CN111199185A - Ground surface temperature downscaling method, system and equipment based on XGboost learning algorithm - Google Patents
Ground surface temperature downscaling method, system and equipment based on XGboost learning algorithm Download PDFInfo
- Publication number
- CN111199185A CN111199185A CN201911170146.XA CN201911170146A CN111199185A CN 111199185 A CN111199185 A CN 111199185A CN 201911170146 A CN201911170146 A CN 201911170146A CN 111199185 A CN111199185 A CN 111199185A
- Authority
- CN
- China
- Prior art keywords
- coverage
- resolution
- vegetation
- albedo
- normalized
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/182—Network patterns, e.g. roads or rivers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
- G06V10/143—Sensing or illuminating at different wavelengths
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- 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
- Y02A30/00—Adapting or protecting infrastructure or their operation
- Y02A30/60—Planning or developing urban green infrastructure
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Processing (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to a ground surface temperature downscaling method, a system and equipment based on an XGboost learning algorithm, the research area is divided according to the impervious surface coverage, and the different influence degrees of the urban impervious surface coverage, the vegetation coverage and the road density with different areas and low resolution on the surface temperature are combined, and the influence of the spatial heterogeneity of the urban underlying surface on the surface temperature is combined, the thermal infrared image data with low resolution is combined, a nonlinear regression model with low resolution is established, and according to the high-resolution surface parameters and the nonlinear regression model, calculating to obtain the high-resolution surface predicted temperature of the whole research area, compared with the prior art, the method realizes the high-resolution prediction of the urban environment surface temperature with complex spatial heterogeneity, can more finely distinguish the surface temperatures of roads, buildings, vegetation and water bodies, and improves the accuracy of the urban surface temperature prediction.
Description
Technical Field
The invention relates to the technical field of geographic information, in particular to a ground surface temperature downscaling method, system and device based on an XGboost learning algorithm.
Background
The surface temperature is an important physical parameter of a land surface system, and has wide application requirements in the research fields of surface evapotranspiration estimation, soil moisture estimation, urban thermal environment and the like. However, with the economic development, population increase and urbanization progress acceleration, urban land utilization and coverage types are changed obviously, the increase of artificial buildings causes the original natural vegetation and bare land to be replaced by water-impermeable underlying surfaces such as buildings, asphalt, cement and the like, and the water-impermeable underlying surfaces store heat in the daytime and release heat at night, particularly have good thermal conductivity and high thermal capacity and are one of the main reasons for the formation of the urban heat island effect.
At present, the method for carrying out spatial dimension reduction on the remote sensing earth surface temperature mainly carries out spatial dimension reduction on earth surface temperatures with low spatial resolution such as AMSR-E, MODIS, FY-3 and the like to obtain earth surface temperatures with resolution of 1km or 30 meters, and the precision of the method cannot meet the requirements of people.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a high-precision and high-resolution earth surface temperature downscaling method, system and device based on an XGboost learning algorithm, which are calculated in different regions.
An earth surface temperature downscaling method based on an XGboost learning algorithm comprises the following steps:
acquiring surface parameters and thermal infrared image data of a first resolution in a research area; wherein the surface parameters comprise impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density;
dividing the research area into high-density, medium-density and low-density subareas by utilizing a spatial clustering algorithm according to the impervious surface coverage;
aiming at each sub-region, constructing and training a non-linear regression model by taking the surface parameters of the first resolution as input variables and the thermal infrared surface temperature of the first resolution as output variables based on an XGboost algorithm to obtain residual values of the non-linear regression model;
inputting the surface parameters of the second resolution in each sub-area into the nonlinear regression model to obtain the initial predicted value of the thermal infrared surface temperature of the second resolution in each sub-area, wherein the second resolution is higher than the first resolution;
carrying out downscaling on the residual value by using a surface-to-point kriging method to obtain a residual of a second resolution of the research area; and adding the residual error of the second resolution to the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain the predicted value of the thermal infrared earth surface temperature of the second resolution.
Compared with the prior art, the method has the advantages that the research area is divided according to the impervious surface coverage, the low-resolution urban impervious surface coverage, the vegetation coverage and the different influence degrees of the road density on the earth surface temperature in different areas are combined, the influence of the spatial heterogeneity of the urban underlying surface on the earth surface temperature is combined, the low-resolution thermal infrared image data is combined, the low-resolution nonlinear regression model is established, the high-resolution earth surface predicted temperature of the whole research area is obtained through calculation according to the high-resolution earth surface parameters and the high-resolution nonlinear regression model, the high-resolution prediction of the urban environment earth surface temperature with complex spatial heterogeneity is realized, the earth surface temperatures of roads, buildings, vegetation and water bodies can be distinguished more finely, and the accuracy of the urban earth surface temperature prediction is improved.
In one embodiment, the acquiring the surface parameters of the research area at the first resolution includes:
acquiring Sentinel-2A remote sensing image data;
extracting the impervious surface coverage and vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method;
resampling near infrared band and mid infrared band data in the Sentinel-2A remote sensing image data to a resolution of 10 meters;
calculating a normalized vegetation index, a normalized construction index, an improved normalized difference water body index and road density with the resolution of 10 meters according to the resampled near-infrared band and mid-infrared band remote sensing image data;
and performing aggregate calculation on the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density with the resolution of 10 meters to obtain surface parameters with the resolution of 30 meters.
In one embodiment, the step of extracting the impervious surface coverage and the vegetation coverage of the second resolution in the remote sensing image data by using the improved linear spectrum unmixing method comprises the following steps:
removing the water body in the remote sensing image data by using the improved normalized difference water body index, selecting a high albedo, a low albedo, a vegetation end member and a soil end member as unmixed objects, and solving the proportion of the high albedo, the low albedo and the vegetation end member in a pixel by using a fully-constrained least square linear spectrum unmixed method to obtain a high albedo coverage, a low albedo coverage and a vegetation coverage;
threshold extraction is carried out on the NDBI image by utilizing an Otsu algorithm, a pixel with a normalized building index larger than a threshold value is used as an impervious surface pixel, a pixel with a normalized building index smaller than the threshold value is used as a non-impervious surface pixel, and the pixel is divided into low albedo coverage of the impervious surface and low albedo coverage of the non-impervious surface by combining the low reflectivity coverage;
further partitioning the low albedo coverage of the non-impervious surface by using a normalized vegetation index: if the normalized vegetation index is less than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of soil, and if the normalized vegetation index is more than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of vegetation;
and adding the high-albedo coverage and the low-albedo coverage of the impervious surface to obtain the impervious surface coverage, and adding the vegetation coverage and the low-albedo coverage of the vegetation to obtain the vegetation coverage.
In one embodiment, the step of constructing and training a non-linear regression model comprises:
and taking the surface parameter of the first resolution as an input variable, and taking the thermal infrared surface temperature of the first resolution as an output variable, and constructing a nonlinear regression model expression:
lst～imper_frac+veg_frac+road+mndwi+ndbi+ndvi；
wherein lst represents the thermal infrared earth surface temperature, impermeant _ frac represents the impervious surface coverage, veg _ frac represents the vegetation coverage, rod represents the road density, MNDWI represents the MNDWI, ndbi represents the normalized building index, and ndvi represents the normalized vegetation index;
setting training parameters of the nonlinear regression model, including iteration times, contraction step length, minimum weight values of sub-nodes, the number of sub-samples, a minimum loss function reduction value required by node splitting and a maximum node splitting depth, and training the nonlinear regression model.
The invention also provides an earth surface temperature downscaling system based on the XGboost learning algorithm, which comprises the following components:
the data acquisition module is used for acquiring surface parameters and thermal infrared image data of a first resolution in a research area; wherein the surface parameters comprise impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density;
the region dividing module is used for dividing the research region into high-density, medium-density and low-density sub-regions by utilizing a spatial clustering algorithm according to the water impermeability coverage;
the model construction module is used for constructing and training a nonlinear regression model by taking the surface parameters of the first resolution as input variables and the thermal infrared surface temperature of the first resolution as output variables on the basis of an XGboost algorithm for each sub-region to obtain residual values of the nonlinear regression model;
the initial value acquisition module is used for inputting the surface parameters of the second resolution in each sub-area into the nonlinear regression model to obtain the initial predicted value of the thermal infrared surface temperature of the second resolution in each sub-area, wherein the second resolution is higher than the first resolution;
the predicted value calculation module is used for carrying out downscaling on the residual value by using a surface-to-point kriging method to obtain a residual error of a second resolution of the research area; and adding the residual error of the second resolution to the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain the predicted value of the thermal infrared earth surface temperature of the second resolution.
In one embodiment, the thermal infrared image data is Landsat 8 thermal infrared image data, the first resolution is 30 meters, and the data acquisition module includes:
the data acquisition unit is used for acquiring Sentinel-2A remote sensing image data;
the extraction unit is used for extracting the impervious surface coverage and the vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method;
the resampling unit is used for resampling the near infrared band and mid infrared band data in the Sentinel-2A remote sensing image data to the resolution of 10 meters;
the index calculation unit is used for calculating a normalized vegetation index, a normalized building index, an improved normalized difference water body index and road density with the resolution of 10 meters according to the resampled near-infrared band and mid-infrared band remote sensing image data;
and the aggregation unit is used for carrying out aggregation calculation on the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density of 10 meters to obtain the land surface parameters with the resolution of 30 meters.
In one embodiment, the index calculation unit includes:
the initial coverage calculation unit is used for removing the water body in the remote sensing image data by utilizing the improved normalized difference water body index, selecting a high albedo, a low albedo, a vegetation end member and a soil end member as unmixed objects, and solving the proportion of the high albedo, the low albedo and the vegetation end member in a pixel by utilizing a fully-constrained least square linear spectrum unmixed method to obtain the high albedo coverage, the low albedo coverage and the vegetation coverage;
the impervious surface pixel extraction unit is used for performing threshold extraction on the NDBI image by utilizing an Otsu algorithm, taking the pixel with the normalized building index larger than the threshold as an impervious surface pixel, taking the pixel with the normalized building index smaller than the threshold as a non-impervious surface pixel, and dividing the pixel into impervious surface low-albedo coverage and non-impervious surface low-albedo coverage by combining the low-reflectivity coverage;
a dividing unit for further dividing the low albedo coverage of the non-impervious surface by a normalized vegetation index: if the normalized vegetation index is less than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of soil, and if the normalized vegetation index is more than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of vegetation;
and the coverage acquisition unit is used for adding the high-albedo coverage and the low-albedo coverage of the impervious surface to obtain the impervious surface coverage, and adding the vegetation coverage and the low-albedo coverage of the vegetation to obtain the vegetation coverage.
In one embodiment, the model building module comprises:
the model expression construction unit is used for constructing a nonlinear regression model expression by taking the surface parameter of the first resolution as an input variable and taking the thermal infrared surface temperature of the first resolution as an output variable:
lst～imper_frac+veg_frac+road+mndwi+ndbi+ndvi；
wherein lst represents the thermal infrared earth surface temperature, impermeant _ frac represents the impervious surface coverage, veg _ frac represents the vegetation coverage, rod represents the road density, MNDWI represents the MNDWI, ndbi represents the normalized building index, and ndvi represents the normalized vegetation index;
and the training unit is used for setting training parameters of the nonlinear regression model, including iteration times, contraction step length, minimum weight values of sub nodes, the number of sub samples, a minimum loss function reduction value required by node splitting and the maximum depth of the node splitting, and training the nonlinear regression model.
The invention also provides a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the steps of the XGBoost learning algorithm-based surface temperature downscaling method according to any one of the above.
The invention also provides a computer device, which comprises a memory, a processor and a computer program stored in the memory and executable by the processor, wherein the processor implements the steps of the surface temperature downscaling method based on the XGBoost learning algorithm according to any one of the above items when executing the computer program.
For a better understanding and practice, the invention is described in detail below with reference to the accompanying drawings.
Drawings
FIG. 1 is a flow chart of a surface temperature downscaling method based on an XGboost learning algorithm in an exemplary embodiment of the invention;
FIG. 2 is a schematic structural diagram of a surface temperature downscaling system based on an XGboost learning algorithm in an exemplary embodiment of the invention;
FIG. 3 is a graphical illustration of a Sentinel-2A image of a region of interest in an exemplary embodiment;
FIG. 4 is a schematic thermal infrared image of the Landsat 8 area of interest in an exemplary embodiment;
FIG. 5 is a schematic representation of the surface temperature of an area of interest extracted using the present invention in an exemplary embodiment.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the invention, as detailed in the appended claims.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
It is to be understood that although the terms first, second, third, etc. may be used herein to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, the second information may also be referred to as the first information, and similarly, the first information may also be referred to as the third information, without departing from the scope of the present invention. The word "if/if" as used herein may be interpreted as "at … …" or "when … …" or "in response to a determination", depending on the context.
Referring to fig. 1, in an embodiment of the present invention, a ground surface temperature downscaling method based on an XGBoost learning algorithm is provided, including the following steps:
step S1: acquiring surface parameters and thermal infrared image data of a first resolution in a research area; wherein the surface parameters comprise impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density;
the research area is a set land surface area, and in the embodiment of the application, the research area mainly refers to an urban area.
The first resolution is spatial image resolution, the image resolution refers to reflection of ground resolution on specific images with different scales, the image resolution changes with different scales of the images, pixels are in direct proportion to the resolution, and the larger the pixel is, the higher the resolution is.
The earth surface parameters are related parameters indicating earth surface environment, wherein the impervious surface coverage and the vegetation coverage are used for indicating the impervious surface and the vegetation coverage in the research area, and the normalized vegetation index (NDVI) refers to the sum of the reflection value of a near infrared band and the reflection value of a red light band in a remote sensing image, and can be used for detecting the vegetation growth state, the vegetation coverage and eliminating partial radiation errors. The normalized building index (NDBI) represents urban building utilization and can be used for extracting urban building areas. The improved normalized difference water body index (MNDWI) is usually used for extracting water bodies in urban areas, can be used for revealing fine characteristics of the water bodies, such as distribution of suspended sediments, change of water quality and the like, and is helpful for distinguishing shadows and the water bodies and improving the water body extraction quality by utilizing the improved normalized difference water body index (MNDWI).
The thermal infrared image data is thermal infrared radiation information which is collected and recorded by a thermal infrared detector and cannot be seen by human eyes and is radiated by ground objects, and the temperature parameter of the ground surface can be acquired by utilizing the thermal infrared information.
Step S2: dividing the research area into high-density, medium-density and low-density subareas by utilizing a spatial clustering algorithm according to the impervious surface coverage;
the spatial clustering algorithm is an unsupervised learning method which divides objects in a spatial data set into classes consisting of similar objects, wherein the similar objects have higher similarity, and the different objects have larger difference. And obtaining the impervious surface aggregation density through a spatial clustering algorithm, wherein the impervious surface aggregation density represents the aggregation degree and distribution density of the impervious surface within a certain radius range by taking the pixel as a center.
In the embodiment of the application, the research area is divided into sub-areas with high density, medium density and low density through the impervious surface gathering density, the impervious surface gathering density can be used for distinguishing the gathering degree of buildings, the impervious surface gathering density of a main urban area is high, and the impervious surface density gathering of a suburban area is sparse.
Step S3: aiming at each sub-region, constructing and training a non-linear regression model by taking the surface parameters of the first resolution as input variables and the thermal infrared surface temperature of the first resolution as output variables based on an XGboost algorithm to obtain residual values of the non-linear regression model;
the extreme gradient boost algorithm (XGboost algorithm) is characterized in that a plurality of decision trees are used as base classifiers, a next decision tree is fitted according to a residual error between an output result of a previous decision tree and an actual value, and a predicted value is obtained by summing output results of the decision trees.
If the dependent variable of the regression model is a function of the independent variable more than once, the regression law is graphically represented by various curves with different forms, and the model is called a nonlinear regression model.
Step S4: inputting the surface parameters of the second resolution in each sub-area into the nonlinear regression model to obtain the initial predicted value of the thermal infrared surface temperature of the second resolution in each sub-area, wherein the second resolution is higher than the first resolution;
step S5: carrying out downscaling on the residual value by using a surface-to-point kriging method to obtain a residual of a second resolution of the research area; and adding the residual error of the second resolution to the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain the predicted value of the thermal infrared earth surface temperature of the second resolution.
The surface-to-point Kriging method is a statistical method for estimating the average section position of a central block section by giving different weights to the taste of each sample according to different sample space positions and different correlation degrees among samples and carrying out sliding weighted average. In the step, the residual error of the first resolution ratio is downscaled into the residual error of the second resolution ratio by using a surface-to-point kriging method, and the residual error of the whole research area is added with the initial predicted values of the thermal infrared surface temperature in each sub-area to obtain the predicted value of the thermal infrared surface temperature of the second resolution ratio.
According to the method, a research area is divided according to the impervious surface coverage, different influence degrees of low-resolution urban impervious surface coverage, vegetation coverage and road density on the surface temperature in different areas are combined, the influence of spatial heterogeneity of an urban underlying surface on the surface temperature is combined, low-resolution thermal infrared image data is combined, a low-resolution nonlinear regression model is established, and the high-resolution surface prediction temperature of the whole research area is calculated according to the high-resolution surface parameters and the nonlinear regression model, so that the high-resolution prediction on the urban environment surface temperature with complicated spatial heterogeneity is realized, the surface temperatures of roads, buildings, vegetation and water bodies can be more finely distinguished, and the accuracy of the urban surface temperature prediction is improved.
In an exemplary embodiment, the first resolution is 30 meters, the thermal infrared image data is Landsat 8 thermal infrared image data, and the Landsat 8 thermal infrared image data refers to thermal infrared image data obtained by a Landsat 8 thermal infrared sensor and has a spatial resolution of 30 meters; the surface parameters of the second resolution ratio refer to impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density with the spatial resolution ratio of 10 meters, the surface parameters of the first resolution ratio are obtained by calculation from remote sensing image data, and specifically, the step of obtaining the surface parameters of the first resolution ratio of the research area comprises the following steps:
s101, acquiring Sentinel-2A remote sensing image data;
the Sentinel-2A data refers to the image data related to the earth surface object acquired by the Sentinel-2A satellite, and comprises near infrared band data and middle infrared band data with the spatial resolution of 20 meters.
Step S102: extracting the impervious surface coverage and vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method;
step S103: resampling near infrared band and mid infrared band data in the Sentinel-2A remote sensing image data to a resolution of 10 meters;
step S104, calculating a normalized vegetation index, a normalized construction index, an improved normalized difference water body index and road density with the resolution of 10 meters according to the resampled near-infrared band and mid-infrared band remote sensing image data;
step S105, carrying out polymerization calculation on the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density with the resolution of 10 meters to obtain a ground surface parameter with the resolution of 30 meters;
and the aggregation calculation arranges a plurality of pixels together to form a new pixel mode, and upscales the surface parameters of the second resolution ratio to obtain the surface parameters consistent with the resolution ratio of the thermal infrared image data, so that the construction and training of subsequent models are facilitated.
In an exemplary embodiment, the step of extracting the impervious surface coverage and the vegetation coverage with the resolution of 10 meters in the remote sensing image data by using the improved linear spectrum unmixing method specifically includes:
removing the water body in the remote sensing image data by using an improved normalized difference water body index, selecting a high albedo, a low albedo and a vegetation end member as unmixed objects, and solving the proportion of the high albedo, the low albedo and the vegetation end member in a pixel by using a fully-constrained least square linear spectrum unmixed method to obtain a high albedo coverage, a low albedo coverage and a vegetation coverage;
threshold extraction is carried out on the NDBI image by utilizing an Otsu algorithm, a pixel with a normalized building index larger than a threshold value is used as an impervious surface pixel, a pixel with a normalized building index smaller than the threshold value is used as a non-impervious surface pixel, and the pixel is divided into low albedo coverage of the impervious surface and low albedo coverage of the non-impervious surface by combining the low reflectivity coverage; the Otsu algorithm is an efficient algorithm for carrying out binarization on an image, and the algorithm assumes that image pixels can be divided into a background part and a target part according to a threshold value, and calculates an optimal threshold value to distinguish the two types of pixels, so that the distinguishing degree of the two types of pixels is maximum.
Further partitioning the low albedo coverage of the non-impervious surface by using a normalized vegetation index: if the normalized vegetation index is less than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of soil, and if the normalized vegetation index is more than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of vegetation;
and adding the high-albedo coverage and the low-albedo coverage of the impervious surface to obtain the impervious surface coverage, and adding the vegetation coverage and the low-albedo coverage of the vegetation to obtain the vegetation coverage.
The improved linear spectrum unmixing method comprises the steps of manually selecting high albedo, low albedo, vegetation end members and soil end members, optimizing end member selection, screening and removing abnormal parts in data obtained by manual end member selection by utilizing the sensitivity of a normalized building index to a building area and the sensitivity of a normalized vegetation index to vegetation, and obtaining high-precision impervious surface coverage and vegetation coverage.
In an exemplary embodiment, the step of constructing and training a non-linear regression model based on the XGBoost algorithm with the surface quantity of the first resolution as an input variable and the thermal infrared surface temperature of the first resolution as an output variable for each sub-region includes:
step S301: and taking the surface parameter of the first resolution as an input variable, and taking the thermal infrared surface temperature of the first resolution as an output variable, and constructing a nonlinear regression model expression:
lst～imper_frac+veg_frac+road+mndwi+ndbi+ndvi；
wherein lst represents the thermal infrared earth surface temperature, impermeant _ frac represents the impervious surface coverage, veg _ frac represents the vegetation coverage, rod represents the road density, MNDWI represents the MNDWI, ndbi represents the normalized building index, and ndvi represents the normalized vegetation index;
step S302: setting training parameters of the nonlinear regression model, including iteration times, contraction step length, minimum weight values of sub-nodes, the number of sub-samples, a minimum loss function reduction value required by node splitting and a maximum node splitting depth, and training the nonlinear regression model.
For example, the nonlinear regression model is set as: XGBoost (data, max _ depth ═ 6, eta ═ 0.3, nround ═ 100, gamma ═ 0, min _ child _ weight ═ 1, and subsample ═ 1), where data represent input variables including Landsat 8 thermal infrared surface temperature, imper _ frac, veg _ frac, road, mdwi, ndbi, and ndvi, nround ═ 100 represents the number of iterations 100, eta ═ 0.3 represents the contraction step size 0.3, min _ child _ weight ═ 1 represents the minimum weight value of a child node 1, subsample ═ 1 represents the number of child samples 1, gamma ═ 0 represents the minimum loss function degradation value required for node splitting 0, and gamma _ depth ═ 6 represents the maximum splitting depth of the node 6.
In a preferred embodiment, the second resolution is 10 meters, the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density with the resolution of 10 meters, which are obtained in the steps S101 to S105, are input into the nonlinear regression model to obtain an initial predicted value of the thermal infrared surface temperature with the resolution of 10 meters in each sub-area, and the residual value is downscaled by using a surface-to-point kriging method to obtain a residual error with the resolution of 10 meters in the whole research area; and adding the residual error of each research area and the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain the predicted value of the thermal infrared earth surface temperature with the resolution of 10 meters.
The invention also provides an earth surface temperature downscaling system based on the XGboost learning algorithm, which comprises the following components:
the data acquisition module 1 is used for acquiring surface parameters and thermal infrared image data of a first resolution in a research area; wherein the surface parameters comprise impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density;
the region dividing module 2 is used for dividing the research region into high-density, medium-density and low-density sub-regions by utilizing a spatial clustering algorithm according to the water impermeability coverage;
the model construction module 3 is used for constructing and training a nonlinear regression model by taking the surface parameters of the first resolution as input variables and the thermal infrared surface temperature of the resolution as output variables on the basis of an XGboost algorithm for each sub-region to obtain residual values of the nonlinear regression model;
the initial value obtaining module 4 is configured to input the surface parameters of the second resolution in each sub-region into the nonlinear regression model to obtain an initial predicted value of the thermal infrared surface temperature in each sub-region, where the second resolution is higher than the first resolution;
the predicted value calculating module 5 is used for carrying out downscaling on the residual value by using a surface-to-point kriging method to obtain a residual error of a second resolution of the whole research area; and adding the residual error of the whole research area and the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain a predicted value of the thermal infrared earth surface temperature with a second resolution.
In an exemplary embodiment, the first resolution is 30 meters, the thermal infrared image data is Landsat 8 thermal infrared image data, and the data acquisition module 1 includes:
the data acquisition unit is used for acquiring Sentinel-2A remote sensing image data;
the extraction unit is used for extracting the impervious surface coverage and the vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method;
the resampling unit is used for resampling the near infrared band and mid infrared band data in the Sentinel-2A remote sensing image data to the resolution of 10 meters;
the index calculation unit is used for calculating a normalized vegetation index, a normalized building index, an improved normalized difference water body index and road density with the resolution of 10 meters according to the resampled near-infrared band and mid-infrared band remote sensing image data;
and the aggregation unit is used for carrying out aggregation calculation on the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density of 10 meters to obtain the land surface parameters with the resolution of 30 meters.
In an exemplary embodiment, the index calculation unit includes:
the initial coverage calculation unit is used for removing the water body in the remote sensing image data by utilizing the improved normalized difference water body index, selecting a high albedo, a low albedo, a vegetation end member and a soil end member as unmixed objects, and solving the proportion of the high albedo, the low albedo and the vegetation end member in a pixel by utilizing a fully-constrained least square linear spectrum unmixed method to obtain the high albedo coverage, the low albedo coverage and the vegetation coverage;
the impervious surface pixel extraction unit is used for performing threshold extraction on the NDBI image by utilizing an Otsu algorithm, taking the pixel with the normalized building index larger than the threshold as an impervious surface pixel, taking the pixel with the normalized building index smaller than the threshold as a non-impervious surface pixel, and dividing the pixel into impervious surface low-albedo coverage and non-impervious surface low-albedo coverage by combining the low-reflectivity coverage;
a dividing unit for further dividing the low albedo coverage of the non-impervious surface by a normalized vegetation index: if the normalized vegetation index is less than 0.2, the low albedo coverage of the non-impervious surface is divided into the low albedo coverage of the soil, and if the normalized vegetation index is more than 0.2, the low albedo coverage of the non-impervious surface is divided into the low albedo coverage of the vegetation;
and the coverage acquisition unit is used for adding the high-albedo coverage and the low-albedo coverage of the impervious surface to obtain the impervious surface coverage, and adding the vegetation coverage and the low-albedo coverage of the vegetation to obtain the vegetation coverage.
In an exemplary embodiment, the model building module 3 includes:
the model expression construction unit is used for constructing a nonlinear regression model expression by taking the surface parameter of the first resolution as an input variable and taking the thermal infrared surface temperature of the first resolution as an output variable:
lst～imper_frac+veg_frac+road+mndwi+ndbi+ndvi；
wherein lst represents the thermal infrared earth surface temperature, impermeant _ frac represents the impervious surface coverage, veg _ frac represents the vegetation coverage, rod represents the road density, MNDWI represents the MNDWI, ndbi represents the normalized building index, and ndvi represents the normalized vegetation index;
and the training unit is used for setting training parameters of the nonlinear regression model, including iteration times, contraction step length, minimum weight values of sub nodes, the number of sub samples, a minimum loss function reduction value required by node splitting and the maximum depth of the node splitting, and training the nonlinear regression model.
The invention also provides a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements any of the above-mentioned steps of the XGBoost learning algorithm-based surface temperature downscaling method.
The present invention may take the form of a computer program product embodied on one or more storage media including, but not limited to, disk storage, CD-ROM, optical storage, and the like, having program code embodied therein. Computer readable storage media, which include both non-transitory and non-transitory, removable and non-removable media, may implement any method or technology for storage of information. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of the storage medium of the computer include, but are not limited to: phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, may be used to store information that may be accessed by a computing device.
The invention also provides a computer device, which comprises a memory, a processor and a computer program stored in the memory and executable by the processor, wherein the processor implements the steps of the surface temperature downscaling method based on the XGBoost learning algorithm according to any one of the above items when executing the computer program.
As shown in fig. 3, which is a schematic diagram of a Sentinel-2A image of a research area, fig. 4 is a schematic diagram of a Landsat 8 thermal infrared image, and fig. 5 is a schematic diagram of a surface temperature extracted by using the surface temperature downscaling method based on the XGBoost learning algorithm according to the present invention, as can be clearly seen from fig. 4-5, the imaging clarity of fig. 5 is significantly improved.
The existing land surface temperature downscaling method only considers empirical regression relations between land surface parameters such as vegetation indexes, DEMs (digital elevation models), albedo and the like and the land surface temperature, neglects the influence of spatial heterogeneity of an urban underlying surface on the land surface temperature, and causes large errors in downscaling results because the influence of different urban impervious surface coverage, vegetation coverage and road density on the land surface temperature is different. According to the method, the research area is divided according to the impervious surface coverage, the high-resolution ground surface prediction temperature of the whole research area is obtained by calculation according to different influence degrees of urban impervious surface coverage, vegetation coverage and road density of different areas on the ground surface temperature and the influence of spatial heterogeneity of an urban underlying surface on the ground surface temperature, so that the high-resolution prediction of the urban environment ground surface temperature with complex spatial heterogeneity is realized, the ground surface temperatures of roads, buildings, vegetation and water bodies can be distinguished more finely, and the accuracy of the urban ground surface temperature prediction is improved.
The present invention is not limited to the above-described embodiments, and various modifications and variations of the present invention are intended to be included within the scope of the claims and the equivalent technology of the present invention if they do not depart from the spirit and scope of the present invention.
Claims (10)
1. An earth surface temperature downscaling method based on an XGboost learning algorithm is characterized by comprising the following steps of:
acquiring surface parameters and thermal infrared image data of a first resolution in a research area; wherein the surface parameters comprise impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density;
dividing the research area into high-density, medium-density and low-density subareas by utilizing a spatial clustering algorithm according to the impervious surface coverage;
aiming at each sub-region, constructing and training a non-linear regression model by taking the surface parameters of the first resolution as input variables and the thermal infrared surface temperature of the first resolution as output variables based on an XGboost algorithm to obtain residual values of the non-linear regression model;
inputting the surface parameters of the second resolution in each sub-area into the nonlinear regression model to obtain the initial predicted value of the thermal infrared surface temperature of the second resolution in each sub-area, wherein the second resolution is higher than the first resolution;
carrying out downscaling on the residual value by using a surface-to-point kriging method to obtain a residual of a second resolution of the research area; and adding the residual error of the second resolution to the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain the predicted value of the thermal infrared earth surface temperature of the second resolution.
2. The earth's surface temperature downscaling method based on the XGBoost learning algorithm of claim 1, wherein the thermal infrared image data is Landsat 8 thermal infrared image data, the first resolution is 30 meters, and acquiring the earth's surface parameters of the first resolution of the research area comprises:
acquiring Sentinel-2A remote sensing image data;
extracting the impervious surface coverage and vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method;
resampling near infrared band and mid infrared band data in the Sentinel-2A remote sensing image data to a resolution of 10 meters;
calculating a normalized vegetation index, a normalized construction index, an improved normalized difference water body index and road density with the resolution of 10 meters according to the resampled near-infrared band and mid-infrared band remote sensing image data;
and performing aggregate calculation on the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density with the resolution of 10 meters to obtain surface parameters with the resolution of 30 meters.
3. The earth surface temperature downscaling method based on the XGboost learning algorithm according to claim 2, wherein the step of extracting the impervious surface coverage and the vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method comprises the following steps:
removing the water body in the remote sensing image data by using the improved normalized difference water body index, selecting a high albedo, a low albedo, a vegetation end member and a soil end member as unmixed objects, and solving the proportion of the high albedo, the low albedo and the vegetation end member in a pixel by using a fully-constrained least square linear spectrum unmixed method to obtain a high albedo coverage, a low albedo coverage and a vegetation coverage;
threshold extraction is carried out on the NDBI image by utilizing an Otsu algorithm, a pixel with a normalized building index larger than a threshold value is used as an impervious surface pixel, a pixel with a normalized building index smaller than the threshold value is used as a non-impervious surface pixel, and the pixel is divided into low albedo coverage of the impervious surface and low albedo coverage of the non-impervious surface by combining the low reflectivity coverage;
further partitioning the low albedo coverage of the non-impervious surface by using a normalized vegetation index: if the normalized vegetation index is less than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of soil, and if the normalized vegetation index is more than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of vegetation;
and adding the high-albedo coverage and the low-albedo coverage of the impervious surface to obtain the impervious surface coverage, and adding the vegetation coverage and the low-albedo coverage of the vegetation to obtain the vegetation coverage.
4. The XGboost learning algorithm-based surface temperature downscaling method of claim 1, wherein the step of constructing and training a nonlinear regression model comprises:
and taking the surface parameter of the first resolution as an input variable, and taking the thermal infrared surface temperature of the first resolution as an output variable, and constructing a nonlinear regression model expression:
lst～imper_frac+veg_frac+road+mndwi+ndbi+ndvi；
wherein lst represents the thermal infrared earth surface temperature, impermeant _ frac represents the impervious surface coverage, veg _ frac represents the vegetation coverage, rod represents the road density, MNDWI represents the MNDWI, ndbi represents the normalized building index, and ndvi represents the normalized vegetation index;
setting training parameters of the nonlinear regression model, including iteration times, contraction step length, minimum weight values of sub-nodes, the number of sub-samples, a minimum loss function reduction value required by node splitting and a maximum node splitting depth, and training the nonlinear regression model.
5. An earth surface temperature downscaling system based on an XGboost learning algorithm is characterized by comprising:
the data acquisition module is used for acquiring surface parameters and thermal infrared image data of a first resolution in a research area; wherein the surface parameters comprise impervious surface coverage, vegetation coverage, normalized vegetation index, normalized building index, improved normalized difference water body index and road density;
the region dividing module is used for dividing the research region into high-density, medium-density and low-density sub-regions by utilizing a spatial clustering algorithm according to the water impermeability coverage;
the model construction module is used for constructing and training a nonlinear regression model by taking the surface parameters of the first resolution as input variables and the thermal infrared surface temperature of the first resolution as output variables on the basis of an XGboost algorithm for each sub-region to obtain residual values of the nonlinear regression model;
the initial value acquisition module is used for inputting the surface parameters of the second resolution in each sub-area into the nonlinear regression model to obtain the initial predicted value of the thermal infrared surface temperature of the second resolution in each sub-area, wherein the second resolution is higher than the first resolution;
the predicted value calculation module is used for carrying out downscaling on the residual value by using a surface-to-point kriging method to obtain a residual error of a second resolution of the research area; and adding the residual error of the second resolution to the initial predicted value of the thermal infrared earth surface temperature in each sub-area to obtain the predicted value of the thermal infrared earth surface temperature of the second resolution.
6. The XGboost learning algorithm-based surface temperature downscaling system according to claim 5, wherein the thermal infrared image data is Landsat 8 thermal infrared image data, the first resolution is 30 meters, and the data acquisition module comprises:
the data acquisition unit is used for acquiring Sentinel-2A remote sensing image data;
the extraction unit is used for extracting the impervious surface coverage and the vegetation coverage with the resolution of 10 meters in the remote sensing image data by using an improved linear spectrum unmixing method;
the resampling unit is used for resampling the near infrared band and mid infrared band data in the Sentinel-2A remote sensing image data to the resolution of 10 meters;
the index calculation unit is used for calculating a normalized vegetation index, a normalized building index, an improved normalized difference water body index and road density with the resolution of 10 meters according to the resampled near-infrared band and mid-infrared band remote sensing image data;
and the aggregation unit is used for carrying out aggregation calculation on the impervious surface coverage, the vegetation coverage, the normalized vegetation index, the normalized building index, the improved normalized difference water body index and the road density of 10 meters to obtain the land surface parameters with the resolution of 30 meters.
7. The XGboost learning algorithm-based surface temperature downscaling system of claim 6, wherein the exponent calculation unit comprises:
the initial coverage calculation unit is used for removing the water body in the remote sensing image data by utilizing the improved normalized difference water body index, selecting a high albedo, a low albedo, a vegetation end member and a soil end member as unmixed objects, and solving the proportion of the high albedo, the low albedo and the vegetation end member in a pixel by utilizing a fully-constrained least square linear spectrum unmixed method to obtain the high albedo coverage, the low albedo coverage and the vegetation coverage;
the impervious surface pixel extraction unit is used for performing threshold extraction on the NDBI image by utilizing an Otsu algorithm, taking the pixel with the normalized building index larger than the threshold as an impervious surface pixel, taking the pixel with the normalized building index smaller than the threshold as a non-impervious surface pixel, and dividing the pixel into impervious surface low-albedo coverage and non-impervious surface low-albedo coverage by combining the low-reflectivity coverage;
a dividing unit for further dividing the low albedo coverage of the non-impervious surface by a normalized vegetation index: if the normalized vegetation index is less than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of soil, and if the normalized vegetation index is more than 0.2, dividing the low albedo coverage of the non-impervious surface into the low albedo coverage of vegetation;
and the coverage acquisition unit is used for adding the high-albedo coverage and the low-albedo coverage of the impervious surface to obtain the impervious surface coverage, and adding the vegetation coverage and the low-albedo coverage of the vegetation to obtain the vegetation coverage.
8. The XGboost learning algorithm-based surface temperature downscaling system of claim 5, wherein the model construction module comprises:
the model expression construction unit is used for constructing a nonlinear regression model expression by taking the surface parameter of the first resolution as an input variable and taking the thermal infrared surface temperature of the first resolution as an output variable:
lst～imper_frac+veg_frac+road+mndwi+ndbi+ndvi；
wherein lst represents the thermal infrared earth surface temperature, impermeant _ frac represents the impervious surface coverage, veg _ frac represents the vegetation coverage, rod represents the road density, MNDWI represents the MNDWI, ndbi represents the normalized building index, and ndvi represents the normalized vegetation index;
and the training unit is used for setting training parameters of the nonlinear regression model, including iteration times, contraction step length, minimum weight values of sub nodes, the number of sub samples, a minimum loss function reduction value required by node splitting and the maximum depth of the node splitting, and training the nonlinear regression model.
9. A computer-readable storage medium having stored thereon a computer program, characterized in that: the computer program when executed by a processor implements the steps of the XGBoost learning algorithm based surface temperature downscaling method of any one of claims 1-4.
10. A computer device, characterized by: comprising a memory, a processor and a computer program stored in the memory and executable by the processor, the processor implementing the steps of the XGBoost learning algorithm based surface temperature downscaling method according to any one of claims 1-4 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911170146.XA CN111199185B (en) | 2019-11-26 | 2019-11-26 | Surface temperature downscaling method, system and equipment based on XGBoost learning algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911170146.XA CN111199185B (en) | 2019-11-26 | 2019-11-26 | Surface temperature downscaling method, system and equipment based on XGBoost learning algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111199185A true CN111199185A (en) | 2020-05-26 |
| CN111199185B CN111199185B (en) | 2023-09-08 |
Family
ID=70746009
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911170146.XA Active CN111199185B (en) | 2019-11-26 | 2019-11-26 | Surface temperature downscaling method, system and equipment based on XGBoost learning algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111199185B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814692A (en) * | 2020-07-10 | 2020-10-23 | 广州地理研究所 | Method for acquiring influence distance of ground object type on ground surface temperature |
| CN111879440A (en) * | 2020-08-06 | 2020-11-03 | 中国科学院西北生态环境资源研究院 | Method and device for calculating surface temperature |
| CN112016052A (en) * | 2020-08-20 | 2020-12-01 | 广东省气象探测数据中心 | Near-surface daily maximum air temperature estimation method, system and terminal based on multi-source data |
| CN112528789A (en) * | 2020-12-02 | 2021-03-19 | 安徽大学 | Remote sensing monitoring method for wheat stripe rust for early and middle stage disease onset analysis |
| CN113221790A (en) * | 2021-05-21 | 2021-08-06 | 广东省科学院广州地理研究所 | Method and device for generating field crop rotation mode based on radar data |
| CN113984212A (en) * | 2021-10-27 | 2022-01-28 | 中国气象科学研究院 | Agricultural irrigated area extraction method and system |
| CN114723166A (en) * | 2022-04-27 | 2022-07-08 | 中国科学院地理科学与资源研究所 | Surface temperature reconstruction method based on XGboost model |
| CN115270638A (en) * | 2022-08-11 | 2022-11-01 | 北华航天工业学院 | Method and system for down-scale time-space analysis and prediction of urban thermal environment |
| CN115859765A (en) * | 2022-09-29 | 2023-03-28 | 中山大学 | Method, device, equipment and storage medium for predicting city expansion |
| CN117148350A (en) * | 2023-10-27 | 2023-12-01 | 航天宏图信息技术股份有限公司 | SAR image-based earth surface temperature space downscaling method, device and equipment |
| CN117218551A (en) * | 2023-08-04 | 2023-12-12 | 华南师范大学 | Estimation algorithm optimization method and device based on error analysis |
| CN117540530A (en) * | 2023-10-10 | 2024-02-09 | 二十一世纪空间技术应用股份有限公司 | Urban earth surface temperature downscaling method and device based on high-resolution satellite images |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104820754A (en) * | 2015-05-13 | 2015-08-05 | 南京信息工程大学 | Space statistical downscaling rainfall estimation method based on geographical difference analysis method |
| CN105095888A (en) * | 2015-09-18 | 2015-11-25 | 中国科学院遥感与数字地球研究所 | Urban main built-up area remote sensing extraction method based on impervious surface aggregation density |
| CN106776481A (en) * | 2016-11-29 | 2017-05-31 | 河海大学 | A kind of NO emissions reduction bearing calibration for acting on satellite precipitation data |
| CN109060133A (en) * | 2018-05-31 | 2018-12-21 | 北京师范大学 | Remote Sensing temperature NO emissions reduction algorithm |
| CN109635309A (en) * | 2018-10-17 | 2019-04-16 | 广州地理研究所 | A kind of surface temperature space NO emissions reduction method |
| US20190228478A1 (en) * | 2016-12-26 | 2019-07-25 | Shanghai Ocean Univeristy | Method for predicting fishing access of katsuwonuspelamis purse seine fishery in the central and western pacific |
-
2019
- 2019-11-26 CN CN201911170146.XA patent/CN111199185B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104820754A (en) * | 2015-05-13 | 2015-08-05 | 南京信息工程大学 | Space statistical downscaling rainfall estimation method based on geographical difference analysis method |
| CN105095888A (en) * | 2015-09-18 | 2015-11-25 | 中国科学院遥感与数字地球研究所 | Urban main built-up area remote sensing extraction method based on impervious surface aggregation density |
| CN106776481A (en) * | 2016-11-29 | 2017-05-31 | 河海大学 | A kind of NO emissions reduction bearing calibration for acting on satellite precipitation data |
| US20190228478A1 (en) * | 2016-12-26 | 2019-07-25 | Shanghai Ocean Univeristy | Method for predicting fishing access of katsuwonuspelamis purse seine fishery in the central and western pacific |
| CN109060133A (en) * | 2018-05-31 | 2018-12-21 | 北京师范大学 | Remote Sensing temperature NO emissions reduction algorithm |
| CN109635309A (en) * | 2018-10-17 | 2019-04-16 | 广州地理研究所 | A kind of surface temperature space NO emissions reduction method |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814692A (en) * | 2020-07-10 | 2020-10-23 | 广州地理研究所 | Method for acquiring influence distance of ground object type on ground surface temperature |
| CN111879440A (en) * | 2020-08-06 | 2020-11-03 | 中国科学院西北生态环境资源研究院 | Method and device for calculating surface temperature |
| CN112016052A (en) * | 2020-08-20 | 2020-12-01 | 广东省气象探测数据中心 | Near-surface daily maximum air temperature estimation method, system and terminal based on multi-source data |
| CN112016052B (en) * | 2020-08-20 | 2021-07-09 | 广东省气象探测数据中心 | Near-surface daily maximum air temperature estimation method, system and terminal based on multi-source data |
| CN112528789A (en) * | 2020-12-02 | 2021-03-19 | 安徽大学 | Remote sensing monitoring method for wheat stripe rust for early and middle stage disease onset analysis |
| CN113221790A (en) * | 2021-05-21 | 2021-08-06 | 广东省科学院广州地理研究所 | Method and device for generating field crop rotation mode based on radar data |
| CN113984212B (en) * | 2021-10-27 | 2023-06-27 | 中国气象科学研究院 | Agricultural irrigation area extraction method and system |
| CN113984212A (en) * | 2021-10-27 | 2022-01-28 | 中国气象科学研究院 | Agricultural irrigated area extraction method and system |
| CN114723166A (en) * | 2022-04-27 | 2022-07-08 | 中国科学院地理科学与资源研究所 | Surface temperature reconstruction method based on XGboost model |
| CN115270638A (en) * | 2022-08-11 | 2022-11-01 | 北华航天工业学院 | Method and system for down-scale time-space analysis and prediction of urban thermal environment |
| CN115270638B (en) * | 2022-08-11 | 2024-02-06 | 北华航天工业学院 | Urban thermal environment downscaling space-time analysis and prediction method and system |
| CN115859765A (en) * | 2022-09-29 | 2023-03-28 | 中山大学 | Method, device, equipment and storage medium for predicting city expansion |
| CN115859765B (en) * | 2022-09-29 | 2023-12-08 | 中山大学 | Urban expansion prediction method, device, equipment and storage medium |
| CN117218551A (en) * | 2023-08-04 | 2023-12-12 | 华南师范大学 | Estimation algorithm optimization method and device based on error analysis |
| CN117218551B (en) * | 2023-08-04 | 2024-04-09 | 华南师范大学 | Estimation algorithm optimization method and device based on error analysis |
| CN117540530A (en) * | 2023-10-10 | 2024-02-09 | 二十一世纪空间技术应用股份有限公司 | Urban earth surface temperature downscaling method and device based on high-resolution satellite images |
| CN117148350A (en) * | 2023-10-27 | 2023-12-01 | 航天宏图信息技术股份有限公司 | SAR image-based earth surface temperature space downscaling method, device and equipment |
| CN117148350B (en) * | 2023-10-27 | 2024-01-30 | 航天宏图信息技术股份有限公司 | SAR image-based earth surface temperature space downscaling method, device and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111199185B (en) | 2023-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111199185A (en) | Ground surface temperature downscaling method, system and equipment based on XGboost learning algorithm | |
| Liu et al. | Annual large-scale urban land mapping based on Landsat time series in Google Earth Engine and OpenStreetMap data: A case study in the middle Yangtze River basin | |
| Liu | Airborne LiDAR for DEM generation: some critical issues | |
| Sharifi | Flood mapping using relevance vector machine and SAR data: A case study from Aqqala, Iran | |
| Verhagen et al. | Object-based landform delineation and classification from DEMs for archaeological predictive mapping | |
| Diaz-Varela et al. | Automatic identification of agricultural terraces through object-oriented analysis of very high resolution DSMs and multispectral imagery obtained from an unmanned aerial vehicle | |
| Ha et al. | A review of downscaling methods for remote sensing-based irrigation management: Part I | |
| Schut et al. | Rapid characterisation of vegetation structure to predict refugia and climate change impacts across a global biodiversity hotspot | |
| Zhang et al. | Assimilating optical satellite remote sensing images and field data to predict surface indicators in the Western US: Assessing error in satellite predictions based on large geographical datasets with the use of machine learning | |
| Silvan-Cardenas et al. | Assessing fine-spatial-resolution remote sensing for small-area population estimation | |
| Lee et al. | Detection of landslides using web-based aerial photographs and landslide susceptibility mapping using geospatial analysis | |
| Abd et al. | Maximum likelihood for land-use/land-cover mapping and change detection using Landsat satellite images: A case study “South of Johor” | |
| Al-Kofahi et al. | Assessment of the urban sprawl on agriculture lands of two major municipalities in Jordan using supervised classification techniques | |
| Weidner et al. | Classifying rock slope materials in photogrammetric point clouds using robust color and geometric features | |
| Gómez‐Gutiérrez et al. | Using visibility analysis to improve point density and processing time of SfM‐MVS techniques for 3D reconstruction of landforms | |
| Shakya et al. | Change over time in grey levels of multispectral landsat 5 TM/8 OLI satellite images | |
| Pan et al. | Remote sensing inversion of soil organic matter by using the subregion method at the field scale | |
| Kolecka et al. | Mapping secondary forest succession on abandoned agricultural land in the Polish Carpathians | |
| Zhang | Multi-scale vegetation-aeolian transport interaction in drylands: remote sensing and modeling | |
| Vahidi et al. | Monitoring land cover changes in Tehran city over 5 years (2018 to 2022) using remote sensing based spatial information | |
| Bigdeli et al. | DTM generation from the point cloud using a progressive geodesic morphology and a modified Particle Swarm Optimization | |
| Cunliffe et al. | Drone-derived canopy height predicts biomass across non-forest ecosystems globally | |
| Jacobson | The effects of endmember selection on modelling impervious surfaces using spectral mixture analysis: a case study in Sydney, Australia | |
| Jaganathan et al. | A localized counter-based filtering approach for generating digital terrain model from medium resolution digital surface model | |
| Aboutalebi et al. | Validation of digital surface models (DSMs) retrieved from unmanned aerial vehicle (UAV) point clouds using geometrical information from shadows |
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 |