CN115690598A - Outdoor environment temperature remote sensing inversion method based on satellite data and random forest model - Google Patents
Outdoor environment temperature remote sensing inversion method based on satellite data and random forest model Download PDFInfo
- Publication number
- CN115690598A CN115690598A CN202211293285.3A CN202211293285A CN115690598A CN 115690598 A CN115690598 A CN 115690598A CN 202211293285 A CN202211293285 A CN 202211293285A CN 115690598 A CN115690598 A CN 115690598A
- Authority
- CN
- China
- Prior art keywords
- data
- outdoor environment
- environment temperature
- random forest
- remote sensing
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 74
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 52
- 238000010801 machine learning Methods 0.000 claims abstract description 33
- 238000007781 pre-processing Methods 0.000 claims abstract description 19
- 238000012549 training Methods 0.000 claims description 18
- 238000003066 decision tree Methods 0.000 claims description 16
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 4
- GNFTZDOKVXKIBK-UHFFFAOYSA-N 3-(2-methoxyethoxy)benzohydrazide Chemical compound COCCOC1=CC=CC(C(=O)NN)=C1 GNFTZDOKVXKIBK-UHFFFAOYSA-N 0.000 claims description 3
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000002310 reflectometry Methods 0.000 claims description 3
- 238000002790 cross-validation Methods 0.000 claims 1
- 238000005259 measurement Methods 0.000 claims 1
- 238000009826 distribution Methods 0.000 abstract description 19
- 238000010586 diagram Methods 0.000 description 8
- 239000003570 air Substances 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 238000003860 storage Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 244000025254 Cannabis sativa Species 0.000 description 1
- 239000012080 ambient air Substances 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 238000009423 ventilation Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Image Processing (AREA)
Abstract
The invention belongs to the technical field of resources and environment, and particularly relates to an outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model, which is implemented by acquiring data required by the outdoor environment temperature remote sensing inversion method; preprocessing the data; acquiring LST, NDVI, MNDWI and NDBI according to the preprocessed data; taking the computed LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables, matching the independent variables with the actually measured outdoor environment temperature of the target variable, making a sample data set, and constructing a random forest machine learning model; and acquiring the outdoor environment temperature OAT based on an optimal random forest machine learning model, and realizing the inversion of large-range and refined outdoor environment temperature based on quasi-real-time satellite remote sensing images, so that people can more specifically and more truly sense the cold and hot conditions of the external environment, the heat distribution condition of a large-range near-ground ecosystem can be more objectively and more finely reflected, and the method has obvious advantages compared with the temperature spatial distribution interpolated on the basis of meteorological observation sites.
Description
Technical Field
The invention belongs to the technical field of resources and environment, and particularly relates to an outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model.
Background
The Outdoor Ambient Temperature (OAT) is a physical quantity representing the cold and hot degree of the Outdoor environment, and reflects the Temperature of the near-ground air under natural conditions, which can be expressed by Temperature in units of ℃. The outdoor environment temperature OAT reflects the temperature condition of the natural and real environment outside the louver box, and can objectively and truly reflect the perception of the human body to the cold and hot degree of the external natural environment.
Meteorological air temperature, usually measured by a meteorological station in a shelter at a height of 1.5 or 2m from the surface of the table, represents the average ambient air temperature on the near-ground over a large range around this height, being ventilated and not directly radiated by the sun at a height of 1.5m from the grass, and is an air temperature close to the ideal.
Due to differences in the underlying surface type, solar radiation, ventilation conditions, etc., there is often a significant difference between the outdoor ambient temperature OAT and the air temperature inside the louvre. In addition, the weather station is limited and the spatial distribution is not uniform, and the range that the temperature observed by the weather station can represent is limited, so that it is difficult to express the horizontal spatial distribution of the ambient temperature at the local position in the area in detail.
Thermal infrared remote sensing images with the characteristic of spatial continuous imaging become important data sources for acquiring temperature information of the subsurface of the earth. The geostationary satellite has high time resolution, the time interval can reach the minute level, the same region can be continuously observed, and the geostationary satellite has unique advantages in the research of near-ground temperature inversion and daily variation. Because the air temperature and the LST have a correlation relationship with definite physical significance, the air temperature can be indirectly estimated through the LST, and the previous research is mostly focused on estimating the near-ground air temperature measured in the louver box by using the LST inverted by a satellite. Due to the lack of observation of the outdoor ambient temperature outside the louvre, no method has been reported for estimating OAT by LST. Therefore, a remote sensing inversion method for rapidly and accurately acquiring the wide-range refined outdoor environment temperature OAT is urgently needed, and a basis is provided for guiding production and life and carrying out scientific health management in a more targeted manner.
Disclosure of Invention
The invention aims to provide an outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model.
In order to solve the technical problem, the invention provides an outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model, which comprises the following steps:
acquiring data required by an outdoor environment temperature remote sensing inversion method;
preprocessing the data;
inverting the surface temperature LST through a window splitting algorithm according to the preprocessed data;
calculating a normalized vegetation index NDVI according to the preprocessed data;
calculating an improved normalized difference water body index MNDWI according to the preprocessed data;
calculating a normalized building index (NDBI) according to the preprocessed data;
taking the computed LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables, matching the independent variables with target variables, making a sample data set, and constructing a random forest machine learning model; and
and acquiring the outdoor environment temperature OAT based on the optimal random forest machine learning model.
Further, the method for acquiring the data required by the outdoor environment temperature remote sensing inversion method comprises the following steps:
himapari-8 satellite data in clear sky and no cloud, outdoor environment temperature ground observation data at a time close to the passing of the satellite and digital elevation DEM data are obtained.
Further, the method for preprocessing the data comprises the following steps:
preprocessing Himapari-8 satellite data to obtain brightness and reflectivity data of 1-16 channels, wherein the projection mode is equal longitude and latitude projection, and the spatial resolution is 2KM.
Further, the method for inverting the surface temperature LST by the split window algorithm according to the preprocessed data comprises the following steps:
inverting the surface temperature LST by a split window algorithm according to the preprocessed Himapari-8 satellite data,
wherein, T i And T j Light temperatures for channels 14 and 15, respectively; t is a unit of s Is the surface temperature LST; epsilon = (epsilon) i +ε j ) 2, is the average emissivity of the 14 and 15 channels; Δ ε = ε i -ε j Is the difference between the emissivity of the 14 and 15 channels; a is a 0 ～a 6 Are algorithmic regression coefficients.
Further, the method of calculating the normalized vegetation index NDVI from the preprocessed data includes:
calculating a normalized vegetation index NDVI according to the preprocessed Himapari-8 satellite data,
NDVI＝(NIR-R)/(NIR+R)；
where NIR is the reflectance of 4 channels and R is the reflectance of 3 channels.
Further, the method for calculating the improved normalized difference water body index MNDWI according to the preprocessed data comprises the following steps:
calculating an improved normalized difference water body index MNDWI according to the preprocessed Himapari-8 satellite data,
MNDWI＝(Geen-MIR)/(Green+MIR)；
where Green is the reflectance of 2 channels and MIR is the reflectance of 7 channels.
Further, the method for calculating the normalized building index NDBI according to the preprocessed data includes:
according to the preprocessed Himapari-8 satellite data, normalizing the building index NDBI,
NDBI＝(SWIR-NIR)/(SWIR+NIR)；
where SWIR is the reflectance of 5 or 6 channels and NIR is the reflectance of 4 channels.
Further, the method for matching the computed LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables with target variables to produce a sample data set and constructing the random forest machine learning model comprises the following steps:
matching the calculated LST, NDVI, MNDWI and NDBI and a pre-acquired DEM as independent variables with the ground measured outdoor environment temperature data, making a sample data set, constructing a random forest machine learning model,
step a, corresponding longitude and latitude coordinates of an outdoor environment temperature observation station to pixel information in a satellite image, taking the pixel as a center, acquiring a 3 x 3 pixel grid point area, and calculating an average value of pixels in the pixel grid point area;
and b, matching the average value of the pixels with the outdoor environment temperature value of the corresponding site at the corresponding moment to form a sample data set of an independent variable and a target variable, wherein the independent variable comprises: LST, NDVI, MNDWI, NDBI, and DEM data, the target variables comprising: actually measuring outdoor environment temperature data on the ground;
step c, dividing the sample data set into 10 subsamples, using one single subsample as a test set for verifying the model, and using the other 9 subsamples for training;
step d, first model training: combining the 9 subsamples to form a training set, randomly extracting N times from the training set to obtain N new training sets, and forming out-of-bag data OOB by the unextracted part;
step e, generating a decision tree for each training set, selecting mtry nodes from the independent variables for each node of the decision tree, and performing branch growth according to the minimum principle of node purity;
step f, repeating the step e for N times to obtain N decision trees to form a random forest;
step g, the result of the random forest is the result obtained by each decision tree through a simple average method, and the prediction precision is determined by the average OOB of each decision tree;
and h, evaluating the models by using the correlation coefficient R and the root mean square error RMSE, applying the models to the rest 1 test set, and averaging the first 10 models with the highest precision to obtain one model.
And step i, repeating the steps d to g for 10 times in sequence by adopting a ten-fold cross verification method, and taking one model with the highest precision as an optimal random forest machine learning model.
Further, the method for obtaining the outdoor environment temperature OAT based on the optimal random forest machine learning model comprises the following steps:
and inputting LST, NDVI, MNDWI, NDBI and DEM data of different times based on the optimal random forest machine learning model, and outputting corresponding time-dependent outdoor environment temperature OAT.
On the other hand, the invention also provides an inversion system adopting the outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model, which comprises the following steps:
the acquisition module acquires data required by an outdoor environment temperature remote sensing inversion method;
the preprocessing module is used for preprocessing the data;
the calculation module is used for inverting the surface temperature LST through a split window algorithm according to the preprocessed data, calculating a normalized vegetation index NDVI according to the preprocessed data, calculating an improved normalized difference water body index MNDWI according to the preprocessed data, and normalizing a building index NDBI according to the preprocessed data;
the model construction module is used for matching the calculated LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables with target variables to produce a sample data set and construct a random forest machine learning model; and
and the acquisition module acquires the outdoor environment temperature OAT based on the optimal random forest machine learning model.
The method has the advantages that data required by the outdoor environment temperature remote sensing inversion method are obtained; preprocessing the data; inverting the surface temperature LST by a window splitting algorithm according to the preprocessed data; calculating a normalized vegetation index NDVI according to the preprocessed data; calculating an improved normalized difference water body index MNDWI according to the preprocessed data; normalizing the building index NDBI according to the preprocessed data; taking DEM data in the LST, NDVI, MNDWI, NDBI and data after calculation as independent variables, matching the independent variables with target variables, making a sample data set, and constructing a random forest machine learning model; and acquiring the outdoor environment temperature OAT based on an optimal random forest machine learning model, and realizing the inversion of large-range and refined outdoor environment temperature based on quasi-real-time satellite remote sensing images, so that people can more specifically and more truly sense the cold and hot conditions of the external environment, the heat distribution condition of a large-range near-ground ecosystem can be more objectively and more finely reflected, and the method has obvious advantages compared with the temperature spatial distribution interpolated on the basis of meteorological observation sites.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
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 embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of an outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model according to the invention;
FIG. 2 is a schematic diagram of a decision tree error analysis of the present invention;
FIG. 3 is a distribution diagram of an outdoor environment temperature observation station in Jiangsu province in the invention;
FIG. 4 is a schematic diagram of model accuracy verification in the present invention;
fig. 5 is a space distribution diagram of remote sensing inversion of outdoor environment temperature in Jiangsu province in the invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
Example 1
As shown in fig. 1 and fig. 2, this embodiment 1 provides an outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model, including: acquiring data required by an outdoor environment temperature remote sensing inversion method; preprocessing the data; inverting the surface temperature LST by a window splitting algorithm according to the preprocessed data; calculating a normalized vegetation index NDVI according to the preprocessed data; calculating an improved normalized difference water body index MNDWI according to the preprocessed data; calculating a normalized building index (NDBI) according to the preprocessed data; taking DEM data in the LST, NDVI, MNDWI, NDBI and data after calculation as independent variables, matching the independent variables with target variables, making a sample data set, and constructing a random forest machine learning model; and acquiring the outdoor environment temperature OAT based on an optimal random forest machine learning model, and realizing the inversion of large-range and refined outdoor environment temperature based on quasi-real-time satellite remote sensing images, so that people can more specifically and more truly sense the cold and hot conditions of the external environment, the heat distribution condition of a large-range near-ground ecosystem can be more objectively and more finely reflected, and the method has obvious advantages compared with the temperature spatial distribution interpolated on the basis of meteorological observation sites.
In this embodiment, the method for acquiring data required by the outdoor environment temperature remote sensing inversion method includes: himapari-8 satellite data when clear sky is clear and no cloud is available, outdoor environment temperature ground observation data at a time close to the passing of the satellite and digital elevation DEM data (such as digital elevation DEM data of Jiangsu province).
In this embodiment, the method for preprocessing data includes: preprocessing Himapari-8 satellite data to obtain brightness and reflectivity data of 1-16 channels, wherein the projection mode is equal longitude and latitude projection, the spatial resolution is 2KM, and the whole Jiangsu province area is covered; preprocessing the Himapari-8 satellite data can facilitate subsequent processing of the data and acquisition of required data according to the preprocessed data.
In this embodiment, the method for inverting the surface temperature LST according to the preprocessed data by the window splitting algorithm includes: inverting the surface temperature LST by a window splitting algorithm according to the preprocessed Himapari-8 satellite data,
wherein, T i And T j Light temperatures of 14 and 15 channels, respectively; t is a unit of s Surface temperature LST; ε = (ε) i +ε j ) 2, is the average emissivity of the 14 and 15 channels; Δ ε = ε i -ε j Is the difference between the emissivity of the 14 and 15 channels; a is 0 ～a 6 Is an algorithm regression coefficient which is obtained by atmospheric radiation transmission simulation calculation and is respectively-204.1781, 2.5512, -0.0029, 2.3468, 0.1376, 39.9991 and-170.1487.
In this embodiment, the method for calculating the normalized vegetation index NDVI according to the preprocessed data includes: calculating a normalized vegetation index NDVI from the preprocessed himwari-8 satellite data, NDVI = (NIR-R)/(NIR + R); where NIR is the reflectance of 4 channels and R is the reflectance of 3 channels.
In this embodiment, the method for calculating an improved normalized difference water body index MNDWI from the preprocessed data includes: calculating an improved normalized difference water body index MNDWI according to the preprocessed Himapari-8 satellite data, wherein the MNDWI is defined as (Geen-MIR)/(Green + MIR); where Green is the reflectance of 2 channels and MIR is the reflectance of 7 channels.
In this embodiment, the method for calculating the normalized building index NDBI according to the preprocessed data includes: calculating a normalized building index NDBI from the preprocessed himwari-8 satellite data, NDBI = (SWIR-NIR)/(SWIR + NIR); where SWIR is the reflectance of 5 or 6 channels and NIR is the reflectance of 4 channels.
In this embodiment, the method for matching the calculated LST, NDVI, MNDWI, NDBI and DEM data in the data as arguments with target variables to produce a sample data set and constructing a random forest machine learning model includes: matching the calculated LST, NDVI, MNDWI and NDBI and a pre-acquired DEM as independent variables with the ground measured outdoor environment temperature data, making a sample data set, constructing a random forest machine learning model,
step a, corresponding longitude and latitude coordinates of an outdoor environment temperature observation station to pixel information in a satellite image, taking the pixel as a center, acquiring a 3 x 3 pixel grid point area, and calculating an average value of pixels of the pixel grid point area;
and b, matching the average value of the pixels with the outdoor environment temperature value of the corresponding time and the corresponding site to form a sample data set of an independent variable and a target variable, wherein the independent variable comprises: LST, NDVI, MNDWI, NDBI, and DEM data, the target variables comprising: actually measuring outdoor environment temperature data on the ground;
step c, dividing the sample data set into 10 subsamples, using one single subsample as a test set for verifying the model, and using the other 9 subsamples for training;
step d, training a model for the first time: combining the 9 subsamples to form a training set, randomly extracting N times from the training set to obtain N new training sets, and forming out-of-bag data OOB by the unextracted part;
step e, generating a decision tree for each training set, selecting mtry from independent variables for each node of the decision tree, and performing branch growth according to the minimum node purity principle, wherein the value of the general mtry is one third of the number of the variables, namely mtry =2;
step f, repeating the step e for N times to obtain N decision trees to form a random forest;
step g, the result of the random forest is the result obtained by each decision tree through a simple averaging method, the prediction precision is determined by using the average OOB of each decision tree, and when N =400, the model tends to be stable, so N can be 400;
step h, evaluating the models by adopting the correlation coefficient R and the root mean square error RMSE, applying the models to the rest 1 test set, and averaging the first 10 models with the highest precision to obtain one model;
and step i, repeating the steps d to g for 10 times in sequence by adopting a ten-fold cross verification method, and taking one model with the highest precision as an optimal random forest machine learning model.
In this embodiment, the method for obtaining the outdoor environment temperature OAT based on the optimal random forest machine learning model includes: based on an optimal random forest machine learning model, LST, NDVI, MNDWI, NDBI and DEM data of different times are input, outdoor environment temperature OAT of corresponding times is output, large-range and refined outdoor environment temperature is inverted based on quasi-real-time satellite remote sensing images, people can more specifically and more truly sense the cold and hot conditions of the external environment, the heat distribution condition of a large-range near-ground ecosystem can be reflected more objectively and more finely, and the temperature spatial distribution interpolation method based on meteorological observation sites has obvious advantages.
Specifically, in this embodiment, an outdoor environment temperature remote sensing inversion method based on himwari-8 satellite data and a random forest model is exemplified by specific data of a part of time in a region of Jiangsu: as shown in fig. 3, a plurality of outdoor ambient temperature observation stations were installed in jiangsu province to monitor the outdoor ambient temperature of each region in real time and according to 2019, 7 monthsThe Himapari-8 satellite data of the whole hour time of 18 days, 7 months, 19 days, 7 months, 20 days, 8 months, 11 days, 8 months, 23 days, 9 months, 4 days, 8 days, 00-16: r 2 =0.8611, rmse =1.24 ℃, MAE =1.0 ℃, wherein R is 2 The method can be used for evaluating the accuracy of the model, the closer the numerical value is to 1, the more accurate the model is, the RMSE is the root mean square error, the MAE is the average absolute error, and the outdoor environment temperature of Jiangsu province of 8 months and 2-15 days in 2022 is inverted to obtain a mean value synthetic graph (as shown in FIG. 5). The outdoor environment temperature shows obvious spatial distribution characteristics, and the spatial distribution of the temperature is finer than that of the temperature interpolated on the basis of meteorological observation stations, so that the heat distribution condition under the influence of different underlying surfaces is reflected.
Example 2
On the basis of embodiment 1, this embodiment 2 further provides an inversion system using the above outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model, including: the acquisition module acquires data required by an outdoor environment temperature remote sensing inversion method; the preprocessing module is used for preprocessing the data; the calculation module is used for inverting the surface temperature LST through a split window algorithm according to the preprocessed data, calculating a normalized vegetation index NDVI according to the preprocessed data, calculating an improved normalized difference water body index MNDWI according to the preprocessed data, and calculating a normalized building index NDBI according to the preprocessed data; the model construction module is used for matching the computed LST, NDVI, MNDWI and NDBI and DEM data in the data as independent variables with target variables, making a sample data set and constructing a random forest machine learning model; and the acquisition module acquires the outdoor environment temperature OAT based on the optimal random forest machine learning model, and realizes the inversion of large-range and refined outdoor environment temperature based on the quasi-real-time satellite remote sensing image, so that people can more specifically and more truly sense the cold and hot conditions of the external environment, the heat distribution condition of a large-range near-ground ecosystem can be more objectively and more finely reflected, and the temperature spatial distribution based on the interpolation of meteorological observation sites has obvious advantages.
In summary, the data required by the outdoor environment temperature remote sensing inversion method are obtained; preprocessing the data; inverting the surface temperature LST by a window splitting algorithm according to the preprocessed data; calculating a normalized vegetation index NDVI according to the preprocessed data; calculating an improved normalized difference water body index MNDWI according to the preprocessed data; calculating a normalized building index (NDBI) according to the preprocessed data; taking DEM data in the LST, NDVI, MNDWI, NDBI and data after calculation as independent variables, matching the independent variables with target variables, making a sample data set, and constructing a random forest machine learning model; and acquiring the outdoor environment temperature OAT based on the optimal random forest machine learning model, and realizing the inversion of large-range and refined outdoor environment temperature based on quasi-real-time satellite remote sensing images, so that people can more specifically and truthfully perceive the cold and hot conditions of the external environment, can more objectively and more finely reflect the heat distribution condition of a large-range near-ground ecosystem, and has obvious advantages compared with the temperature spatial distribution interpolated on the basis of meteorological observation sites.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. 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: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.
Claims (10)
1. An outdoor environment temperature remote sensing inversion method based on satellite data and a random forest model is characterized by comprising the following steps:
acquiring data required by an outdoor environment temperature remote sensing inversion method;
preprocessing the data;
inverting the surface temperature LST by a window splitting algorithm according to the preprocessed data;
calculating a normalized vegetation index NDVI according to the preprocessed data;
calculating an improved normalized difference water body index MNDWI according to the preprocessed data;
calculating a normalized building index (NDBI) according to the preprocessed data;
taking the computed LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables, matching the independent variables with target variables, making a sample data set, and constructing a random forest machine learning model; and
and acquiring the outdoor environment temperature OAT based on the optimal random forest machine learning model.
2. The remote sensing inversion method of outdoor environment temperature according to claim 1,
the method for acquiring the data required by the outdoor environment temperature remote sensing inversion method comprises the following steps:
and acquiring Himapari-8 satellite data when clear sky is clear and no cloud, outdoor environment temperature measured data at a moment adjacent to the satellite transit and digital elevation DEM data.
3. The remote sensing inversion method of outdoor environment temperature according to claim 2,
the method for preprocessing the data comprises the following steps:
preprocessing Himapari-8 satellite data to obtain brightness and reflectivity data of 1-16 channels, wherein the projection mode is equal longitude and latitude projection, and the spatial resolution is 2KM.
4. The remote sensing inversion method of outdoor environment temperature according to claim 3,
the method for inverting the surface temperature LST through the split window algorithm according to the preprocessed data comprises the following steps:
inverting the surface temperature LST by a split window algorithm according to the preprocessed Himapari-8 satellite data,
wherein, T i And T j Light temperatures of 14 and 15 channels, respectively; t is s Is the surface temperature LST; ε = (ε) i +ε j ) 2, is the average emissivity of the 14 and 15 channels; Δ ε = ε i -ε j Is the difference between the emissivity of the 14 and 15 channels; a is 0 ～a 6 Are algorithmic regression coefficients.
5. The remote sensing inversion method of outdoor environment temperature according to claim 4,
the method for calculating the normalized vegetation index NDVI according to the preprocessed data comprises the following steps:
calculating a normalized vegetation index NDVI according to the preprocessed Himapari-8 satellite data,
NDVI＝(NIR-R)/(NIR+R)；
where NIR is the reflectance of 4 channels and R is the reflectance of 3 channels.
6. The remote sensing inversion method of outdoor environment temperature according to claim 5,
the method for calculating the improved normalized difference water body index MNDWI according to the preprocessed data comprises the following steps:
calculating an improved normalized difference water body index MNDWI according to the preprocessed Himapari-8 satellite data,
MNDWI＝(Geen-MIR)/(Green+MIR)；
where Green is the reflectance of 2 channels and MIR is the reflectance of 7 channels.
7. The remote sensing inversion method of outdoor environment temperature according to claim 6,
the method for calculating the normalized building index NDBI according to the preprocessed data comprises the following steps:
according to the preprocessed Himapari-8 satellite data, normalizing the building index NDBI,
NDBI＝(SWIR-NIR)/(SWIR+NIR)；
where SWIR is the reflectance of 5 or 6 channels and NIR is the reflectance of 4 channels.
8. The remote sensing inversion method of outdoor environment temperature according to claim 7,
the method for matching the computed LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables with target variables to produce a sample data set and constructing the random forest machine learning model comprises the following steps:
matching the calculated LST, NDVI, MNDWI and NDBI and a pre-acquired DEM as independent variables with ground actual measurement outdoor environment temperature data, making a sample data set, constructing a random forest machine learning model,
step a, corresponding longitude and latitude coordinates of an outdoor environment temperature observation station to pixel information in a satellite image, taking the pixel as a center, acquiring a 3 x 3 pixel grid point area, and calculating an average value of pixels in the pixel grid point area;
and b, matching the average value of the pixels with the outdoor environment temperature value of the corresponding time and the corresponding site to form a sample data set of an independent variable and a target variable, wherein the independent variable comprises: LST, NDVI, MNDWI, NDBI, and DEM data, the target variables comprising: actually measuring outdoor environment temperature data on the ground;
step c, dividing the sample data set into 10 subsamples, taking one single subsample as a test set to verify the model, and using the other 9 subsamples for training;
step d, first model training: combining the 9 subsamples to form a training set, randomly extracting N times from the training set to obtain N new training sets, and forming out-of-bag data OOB by the unextracted part;
step e, generating a decision tree for each training set, selecting mtry nodes from the independent variables for each node of the decision tree, and performing branch growth according to the minimum principle of node purity;
step f, repeating the step e for N times to obtain N decision trees to form a random forest;
step g, the result of the random forest is the result obtained by each decision tree through a simple average method, and the prediction precision is determined by the average OOB of each decision tree;
and h, evaluating the models by adopting the correlation coefficient R and the root mean square error RMSE, applying the models to the rest 1 test set, and averaging the first 10 models with the highest precision to obtain one model.
And step i, repeating the steps d to g for 10 times in sequence by adopting a ten-fold cross-validation method, and taking one model with the highest precision as an optimal random forest machine learning model.
9. The remote sensing inversion method of outdoor environment temperature according to claim 8,
the method for acquiring the outdoor environment temperature OAT based on the optimal random forest machine learning model comprises the following steps:
and inputting LST, NDVI, MNDWI, NDBI and DEM data of different times based on the optimal random forest machine learning model, and outputting corresponding time outdoor environment temperature OAT.
10. An inversion system using the satellite data and random forest model-based outdoor environment temperature remote sensing inversion method according to claim 1, comprising:
the acquisition module acquires data required by an outdoor environment temperature remote sensing inversion method;
the preprocessing module is used for preprocessing the data;
the calculation module is used for inverting the surface temperature LST through a split window algorithm according to the preprocessed data, calculating a normalized vegetation index NDVI according to the preprocessed data, calculating an improved normalized difference water body index MNDWI according to the preprocessed data, and normalizing a building index NDBI according to the preprocessed data;
the model construction module is used for matching the calculated LST, NDVI, MNDWI, NDBI and DEM data in the data as independent variables with target variables to produce a sample data set and construct a random forest machine learning model; and
and the acquisition module acquires the outdoor environment temperature OAT based on the optimal random forest machine learning model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211293285.3A CN115690598B (en) | 2022-10-21 | 2022-10-21 | Outdoor environment temperature remote sensing inversion method based on satellite data and random forest model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211293285.3A CN115690598B (en) | 2022-10-21 | 2022-10-21 | Outdoor environment temperature remote sensing inversion method based on satellite data and random forest model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115690598A true CN115690598A (en) | 2023-02-03 |
| CN115690598B CN115690598B (en) | 2024-06-14 |
Family
ID=85067353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211293285.3A Active CN115690598B (en) | 2022-10-21 | 2022-10-21 | Outdoor environment temperature remote sensing inversion method based on satellite data and random forest model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115690598B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117828969A (en) * | 2023-08-29 | 2024-04-05 | 北华航天工业学院 | Satellite remote sensing earth surface temperature intelligent inversion method and system integrating DSM |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108985959A (en) * | 2018-08-09 | 2018-12-11 | 安徽大学 | Wheat powdery mildew remote sensing monitoring method based on surface temperature inversion technology |
| US20190057171A1 (en) * | 2017-02-10 | 2019-02-21 | Guangxi Institute of Meteorlogical Disaster- Reducing Research | Novel nonlinear method for area-wide near surface air temperature precision retrieval |
| CN109509319A (en) * | 2018-12-29 | 2019-03-22 | 北京恒泰实达科技股份有限公司 | Transmission line forest fire monitoring and pre-alarming method based on fixed statellite monitoring materials |
| US20210239606A1 (en) * | 2020-02-04 | 2021-08-05 | Andrea Gabrieli | Computationally efficient method for retrieving physical properties from 7-14 um hyperspectral imaging data under clear and cloudy background conditions |
| CN113408742A (en) * | 2021-06-24 | 2021-09-17 | 桂林理工大学 | High-precision sea surface temperature inversion method based on machine learning |
| US20220067553A1 (en) * | 2020-08-25 | 2022-03-03 | Institute Of Soil Science, Chinese Academy Of Sciences | Method for predictive soil mapping based on solar radiation in large flat area |
| CN114139444A (en) * | 2021-11-24 | 2022-03-04 | 大连海事大学 | Offshore sea surface temperature inversion method based on machine learning |
| CN114563353A (en) * | 2022-01-21 | 2022-05-31 | 河海大学 | Soil heat flux prediction method based on multi-source satellite remote sensing data |
| CN115204691A (en) * | 2022-07-13 | 2022-10-18 | 中国科学院地理科学与资源研究所 | Urban artificial heat emission estimation method based on machine learning and remote sensing technology |
-
2022
- 2022-10-21 CN CN202211293285.3A patent/CN115690598B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190057171A1 (en) * | 2017-02-10 | 2019-02-21 | Guangxi Institute of Meteorlogical Disaster- Reducing Research | Novel nonlinear method for area-wide near surface air temperature precision retrieval |
| CN108985959A (en) * | 2018-08-09 | 2018-12-11 | 安徽大学 | Wheat powdery mildew remote sensing monitoring method based on surface temperature inversion technology |
| CN109509319A (en) * | 2018-12-29 | 2019-03-22 | 北京恒泰实达科技股份有限公司 | Transmission line forest fire monitoring and pre-alarming method based on fixed statellite monitoring materials |
| US20210239606A1 (en) * | 2020-02-04 | 2021-08-05 | Andrea Gabrieli | Computationally efficient method for retrieving physical properties from 7-14 um hyperspectral imaging data under clear and cloudy background conditions |
| US20220067553A1 (en) * | 2020-08-25 | 2022-03-03 | Institute Of Soil Science, Chinese Academy Of Sciences | Method for predictive soil mapping based on solar radiation in large flat area |
| CN113408742A (en) * | 2021-06-24 | 2021-09-17 | 桂林理工大学 | High-precision sea surface temperature inversion method based on machine learning |
| CN114139444A (en) * | 2021-11-24 | 2022-03-04 | 大连海事大学 | Offshore sea surface temperature inversion method based on machine learning |
| CN114563353A (en) * | 2022-01-21 | 2022-05-31 | 河海大学 | Soil heat flux prediction method based on multi-source satellite remote sensing data |
| CN115204691A (en) * | 2022-07-13 | 2022-10-18 | 中国科学院地理科学与资源研究所 | Urban artificial heat emission estimation method based on machine learning and remote sensing technology |
Non-Patent Citations (3)
| Title |
|---|
| MANICKAVASAGAM, S: "Processing of Remote Sensing Ocean Parameter Using Downscaling and Machine Learning Techniques", 《 2021 2ND INTERNATIONAL CONFERENCE ON COMPUTATION, AUTOMATION AND KNOWLEDGE MANAGEMENT (ICCAKM)》, 21 March 2021 (2021-03-21) * |
| 华俊玮;祝善友;张桂欣;: "基于随机森林算法的地表温度降尺度研究", 国土资源遥感, no. 01, 3 February 2018 (2018-02-03) * |
| 张飞: "近20年城市遥感研究现状及其发展趋势", 《生态学报》, vol. 41, no. 8, 31 December 2021 (2021-12-31) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117828969A (en) * | 2023-08-29 | 2024-04-05 | 北华航天工业学院 | Satellite remote sensing earth surface temperature intelligent inversion method and system integrating DSM |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115690598B (en) | 2024-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| De Keukelaere et al. | Atmospheric correction of Landsat-8/OLI and Sentinel-2/MSI data using iCOR algorithm: validation for coastal and inland waters | |
| Wan et al. | Quality assessment and validation of the MODIS global land surface temperature | |
| Cheng et al. | Estimating the hemispherical broadband longwave emissivity of global vegetated surfaces using a radiative transfer model | |
| Sun et al. | Satellite data cloud detection using deep learning supported by hyperspectral data | |
| Huang et al. | A physically based algorithm for retrieving land surface temperature under cloudy conditions from AMSR2 passive microwave measurements | |
| Landier et al. | Calibration of urban canopies albedo and 3D shortwave radiative budget using remote-sensing data and the DART model | |
| Zhang et al. | Development of a high spatiotemporal resolution cloud-type classification approach using Himawari-8 and CloudSat | |
| Polyakov et al. | Using artificial neural networks in the temperature and humidity sounding of the atmosphere | |
| CN115690598B (en) | Outdoor environment temperature remote sensing inversion method based on satellite data and random forest model | |
| Schlager et al. | Generation of high-resolution wind fields from the WegenerNet dense meteorological station network in southeastern Austria | |
| Osmond | Hemispherical photography as a tool for urban sustainability evaluation and design | |
| Aires et al. | An innovative calibration method for the inversion of satellite observations | |
| Zhou et al. | Comparison of Machine‐Learning Algorithms for Near‐Surface Air‐Temperature Estimation from FY‐4A AGRI Data | |
| CN113836731B (en) | Construction method and device of land surface stable target atmosphere top reflectivity model | |
| Yan et al. | Gap filling of advanced technology microwave sounder data as applied to hurricane warm core animations | |
| CN115937714A (en) | Net Primary Productivity Estimation Method Based on Neural Network and CASA Model | |
| Mohia et al. | Combination of spectral and textural features in the MSG satellite remote sensing images for classifying rainy area into Different classes | |
| CN111274535B (en) | Dynamic evaluation method for spatial representativeness of ground station | |
| Nassar et al. | Implications of soil and canopy temperature uncertainty in the estimation of surface energy fluxes using TSEB2T and high-resolution imagery in commercial vineyards | |
| Grasso et al. | An example of the use of synthetic 3.9 μm GOES-12 imagery for two-moment microphysical evaluation | |
| Pasillas | Turning Night Into Day: The Creation, Validation, and Application of Synthetic Lunar Reflectance Values From the Day/Night Band and Infrared Sensors for Use With JPSS VIIRS and GOES ABI | |
| Hena et al. | A simple statistical model to estimate incident solar radiation at the surface from NOAA AVHRR satellite data | |
| Cheng et al. | Obtaining global land-surface broadband emissivity from MODIS collection 5 spectral albedos using a dynamic learning neural network | |
| Wu et al. | Single Field-of-view Sounding Atmospheric Products | |
| Jiang et al. | Surface global and diffuse solar radiation over China acquired from geostationary Multi-functional Transport Satellite data |
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 |