CN112836610B - Land use change and carbon reserve quantitative estimation method based on remote sensing data - Google Patents
Land use change and carbon reserve quantitative estimation method based on remote sensing data Download PDFInfo
- Publication number
- CN112836610B CN112836610B CN202110100517.8A CN202110100517A CN112836610B CN 112836610 B CN112836610 B CN 112836610B CN 202110100517 A CN202110100517 A CN 202110100517A CN 112836610 B CN112836610 B CN 112836610B
- Authority
- CN
- China
- Prior art keywords
- carbon
- remote sensing
- land
- image
- reserves
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
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
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention discloses a land use change and carbon reserve quantitative estimation method based on remote sensing data, which comprises the following steps: downloading the image; image preprocessing; land utilization classification; calculating the carbon density of the ground object according to ground survey data; carrying out correlation analysis on the carbon reserves and each characteristic value in various places, and selecting the characteristic value with obvious correlation for modeling; and performing normalization processing on the screened characteristic values to serve as an input layer of a convolutional neural network, putting the carbon density calculated in various ways into a network output layer, performing network training, and performing quantitative estimation on the carbon reserves of the area to be researched by using a trained model. The invention is based on the hierarchical learning architecture of the multi-scale convolutional neural network, so that the land utilization classification result is better. Based on different characteristic values in the image and the carbon density obtained from ground survey data, the nonlinear relation between characteristic variables and carbon reserves is better fitted, and the final regional carbon reserve quantitative estimation result is improved.
Description
Technical Field
The invention relates to the field of remote sensing, in particular to a land use change and carbon reserve quantitative estimation method based on remote sensing data.
Background
Land use changes are the largest uncertainty factor in estimating land ecosystem carbon storage and release. The land utilization change influences the fixation, accumulation and release of ecological system soil and vegetation carbon, further influences the carbon cycle process of the whole ecological system, and changes the original carbon storage and carbon release mode of the ecological system.
The method for accurately predicting the influence of future land use change on regional carbon reserves has important significance in land use decision and city expansion planning. The existing carbon reserve calculation methods mainly comprise a remote sensing image direct estimation method, a vegetation index estimation method, a spectral measurement analysis method, an InVEST model, a CASA model and the like.
Due to the fact that land utilization/coverage types in the region are complex and diverse, most of existing models only consider local features of images and do not consider feature information in different scale ranges, and therefore the land utilization classification effect is poor. The correlation and the determination coefficient of the model result calculated based on various carbon reserves are relatively low, and an overfitting phenomenon easily exists.
The high-resolution remote sensing image improves the image quality and simultaneously causes the land use classification to need multi-dimensional characteristic information. On one hand, the land use type is a combination of different land features, and the land use type needs to be acquired in a larger scale range in order to consider the spatial distribution characteristics of the image. For such a complex scene, more high-level semantic features and underlying spatial information need to be considered. On the other hand, the texture shape difference is large for different ground object sizes, and the characteristic information in different scale ranges is needed.
Although the conventional regression model for regional carbon reserve research in China is simple to operate, the co-linearity problem easily exists among variables.
Disclosure of Invention
In order to overcome the defects of the technology, the invention provides a land use change and carbon reserve quantitative estimation method based on remote sensing data.
In order to solve the technical problems, the invention adopts the technical scheme that: a land use change and carbon reserve quantitative estimation method based on remote sensing data comprises the following steps:
step one, downloading an image: downloading remote sensing images of a designated research area over the years according to research requirements, and acquiring a labeled land classification data set;
step two, image preprocessing: carrying out geometric correction, inlaying, cutting, radiometric calibration and atmospheric correction preprocessing operations on the remote sensing image;
step three, land utilization classification: an end-to-end convolutional neural network is adopted, the same area picture is respectively expanded into three different scales of nxn, 2 nx2 n and 3 nx3 n, and nxn blocks are taken as the network from the same area picture and input into the three convolutional neural networks; in order to reduce the parameters during training, the parameters are shared among the three convolutional neural networks. Extracting high-level semantic features from each convolutional neural network through a convolutional layer and a pooling layer, combining all generated feature maps through the convolutional layer, and finally restoring the sizes of the feature maps after convolution to the size of n multiplied by n of the original feature maps, namely, a classification result; taking 80% of the land classification data set as a training set, and taking the other 20% as a testing set;
step four, calculating the carbon density of the ground objects according to the ground survey data: determining a typical sampling area in the research area according to the land use type classification result of the research area, sampling and surveying a plurality of sample plots in the sampling area range, and calculating sample plot carbon reserves and average carbon reserves in unit area, namely carbon density, through data acquired and recorded by ground sample plot survey; dividing the carbon reserves into soil carbon reserves and vegetation carbon reserves according to different land types;
step five, reading the image through a GDAL function library by using the downloaded high-resolution remote sensing image based on python language, acquiring a plurality of characteristic values of the image in a research area, extracting texture characteristics by using a gray level co-occurrence matrix method, carrying out correlation analysis on the carbon reserves and each characteristic value in each sample land, and selecting characteristic values with obvious correlation for modeling;
and step six, performing normalization processing on the characteristic values screened out in the step five to serve as an input layer of the convolutional neural network, putting the carbon density calculated in various ways into a network output layer, performing network training, and performing quantitative carbon reserve estimation on the area to be researched by using the trained model.
Further, in the third step, the convolved feature map is subjected to bilinear interpolation upsampling to restore the size.
Further, the carbon reserve of the soil in the fourth step is measured by a potassium dichromate external heating method, and the formula is the content of organic carbon in the soil:
in the formula: c is 0.8000mol/L potassium dichromate solution; v0Blank titration with ferrous sulfate volume; v is the volume of ferrous sulfate for sample titration; m is the mass of the blank soil sample; and k is the coefficient of converting the air-dried soil into the dried soil.
Further, the estimation of the carbon reserve of the vegetation adopts a biomass spreading factor method:
C=(44/12)V*N*WD*BEF*(1+R)*CF。
wherein: c is fixed CO2V is the volume of the individual plant, N is the number of plants, WD is the wood density of the tree species, BEF is the biomass expansion factor for the conversion of the trunk biomass of the tree species to the above-ground biomass, CF is the average carbon content of the tree species, R is the biomass to root ratio of the tree species, 44/12 is CO2Molecular weight ratio to C.
In order to better quantify local features and global features of different ground features, the multi-scale semantic features are learned from multi-scale input images based on a hierarchical learning framework of a multi-scale convolutional neural network, so that a land utilization classification result is better. Based on different characteristic values in the image and the carbon density obtained from ground survey data, a convolutional neural network is used for modeling, and nonlinear relations between characteristic variables and carbon reserves are better fitted. And improving the quantitative estimation result of the carbon reserves in the final area.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a diagram of a convolutional neural network structure at multiple scales.
Detailed Description
The invention is described in further detail below with reference to the drawings and the detailed description.
The land use change and carbon reserve quantitative estimation method based on remote sensing data as shown in figure 1 comprises the following steps:
step one, downloading an image: and downloading the remote sensing images of the designated research area over the years according to the research requirements. Acquiring a labeled land classification data set;
step two, image preprocessing: carrying out preprocessing operations such as geometric correction, embedding, cutting, radiometric calibration, atmospheric correction and the like on the remote sensing image;
and (3) geometric correction: and (3) correcting and eliminating geometric errors of the remote sensing image by using a series of mathematical models.
Radiation calibration: the value or voltage recorded by the sensor is converted into absolute radiance.
Atmospheric correction: the apparent reflectivity is converted into the actual reflectivity of the earth surface, and errors caused by atmospheric scattering, hand washing and reflection are eliminated.
Step three, land utilization classification:
with an end-to-end convolutional neural network, as shown in fig. 2, three patches with different scales (in this embodiment, 30 × 30, 60 × 60, and 90 × 90 input pictures, and a 30 × 30 block taken from the pictures as a network input) are input into the three convolutional neural networks in parallel, and the network structure is as shown in fig. 2. To reduce the parameters during training, we share the parameters between the three convolutional neural networks. Each convolutional neural network extracts context information under high-level semantic features and different scales through a convolutional layer and a pooling layer.
The pooling operation is another important component of the convolutional neural network, which is a down-sampling operation in both the length and width dimensions, typically after the convolution operation. After the convolution layer is input into the image or the feature map, a group of feature maps with the same length and width as the input feature map is generated, and with the increase of the number of channels, if the feature maps are not subjected to length and width dimension down-sampling, the generated feature map parameters are increased sharply, so that the calculation overhead is not borne, and meanwhile, the feature with larger scale is difficult to generate. The pooling operation is performed by scanning the eigen map tensor using a matrix window, taking the maximum or average value in each matrix as a result to reduce the number of elements, selecting the maximum value for pooling is called maximum pooling, and selecting the average value for pooling is called average pooling. The pooling operation enables the feature map to be reduced in dimension in length and width, and the channel dimension can be increased in dimension, so that richer feature information can be obtained, and the whole network model has certain space invariance.
And combining all the generated feature maps through convolution layers, and finally restoring the dimension of the convolved feature maps to the original feature map size (30 multiplied by 30) through a bilinear interpolation upsampling mode, namely a classification result.
80% of the land classification data set is used as a training set, and the other 20% is used as a testing set.
Step four, calculating the carbon density of the ground objects according to the ground survey data: and determining a typical sampling area in the research area according to the land use type classification result of the research area. Sampling investigation is carried out by taking a plurality of sample areas in the sampling area range by using a system sampling method. The plot carbon reserve calculation is data (plot number, land use type, etc.) of records obtained by a ground plot survey. The carbon reserves in the respective plots and the average unit area, that is, the carbon density were calculated.
The carbon reserve of the soil is determined by a potassium dichromate external heating method, and the formula is the organic carbon content of the soil:
in the formula: c is 0.8000mol/L potassium dichromate solution; v0Volume of ferrous sulfate (ml) for blank titration; v is the volume of ferrous sulfate (ml) used for titration of the sample; m is the mass (g) of the blank soil sample; and k is the coefficient of converting the air-dried soil into the dried soil.
The estimation of the carbon reserve of the vegetation adopts a biomass spreading factor method, and the formula is as follows:
C=(44/12)V*N*WD*BEF*(1+R)*CF。
wherein: c is fixed CO2The mass of (c), V is the volume of the individual plant (m)3Strain), N is the number of strains, WD is the wood density (dry weight, t m) of the tree species-3) BEF is a biomass expansion factor for converting the biomass of the tree trunk to the above-ground biomass, CF is the average carbon content of the tree, R is the biomass-to-rhizome ratio of the tree (i.e., the ratio of the underground biomass to the above-ground biomass, without units), 44/12 is CO2Molecular weight ratio to C.
And fifthly, reading the image through a GDAL function library by using the downloaded high-resolution remote sensing image based on a python language, acquiring a plurality of characteristic values of the image in the research area, such as image band data, a band ratio, a normalized vegetation index NDVI, a ratio vegetation index RVI, an enhanced vegetation index EVI and a humidity vegetation index WVI, and extracting texture characteristics (mean value, variance, entropy, contrast and uniformity) by using a gray level co-occurrence matrix method. Performing correlation analysis on the carbon reserves in various places and each characteristic value through SPSS analysis data, and selecting the characteristic value with obvious correlation for modeling;
step six, dividing the research area into 86 sample areas, setting the training samples to be 69 and setting the verification samples to be 17. And (4) performing normalization processing on the characteristic values screened out in the step five to serve as an input layer of the convolutional neural network, putting the carbon density calculated in various ways into a network output layer, and performing network training. And (5) carrying out quantitative estimation on the carbon reserves of the area to be researched by using the trained model.
The above embodiments are not intended to limit the present invention, and the present invention is not limited to the above examples, and those skilled in the art may make variations, modifications, additions or substitutions within the technical scope of the present invention.
Claims (4)
1. A land use change and carbon reserve quantitative estimation method based on remote sensing data is characterized by comprising the following steps: the method comprises the following steps:
step one, downloading an image: downloading remote sensing images of a designated research area over the years according to research requirements, and acquiring a labeled land classification data set;
step two, image preprocessing: carrying out geometric correction, inlaying, cutting, radiometric calibration and atmospheric correction preprocessing operations on the remote sensing image;
step three, land utilization classification: an end-to-end convolutional neural network is adopted, the same area picture is respectively expanded into three different scales of nxn, 2 nx2 n and 3 nx3 n, and nxn blocks are taken as the network from the same area picture and input into the three convolutional neural networks; in order to reduce parameters during training, parameters are shared among the three convolutional neural networks; extracting high-level semantic features and context information with different scales from each convolutional neural network through a convolutional layer and a pooling layer, combining all generated feature maps through the convolutional layers, and finally restoring the sizes of the feature maps after convolution to the size of n multiplied by n of the original feature maps, namely a classification result; 80% of the land classification data set is used as a training set, and the other 20% is used as a testing set;
step four, calculating the carbon density of the ground objects according to the ground survey data: determining a typical sampling area in the research area according to the land use type classification result of the research area, sampling and surveying a plurality of sample plots in the sampling area range, and calculating sample plot carbon reserves and average carbon reserves in unit area, namely carbon density, through data acquired and recorded by ground sample plot survey; dividing the carbon reserves into soil carbon reserves and vegetation carbon reserves according to different land types;
step five, reading the image through a GDAL function library by using the downloaded high-resolution remote sensing image based on python language, acquiring a plurality of characteristic values of the image in a research area, extracting texture characteristics by using a gray level co-occurrence matrix method, carrying out correlation analysis on the carbon reserves and each characteristic value in each sample land, and selecting characteristic values with obvious correlation for modeling;
and step six, performing normalization processing on the characteristic values screened out in the step five to serve as an input layer of the convolutional neural network, putting the carbon density calculated in various ways into a network output layer, performing network training, and performing quantitative carbon reserve estimation on the area to be researched by using the trained model.
2. The remote sensing data-based land use change and carbon reserve quantitative estimation method according to claim 1, characterized in that: and in the third step, recovering the size of the convolved characteristic diagram in a bilinear interpolation upsampling mode.
3. The remote sensing data-based land use change and carbon reserve quantitative estimation method according to claim 1, characterized in that: and in the fourth step, the carbon reserve of the soil is determined by a potassium dichromate external heating method, and the formula is the organic carbon content of the soil:
in the formula: c is 0.8000mol/L potassium dichromate solution; v0Blank titration with ferrous sulfate volume; v is the volume of ferrous sulfate for sample titration; m is the mass of the blank soil sample; and k is the coefficient of converting the air-dried soil into the dried soil.
4. The remote sensing data-based land use change and carbon reserve quantitative estimation method according to claim 3, characterized in that: the estimation of the carbon reserves of the vegetation adopts a biomass spreading factor method:
C=(44/12)V*N*WD*BEF*(1+R)*CF;
wherein: c is fixed CO2V is the volume of the individual plant, N is the number of plants, WD is the wood density of the tree species, BEF is the biomass expansion factor for the conversion of the trunk biomass of the tree species to the above-ground biomass, CF is the average carbon content of the tree species, R is the biomass to root ratio of the tree species, 44/12 is CO2Molecular weight ratio to C.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110100517.8A CN112836610B (en) | 2021-01-26 | 2021-01-26 | Land use change and carbon reserve quantitative estimation method based on remote sensing data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110100517.8A CN112836610B (en) | 2021-01-26 | 2021-01-26 | Land use change and carbon reserve quantitative estimation method based on remote sensing data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112836610A CN112836610A (en) | 2021-05-25 |
CN112836610B true CN112836610B (en) | 2022-05-27 |
Family
ID=75931579
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110100517.8A Active CN112836610B (en) | 2021-01-26 | 2021-01-26 | Land use change and carbon reserve quantitative estimation method based on remote sensing data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112836610B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113177744B (en) * | 2021-06-09 | 2024-03-01 | 西安建筑科技大学 | Urban green land system carbon sink estimation method and system |
CN113449976A (en) * | 2021-06-21 | 2021-09-28 | 广东翁源滃江源国家湿地公园管理处 | Forestry carbon metering method based on ecological process model |
CN114021371B (en) * | 2021-11-16 | 2023-03-03 | 中国科学院西北生态环境资源研究院 | Carbon reserve influence estimation method and device, electronic equipment and storage medium |
CN114648705B (en) * | 2022-03-28 | 2022-11-22 | 王大成 | Carbon sink monitoring system and method based on satellite remote sensing |
CN114740180A (en) * | 2022-04-07 | 2022-07-12 | 中山大学 | Soil organic carbon estimation method and device based on multi-source remote sensing data |
CN114529838B (en) * | 2022-04-24 | 2022-07-15 | 江西农业大学 | Soil nitrogen content inversion model construction method and system based on convolutional neural network |
CN114896561B (en) * | 2022-05-07 | 2023-06-16 | 安徽农业大学 | Wetland carbon reserve calculation method based on remote sensing algorithm |
CN114819737B (en) * | 2022-05-26 | 2023-10-17 | 中交第二公路勘察设计研究院有限公司 | Method, system and storage medium for estimating carbon reserves of highway road vegetation |
CN114969665B (en) * | 2022-06-02 | 2023-02-10 | 中国地质科学院矿产资源研究所 | Method and system for estimating carbon emission and carbon reserve of mineral resource base |
CN114937029B (en) * | 2022-06-21 | 2023-01-31 | 西南林业大学 | Forest carbon storage amount sampling estimation method, device, equipment and storage medium |
CN115730833B (en) * | 2022-10-31 | 2024-03-12 | 湖北省规划设计研究总院有限责任公司 | Land ecological system carbon reserve estimation method based on InVEST model |
CN115796712B (en) * | 2023-02-07 | 2023-04-18 | 北京师范大学 | Regional land ecosystem carbon reserve estimation method and device and electronic equipment |
CN116563718B (en) * | 2023-07-11 | 2023-09-05 | 成都垣景科技有限公司 | Remote sensing mapping-based carbon reserve estimation method |
CN117313959A (en) * | 2023-11-28 | 2023-12-29 | 吉林省林业科学研究院(吉林省林业生物防治中心站) | Forestry carbon sink monitoring method and system based on big data |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106384171A (en) * | 2016-09-29 | 2017-02-08 | 四川师范大学 | Fir forest carbon reserve estimation model suitable for local research area |
CN111428781A (en) * | 2020-03-20 | 2020-07-17 | 中国科学院深圳先进技术研究院 | Remote sensing image ground object classification method and system |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA2709377A1 (en) * | 2007-12-20 | 2009-07-09 | Imagetree Corp. | Remote sensing and probabilistic sampling based method for determining the carbon dioxide volume of a forest |
US20120287273A1 (en) * | 2011-05-09 | 2012-11-15 | Abengoa Bioenergia Nuevas Tecnologias, S.A. | System for identifying sustainable geographical areas by remote sensing techniques and method thereof |
US20150371161A1 (en) * | 2013-01-30 | 2015-12-24 | The Board Of Trustees Of The University Of Illinois | System and methods for identifying, evaluating and predicting land use and agricultural production |
US9946931B2 (en) * | 2015-04-20 | 2018-04-17 | Los Alamos National Security, Llc | Change detection and change monitoring of natural and man-made features in multispectral and hyperspectral satellite imagery |
CN104361338B (en) * | 2014-10-17 | 2017-11-28 | 中国科学院东北地理与农业生态研究所 | A kind of peat bog information extracting method based on ENVISAT ASAR, Landsat TM and dem data |
FR3069940B1 (en) * | 2017-08-03 | 2019-09-06 | Universite D'orleans | METHOD AND SYSTEM FOR CARTOGRAPHY OF THE HEALTH CONDITION OF CULTURES |
CN108647623A (en) * | 2018-05-04 | 2018-10-12 | 中国科学院遥感与数字地球研究所 | A kind of potential organic C storage remote sensing estimation method of forest based on resource constraint |
CN112183909A (en) * | 2019-07-01 | 2021-01-05 | 苏州五蕴明泰科技有限公司 | Method and apparatus for determining carbon change for forestry and land use |
CN110750904B (en) * | 2019-10-22 | 2021-07-23 | 南京信大气象科学技术研究院有限公司 | Regional carbon reserve space pattern monitoring system and method based on remote sensing data |
CN111488902A (en) * | 2020-01-14 | 2020-08-04 | 沈阳农业大学 | Method and system for quantitatively estimating carbon reserves of ecosystem of primary coastal wetland |
CN111289725B (en) * | 2020-03-17 | 2022-06-03 | 江苏农林职业技术学院 | Farmland soil organic carbon reserve estimation method and system combining model and time sequence sampling |
CN111553289A (en) * | 2020-04-29 | 2020-08-18 | 中国科学院空天信息创新研究院 | Remote sensing image cloud detection method and system |
-
2021
- 2021-01-26 CN CN202110100517.8A patent/CN112836610B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106384171A (en) * | 2016-09-29 | 2017-02-08 | 四川师范大学 | Fir forest carbon reserve estimation model suitable for local research area |
CN111428781A (en) * | 2020-03-20 | 2020-07-17 | 中国科学院深圳先进技术研究院 | Remote sensing image ground object classification method and system |
Also Published As
Publication number | Publication date |
---|---|
CN112836610A (en) | 2021-05-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112836610B (en) | Land use change and carbon reserve quantitative estimation method based on remote sensing data | |
CN113128134B (en) | Mining area ecological environment evolution driving factor weight quantitative analysis method | |
CN112070078B (en) | Deep learning-based land utilization classification method and system | |
CN107230197B (en) | Tropical cyclone objective strength determination method based on satellite cloud image and RVM | |
Biard et al. | Automated detection of weather fronts using a deep learning neural network | |
CN113935249B (en) | Upper-layer ocean thermal structure inversion method based on compression and excitation network | |
Liang et al. | Maximum likelihood classification of soil remote sensing image based on deep learning | |
CN112001293A (en) | Remote sensing image ground object classification method combining multi-scale information and coding and decoding network | |
CN114005048A (en) | Multi-temporal data-based land cover change and thermal environment influence research method | |
CN114937173A (en) | Hyperspectral image rapid classification method based on dynamic graph convolution network | |
US11379742B2 (en) | Method for predictive soil mapping based on solar radiation in large flat area | |
CN115205703A (en) | Multi-feature blue-green algae extraction method and device, electronic equipment and storage medium | |
Shang et al. | Spatiotemporal reflectance fusion using a generative adversarial network | |
CN116863341B (en) | Crop classification and identification method and system based on time sequence satellite remote sensing image | |
CN107358625B (en) | SAR image change detection method based on SPP Net and region-of-interest detection | |
doninck et al. | Multispectral canopy reflectance improves spatial distribution models of Amazonian understory species | |
Jadhav et al. | Segmentation analysis using particle swarm optimization-self organizing map algorithm and classification of remote sensing data for agriculture | |
CN115797501A (en) | Time-series forest age mapping method combining forest disturbance and recovery events | |
He et al. | Bayesian temporal tensor factorization-based interpolation for time-series remote sensing data with large-area missing observations | |
Zhou et al. | Superpixel-based time-series reconstruction for optical images incorporating SAR data using autoencoder networks | |
CN114612315A (en) | High-resolution image missing region reconstruction method based on multi-task learning | |
CN114419463A (en) | Cloud platform-based global solar photovoltaic panel remote sensing automatic identification method | |
CN111325384B (en) | NDVI prediction method combining statistical characteristics and convolutional neural network model | |
CN113077458A (en) | Cloud and shadow detection method and system in remote sensing image | |
Jing et al. | A Rigorously-Incremental Spatiotemporal Data Fusion Method for Fusing Remote Sensing Images |
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 |