflectivity of the water body to be detected acquired by the satellite is detected through the screened actual measurement water body spectral reflectivity, whether the error between the screened actual measurement water body spectral reflectivity and the spectral reflectivity of the water body to be detected acquired through satellite inversion is larger than a preset threshold value or not is judged, if the error is within the preset threshold value range, the spectral reflectivity of the water body to be detected acquired by the satellite is high in precision, and the spectral reflectivity can be used as a basis for subsequent quantitative remote sensing inversion and remote sensing parameter extraction; if the error exceeds the preset threshold range, the accuracy of the spectral reflectivity of the water body to be detected acquired by the satellite is low, the inversion model of the spectral reflectivity of the satellite needs to be corrected, and under the condition of uncorrecting, the error needs to be carefully considered when the image of the satellite is used for quantitative remote sensing inversion and water body parameter extraction.
For example, whether the error between the spectral reflectance of the actual measurement water after screening and the spectral reflectance of the water to be measured collected by the satellite is greater than a preset threshold value is judged, wherein the preset threshold value comprises: r2Threshold, RMSE threshold and AE threshold, and the actual measurement is determined by these three thresholdsWhether the error between the spectral reflectivity of the water body and the spectral reflectivity of the water body to be detected acquired by the satellite meets the requirement or not. Wherein R is2The threshold is a determinant coefficient threshold, the RMSE threshold is a root mean square error threshold, and the AE threshold is a mean absolute relative error threshold.
And if the error value between the filtered actually-measured water body spectral reflectivity and the spectral reflectivity of the water body to be measured is larger than a preset threshold value, the accuracy of the spectral reflectivity of the water body to be measured is lower, and then the filtered actually-measured water body spectral reflectivity is used as truth value data to perform modeling analysis on the water body reflectivity of the remote sensing satellite, so that the inversion precision of the water body reflectivity is improved.
For example, if the error value between the spectral reflectance of the screened actually measured water body and the spectral reflectance of the water body to be measured is 0.5, and the error value between the spectral reflectance of the preset actually measured water body and the spectral reflectance of the water body to be measured is 0.2, it is determined that the error of the spectral reflectance of the water body to be measured is large, and the error needs to be corrected. And modeling and analyzing the water body reflectivity of the remote sensing satellite by using the screened actually-measured water body spectral reflectivity as true value data, and acquiring the spectral reflectivity of the water body to be detected by using a corrected satellite inversion algorithm, so that the inversion accuracy of the water body spectral reflectivity is improved.
Optionally, on the basis of the foregoing embodiments, the screening of the actually measured water spectral reflectance by using the preset water spectral reflectance prediction model specifically includes:
inputting the target related variable into the preset water body spectral reflectivity prediction model, and outputting the predicted water body spectral reflectivity corresponding to the target related variable;
and screening the actually measured water body spectral reflectivity according to the predicted water body spectral reflectivity.
Specifically, the target related variable is an input value of a preset water spectral reflectance prediction model, the output value is a predicted water spectral reflectance corresponding to the target related variable, the type of the target related variable can be set according to requirements, and the more obvious the connection between the type of the target related variable and the predicted water spectral reflectance is, the better the effect is.
For example, for a specific wave band of a satellite to be detected, a water body and an environment variable which have strong correlation with the reflectivity of the wave band are used as target correlation variables, the reflectivity of the wave band is used as a dependent variable, BP neural network and GRNN neural network modeling is carried out, and R is utilized2And RMSE and AE were evaluated for accuracy.
Taking MODIS observation band 438-448nm as an example, the variables with higher correlation with the reflectivity of the equivalent band 440-450nm include 8 variables of water temperature, specific conductance, salinity, chlorophyll concentration, FDOM, average wind direction, air temperature and humidity. And (3) modeling the BP neural network and the GRNN neural network by using the 8 variables as target related variables and using the spectral reflectance of the water body with the equivalent wave band of 440-450nm as a dependent variable.
And obtaining available data 366 groups in total through data preprocessing and one-to-one correspondence between the water body spectral reflectivity data and the water body and environment parameter data observation time. And randomly selecting 90% of the data, namely 329 groups of data, as training data, and respectively modeling by using BP and GRNN neural networks. The remaining 37 groups of data are used as a test set, the test data are used for respectively verifying two neural network models, and R is used2RMSE and AE evaluate model accuracy. Fig. 4 and 5 show the test results of the BP neural network and GRNN neural network full-band models, respectively (11 test samples were randomly selected for display), and the precision evaluation results are shown in table 1.
TABLE 1 test results of single band neural network model
For another example, the correlation between the reflectivity of part of equivalent wave bands and water and environmental variables is easy to know, and 10 variables of water temperature, specific conductance, salinity, chlorophyll concentration, blue algae protein, FDOM, suspended matter concentration, average wind direction, air temperature and humidity still have higher sensitivity in the whole wave band. Although the correlation coefficient between chlorophyll concentration and suspended matter concentration is low in the partial band, it is reasonable to select these 10 variables as the independent variables in this step. We use these 10 variables as target dependent variables.
And (3) modeling by using the BP neural network and the GRNN neural network by using the 10 variables as target related variables and using the reflectivity of 50 equivalent wave bands between 400-900nm as dependent variables. And obtaining a usable data 366 group through data preprocessing and the corresponding of the observation time of the water body reflectivity data and the independent variable observation time. Randomly selecting 90% of data, i.e. 329 groups of data as training set, and the rest 37 groups of data as test set, testing the model, and using R2And RMSE and AE were evaluated for accuracy. Fig. 6 and 7 show the test results of the BP neural network and GRNN neural network full-band models, respectively (one randomly selected from 37 test spectral curves is shown). The results of the precision evaluation are shown in Table 2.
TABLE 2 test results of full band neural network model
The spectral reflectivity of the water body to be detected acquired by the satellite is detected according to the screened actually-measured spectral reflectivity of the water body, so that the defect of low reliability of the existing satellite water body spectral reflectivity authenticity detection and verification method is overcome, and the reliability of the remote sensing-based spectral reflectivity of the water body to be detected is improved.
Optionally, on the basis of the foregoing embodiments, before the inputting the target related variable into the preset water spectral reflectance prediction model, the method further includes:
acquiring a correlation value between each water body parameter sample data and the corresponding water body sample spectral reflectivity and a correlation value between each environment sample parameter data and the corresponding water body sample spectral reflectivity based on the Pearson correlation coefficient;
acquiring sample related variables according to the correlation value between the water body sample parameter data and the corresponding water body sample spectral reflectivity and the correlation value between the environment sample parameter data and the corresponding water body sample spectral reflectivity;
and screening out target related variables from the target water body parameter data and the target environment parameter data according to the types of the sample related variables.
Specifically, as the water body parameter sample data has various types of data, the environment sample parameter data also includes various types of data, in order to screen out the data with the closest correlation value between each water body parameter sample data and the corresponding water body sample spectral reflectance and the data with the closest correlation value between each environment sample parameter data and the corresponding water body sample spectral reflectance, it is necessary to screen out each acquired correlation value according to a preset correlation value threshold, and acquire a target correlation variable according to the type of the screened sample correlation variable, where the type of the sample correlation variable is the same as the type of the target correlation variable.
For example, common environmental parameters such as temperature, humidity and atmospheric pressure and water parameters such as chlorophyll concentration, FDOM, blue-green algae protein and suspended matter concentration affect the spectral reflectance of water, and the influence intensities are different in different wavelength ranges. In addition, observation wave bands of different satellites and sensors are all specific water body parameter sensitive wave bands, in authenticity inspection experiments, the inspection of the water body reflectivity is performed according to the observation wave bands of the satellites to be inspected, and the observation wave bands of common satellites/sensors are listed in table 3. In summary, before performing correlation analysis on the water spectral reflectance and the water and environment parameters, equivalent processing needs to be performed on the reflectance data with the resolution of 1nm in the wavelength range of 400 plus 900nm, and the water spectral reflectance data between 400 plus 900nm is divided into 50 equivalent bands according to the step length of 10 nm. Compared with the original water spectral reflectivity, the equivalent reflectivity data has reduced smoothness and can better reflect the characteristic information of the water.
TABLE 3 Observation band of common satellites/Sensors
By screening out sample related variables from the water body parameter sample data and the environment sample parameter data, the accuracy of constructing the water body spectral reflectivity prediction model is improved, the accuracy of the obtained predicted water body spectral reflectivity can be further improved, and the screening of the actually measured water body spectral reflectivity is more accurate.
Optionally, on the basis of the foregoing embodiments, the obtaining of the sample-related variable specifically includes:
and screening out water body sample parameter data and environment sample parameter data which have correlation values between the water body sample parameter data and the corresponding water body sample spectral reflectivity larger than a preset correlation value threshold value from the water body sample parameter data and the environment sample parameter data as sample correlation variables.
Specifically, the method for screening out the relevant variables from the water body sample parameter data and the environment sample parameter data is to compare the relevant values with preset relevant value thresholds, and to use the water body sample parameter data and the environment sample parameter data with the relevant values larger than the preset relevant value thresholds as the sample relevant variables.
For example, the correlation of two sets of variables can be measured by the Pearson correlation coefficient r (X, Y). When the value range of the correlation coefficient | r | is between 0.2 and 0.8, the two variables are considered to have strong correlation, and water and environment variables which respectively have strong correlation with the reflectivity of 50 equivalent wave bands are sequentially screened out. Table 4 lists only some of the equivalent bands and the correlation coefficients of all variables, subject to the space of the table.
TABLE 4 correlation coefficient of partial equivalent wave band with water and environment variable
By selecting the related variables of the samples, the sample parameter data with the related values not meeting the standard can be screened from the water body sample parameter data and the environment sample parameter data, the pressure of the water body spectral reflectivity prediction model for processing the data is reduced, and the prediction accuracy is improved.
Optionally, on the basis of the foregoing embodiments, the specific steps of obtaining the water spectral reflectance prediction model are as follows:
acquiring the sample related variable and the spectral reflectivity of the water body sample corresponding to the sample related variable;
and generating model parameters of the water body spectral reflectivity prediction model based on the sample related variables and the water body sample spectral reflectivity corresponding to the sample related variables, and obtaining the water body spectral reflectivity prediction model.
Specifically, in order to generate and train a water body spectral reflectance prediction model, the sample related variables and the water body sample spectral reflectances corresponding to the sample related variables need to be obtained. And generating model parameters of a water body spectral reflectivity prediction model according to the sample related variables and the water body sample spectral reflectivity corresponding to the sample related variables, and further acquiring the water body spectral reflectivity prediction model according to the generated model parameters, wherein the model can be various neural network models, such as a BP neural network, a GRNN neural network and the like.
For example, a representative lake in China, Taihu, is taken as a research object, and the long-time sequence water spectral reflectivity, water and environment parameter data are obtained. Firstly, an automatic observation buoy system is established in the Taihu lake, and an automatic observation spectrometer, a water quality instrument and an meteorological station are arranged on a buoy. The automatic observation spectrometer observes once every 30 minutes from 10 am to 15 pm, continuously measures 10 spectral curves for standby each time, and in addition, the automatic observation spectrometer takes a picture of the lake surface and the sky before observing each time for recording weather conditions; the water quality instrument and the meteorological station work all weather, and 14 kinds of water body and environment parameter data of water temperature, specific conductance, conductivity, salinity, turbidity, chlorophyll concentration, blue algae protein, fluorescence soluble organic matter (FDOM), suspended matter concentration, average wind direction, average wind speed, air temperature, humidity and atmospheric pressure are respectively measured every 30 minutes from 0 hour to 24 hours. In the invention, water and environment variables between 10 am and 15 pm in a sunny day (which can be judged according to the pictures) are selected, on one hand, the water and environment variables correspond to the acquisition time of the spectral reflectivity data of the water, on the other hand, the requirements of an authenticity inspection experiment on weather conditions and time windows are met, and a matching data 366 group is obtained in total. Sample dependent variables may be obtained from the acquired parametric data.
According to the specific situation of the obtained parameter data, the obtained parameter data can be preprocessed according to the needs, and the preprocessing mainly comprises deleting obvious gross error data and reducing data errors.
Model parameters of the water body spectral reflectivity prediction model are generated through the sample related variables and the water body sample spectral reflectivity corresponding to the sample related variables, the water body spectral reflectivity prediction model can be obtained, the actually measured water body spectral reflectivity can be screened through the water body spectral reflectivity prediction model, and the accuracy of the actually measured water body spectral reflectivity is improved.
Optionally, on the basis of each of the above embodiments, the screening of the actually measured water spectral reflectance according to the predicted water spectral reflectance specifically includes:
calculating a first error value between the predicted water spectral reflectivity and the measured water spectral reflectivity;
and deleting the data of which the first error value exceeds a first preset threshold value from the actually measured water body spectral reflectivity.
Specifically, the screening of the actually measured water spectral reflectivity is to determine whether an error value between the actually measured water spectral reflectivity and the predicted water spectral reflectivity exceeds a preset threshold, determine that the actually measured water spectral reflectivity exceeding the preset threshold is unqualified if the error value exceeds the preset threshold, and delete the unqualified actually measured water spectral reflectivity.
For example, when an error value between a certain measured water spectral reflectance and a corresponding predicted water spectral reflectance is 0.5 and a preset threshold of the error value is 0.2, the measured water spectral reflectance exceeds the preset threshold, and the measured water spectral reflectance is deleted.
By deleting unqualified actually-measured water body spectral reflectivity, the precision of the actually-measured water body spectral reflectivity is improved, the spectral reflectivity of the water body to be detected collected by the satellite is further detected according to the actually-measured water body spectral reflectivity, and the accuracy and reliability of the detection of the spectral reflectivity authenticity of the water body to be detected obtained by the satellite are improved.
The device for testing the spectral reflectivity of the lake water body provided by the embodiment of the invention is described below, and the device for testing the spectral reflectivity of the lake water body described below and the method for testing the spectral reflectivity of the lake water body described above can be referred to correspondingly.
Fig. 2 is a schematic structural diagram of a device for testing spectral reflectance of a lake water body according to an embodiment of the present invention, as shown in fig. 2, specifically including: the system comprises a screening module 201 and a checking module 202, wherein the screening module 201 is used for screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment; the inspection module 202 is configured to inspect and correct the satellite inversion algorithm according to the screened actually-measured water body spectral reflectance, and acquire the water body spectral reflectance to be measured by using the corrected satellite inversion algorithm.
Specifically, the actually measured water spectral reflectance is the spectral reflectance obtained by sampling and detecting the water to be measured through the measuring equipment, so that the problems of inaccurate measured value and low reliability may exist in the actually measured water spectral reflectance. The screening module 201 is configured to screen the actually measured water spectral reflectance through a preset water spectral reflectance prediction model, that is, perform error determination on the actually measured water spectral reflectance and the predicted water spectral reflectance generated in the preset water spectral reflectance prediction model, determine that the actually measured water spectral reflectance meets the requirement if the error is within a preset range, and correct the actually measured water spectral reflectance if the error exceeds the preset range.
The inspection module 202 is configured to inspect and correct the satellite inversion algorithm according to the screened actually-measured water body spectral reflectance, and acquire the water body spectral reflectance to be measured by using the corrected satellite inversion algorithm.
And if the error value between the filtered actually-measured water body spectral reflectivity and the spectral reflectivity of the water body to be measured is larger than a preset threshold value, the accuracy of the spectral reflectivity of the water body to be measured is lower, and then the filtered actually-measured water body spectral reflectivity is used as truth value data to perform modeling analysis on the water body reflectivity of the remote sensing satellite, so that the inversion precision of the water body reflectivity is improved.
For example, if the error value between the spectral reflectance of the screened actually measured water body and the spectral reflectance of the water body to be measured is 0.5, and the error value between the spectral reflectance of the preset actually measured water body and the spectral reflectance of the water body to be measured is 0.2, it is determined that the error of the spectral reflectance of the water body to be measured is large, and the error needs to be corrected. And modeling and correcting the spectral reflectivity of the water body to be detected acquired by satellite inversion by using the screened actually-measured water body spectral reflectivity. The remote sensing satellite water body reflectivity is modeled and analyzed by using the screened actually-measured water body spectral reflectivity as truth data, so that the inversion precision of the water body reflectivity is improved, and the accuracy and reliability of the spectral reflectivity of the water body to be detected, which is acquired by the satellite, are improved.
Fig. 3 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 3: a processor (processor)301, a communication Interface (communication Interface)302, a memory (memory)303 and a communication bus 304, wherein the processor 301, the communication Interface 302 and the memory 303 complete communication with each other through the communication bus 304. The processor 301 may invoke logic instructions in the memory 303 to perform a method of water spectral reflectance detection, the method comprising: screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment; and checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the water body spectral reflectivity to be measured by using the corrected satellite inversion algorithm.
In addition, the logic instructions in the memory 303 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, an embodiment of the present invention further provides a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by a computer, the computer can execute the method for detecting the spectral reflectivity of the water body provided by the above-mentioned method embodiments, the method includes: screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment; and checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the water body spectral reflectivity to be measured by using the corrected satellite inversion algorithm.
In yet another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented by a processor to execute the method for detecting spectral reflectivity of a water body provided by the foregoing embodiments, and the method includes: screening the actually measured water spectral reflectivity by using a preset water spectral reflectivity prediction model; the water body spectral reflectivity prediction model is obtained by training based on water body parameter sample data, environment parameter sample data and corresponding water body sample spectral reflectivity; the actually measured water body spectral reflectivity is obtained by sampling and detecting the water body to be measured through measuring equipment; and checking and correcting the satellite inversion algorithm according to the screened actually-measured water body spectral reflectivity, and acquiring the water body spectral reflectivity to be measured by using the corrected satellite inversion algorithm.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.