CN110287455A - A kind of PM2.5 deep learning inversion method of combination remotely-sensed data and social perception data - Google Patents
A kind of PM2.5 deep learning inversion method of combination remotely-sensed data and social perception data Download PDFInfo
- Publication number
- CN110287455A CN110287455A CN201910451339.6A CN201910451339A CN110287455A CN 110287455 A CN110287455 A CN 110287455A CN 201910451339 A CN201910451339 A CN 201910451339A CN 110287455 A CN110287455 A CN 110287455A
- Authority
- CN
- China
- Prior art keywords
- data
- variable
- deep learning
- remotely
- social perception
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000008566 social perception Effects 0.000 title claims abstract description 27
- 238000013135 deep learning Methods 0.000 title claims abstract description 13
- 238000004458 analytical method Methods 0.000 claims abstract description 14
- 238000009826 distribution Methods 0.000 claims abstract description 10
- 238000013136 deep learning model Methods 0.000 claims abstract description 7
- 230000002123 temporal effect Effects 0.000 claims abstract description 5
- 238000010606 normalization Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims abstract description 4
- 238000013480 data collection Methods 0.000 claims abstract description 3
- 239000000284 extract Substances 0.000 claims abstract description 3
- 239000000443 aerosol Substances 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 230000003287 optical effect Effects 0.000 claims description 6
- 238000001556 precipitation Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000012952 Resampling Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008447 perception Effects 0.000 claims description 3
- FNMKZDDKPDBYJM-UHFFFAOYSA-N 3-(1,3-benzodioxol-5-yl)-7-(3-methylbut-2-enoxy)chromen-4-one Chemical compound C1=C2OCOC2=CC(C2=COC=3C(C2=O)=CC=C(C=3)OCC=C(C)C)=C1 FNMKZDDKPDBYJM-UHFFFAOYSA-N 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 6
- 230000007812 deficiency Effects 0.000 abstract description 4
- 230000008901 benefit Effects 0.000 description 4
- 241001269238 Data Species 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 230000035935 pregnancy Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N15/06—Investigating concentration of particle suspensions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Dispersion Chemistry (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Algebra (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Operations Research (AREA)
- Analytical Chemistry (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biochemistry (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses the PM2.5 deep learning inversion method of a kind of combination remotely-sensed data and social perception data, comprising: pre-processes to ground station PM2.5 data, remotely-sensed data, social perception data and auxiliary data;Characteristic variable is carried out to multi-source data and extracts and calculates using ground spatial statistics and analysis method, remote sensing information process means;Time-space registration is carried out to multi-source data using grid mode, has the grid of ground station true value that will generate the multi-source data collection of space-time uniformity as training sample;It is trained being input in deep learning model after the Grid square collection for having website PM2.5 true value normalization, inverting is carried out to unknown grid PM2.5 concentration by the model after being verified;Fine PM2.5 spatial and temporal distributions drawing is carried out to inversion result.The present invention can effectively excavate multi-source information using depth learning technology, compensate for deficiency of the conventional statistics model in nonlinear problem, obtain higher inversion accuracy and more fine space-time PM2.5 distribution.
Description
Technical field
The invention belongs to remote sensing image processings and Information application field, are related to a kind of method for obtaining PM2.5 concentration, specifically
It is related to a kind of method based on deep learning combination remotely-sensed data and the fine space-time PM2.5 concentration of social perception data inverting.
Background technique
Fine space-time PM2.5 concentration distribution plays a significant role in environmental monitoring, health evaluating application.Existing website point
The sparse unevenness of cloth, website observation are unable to satisfy application demand, obtain PM2.5 space and time continuous distributed data and are widely noticed.PM2.5 at
Quickly because of complicated, variation, concentration is influenced by natural, artificial multi-party factor simultaneously, and is existed between each factor and PM2.5
Complicated non-linear relation.Therefore, how multi-resources Heterogeneous data, the abundant fine inverting PM2.5 technology of mined information effectively to be combined
Urgently explore.
PM2.5 retrieving concentration method mainly includes physical-chemical structure simulation and statistical model method at present, the former relies on
The input of more model parameter is realized complicated;And the technology of statistical model method estimation PM2.5 concentration is high by its precision, easy
The advantage of realization is widely applied.It is insufficient that existing related art method is primarily present two aspects: on the one hand consider that influence factor is insufficient,
More solely using satellite remote sensing date or only with social perception data, the application of both rare combinations;On the other hand,
When the case where conventional statistics model explains deficiency to non-linear relation, especially combines in face of multi-resources Heterogeneous big data, shallow Model
Feature extraction and information excavating scarce capacity.
Summary of the invention
It is an object of the present invention to be directed to the above-mentioned deficiency of the prior art, a kind of combination remotely-sensed data and social feeling are provided
The method of the PM2.5 deep learning inverting of primary data.
The technical scheme adopted by the invention is that: a kind of PM2.5 depth of combination remotely-sensed data and social perception data
Practise inversion method, comprising the following steps:
Step 1, ground station PM2.5 data, remotely-sensed data, social perception data and auxiliary data are pre-processed;
Step 2, feature change is carried out to multi-source data using ground spatial statistics and analysis method, remote sensing information process means
Amount is extracted and is calculated, and specific implementation includes following sub-step:
Step 2.1, for ground station PM2.5 data, it is dense that temporal-spatial interpolating initial fields are calculated according to space-time autocorrelation performance
Angle value obtains PMS, PMT;
Step 2.2, for remotely-sensed data and auxiliary data, choose aerosol optical depth, vegetation index, temperature, wind speed,
Relative humidity, air pressure, precipitation, atmospheric boundary layer height parameter by image projecting conversion, resampling, cut process, obtain phase
Aerosol optical depth variables A OD, vegetation index variable NDVI, temperature variables Temp, the wind speed variable WS, relative humidity change answered
Measure RH, pressure variation PS, precipitation variable Pre, atmospheric boundary layer height variable PBLH;
Step 2.3, for social perception data, spatial statistics are learned with taking and analysis method extraction is relevant to PM2.5
Characteristic variable, obtain traffic index variable Tindex, real-time population register quantitative variation RTLD, point of interest number variable POIs,
Road mileage variable R oad;
Step 3, time-space registration is carried out to multi-source data using grid mode, has the grid of ground station true value by conduct
Training sample generates the multi-source data collection of space-time uniformity;
Step 4, it will be input in deep learning model and instruct after the Grid square collection for having website PM2.5 true value normalization
Practice, inverting is carried out to unknown grid PM2.5 concentration by the model after being verified;
Step 5, fine PM2.5 spatial and temporal distributions are carried out to inversion result to chart.
Further, it is pre-processed described in step 1, including rejects exceptional value in website PM2.5 data;To remote sensing image number
Unified tiff image file format is converted into according to format;Format is carried out to social perception data using Python, ArcGIS to turn
It changes, after data screening, then carries out geographical coordinate conversion, vectoring operations.
Further, the calculation of PMS and PMT is as follows in step 2.1,
Wherein n, m are respectively spatial neighbor station number, time interval number, and w is weight, d, Δ t be respectively space away from
From, time interval.
Further, cuclear density point is carried out to original traffic exponent data by ArcGIS batch processing tool in step 2.3
Analysis obtains traffic index variable Tindex, road mileage variable R oad;Real-time population is obtained by spatial interpolation methods to register number
Quantitative change amount RTLD;Point of interest number variable POIs is extracted by buffer zone analysis.
Further, the data input of deep learning model described in step 4 uses maximin normalized, mould
Type frame are as follows:
PM2.5=f (PM2.5 space-time initial fields, remote sensing variable, society's perception variable, auxiliary data variable) (3).
Further, auxiliary data described in step 1 is meteorological data.
The present invention has the advantages that
(1) innovation proposes to combine remotely-sensed data and social perception data inverting PM2.5, has fully considered natural, society's warp
Ji factor;
(2) multi-source heterogeneous data are overcome and are difficult to matching problem, use the matched process flow of space-time gridization, are simplified
Operation;
(3) multi-source information can be effectively excavated using depth learning technology, compensates for conventional statistics model in nonlinear problem
In deficiency, obtain higher inversion accuracy and more fine space-time PM2.5 distribution.
In short, method proposed by the present invention can effectively combine remotely-sensed data, social perception data and other auxiliary datas, obtain
More accurate inversion result is obtained, realizes the fine space-time monitoring of PM2.5.
Detailed description of the invention
Fig. 1 is the method flow diagram of the embodiment of the present invention.
Specific embodiment
Understand for the ease of those of ordinary skill in the art and implement the present invention, with reference to the accompanying drawings and embodiments to this hair
It is bright to be described in further detail, it should be understood that implementation example described herein is merely to illustrate and explain the present invention, not
For limiting the present invention.
Influence PM2.5 concentration factor is numerous and relationship is complicated, and the PM2.5 of the fine spatial and temporal distributions of inverting is caused to face the challenge.
Using the natural characteristic of remotely-sensed data and social economy's attribute of social perception data, sufficiently excavated based on deep learning useful
Information, to realize the inverting of fine space-time PM2.5 distribution.
Referring to Fig.1, the PM2.5 deep learning inverting of a kind of combination remotely-sensed data provided by the invention and social perception data
Method, comprising the following steps:
Step 1: to ground station PM2.5 data, remotely-sensed data, social perception data and other auxiliary datas (such as meteorology
Data) it is pre-processed.It follows that highly reliable, relevance is big and obtainable condition early period, collects multi-source data extensively;For
Different types of data take corresponding pretreatment, as rejected exceptional value in website PM2.5 data;Remote sensing image data format turns
Change unified tiff image file format into;Social perception data progress JSON resolves to the conversion of TXT text formatting, removal lacks
The data screening process of few attribute value, error value and outlier carries out the operation such as geographical coordinate conversion, vector quantization later, is convenient for
Subsequent analysis calculates, and handling implement and software are Python, ArcGIS etc.;Meteorological data uses assimilation data,
Processing of the processing mode with remote sensing image data.
Step 2: learning spatial statistics using ground and analysis method, remote sensing information process means carry out feature change to multi-source data
Amount is extracted and is calculated, and specific implementation includes following sub-step:
Step 2.1: for ground station PM2.5 data, it is dense to calculate temporal-spatial interpolating initial fields according to space-time autocorrelation performance
Angle value, the present invention calculate separately to obtain space PM2.5 initial concentration field PMS and time PM2.5 initially dense using inverse distance-weighting
Field PMT is spent, calculation method is shown in formula (1), (2), and wherein n, m are respectively spatial neighbor station number, time interval number, and w is
Weight, d, Δ t are respectively space length, time interval.
Step 2.2: for remotely-sensed data and other auxiliary datas, choose aerosol optical depth, vegetation index, temperature,
The parameters such as wind speed, relative humidity, air pressure, precipitation, atmospheric boundary layer height carry out shadow for the application scenarios in practical study region
As projection transform unifies geographic coordinate system and projected coordinate system, resampling unified resolution, the volume for being cropped to survey region range
Journey batch operation obtains corresponding aerosol optical depth variables A OD, vegetation index variable NDVI, temperature variables Temp, wind
Fast variable WS, relative humidity variable R H, pressure variation PS, precipitation variable Pre, atmospheric boundary layer height variable PBLH etc.;
Step 2.3: learning spatial statistics for social perception data with taking and analysis method extracts spy relevant to PM2.5
Sign variable: it analyzes to obtain traffic index variable Tindex, road mileage change by carrying out cuclear density to original traffic exponent data
Measure Road;Real-time population is obtained by spatial interpolation methods to register quantitative variation RTLD;Point of interest is extracted by buffer zone analysis
Number variable POIs, the above analysis and calculating are realized by ArcGIS batch processing tool;
Step 3: time-space registration is carried out to multi-source data using grid mode, all data are unified for grid, into
Row single-frame net matching has the grid of ground station true value that will generate the multi-source data of space-time uniformity as training sample, the step
Collection, implementation are ArcGIS secondary development;
Step 4: being instructed being input in deep learning model after the Grid square collection for having website PM2.5 true value normalization
Practice, inverting is carried out to grid PM2.5 concentration to be evaluated by the model after being verified.Model data input uses minimax
It is worth normalized, eliminates dimension to acceleration model convergence rate;Using deepness belief network model.Deep learning model frame
Frame are as follows:
PM2.5=f (PM2.5 space-time initial fields, remote sensing variable, society's perception variable, auxiliary data variable) (3);
Step 5: fine PM2.5 distribution drawing being carried out to inversion result and is monitored to space-time.
The present invention can get, under the background of artificial intelligence technology rapid development in multi-source data, it is contemplated that atmosphere PM2.5 is dirty
Dye is to ecological environment and the multiple harm of human health bring, and innovation combines remotely-sensed data and social perception data, with ground
Spatial statistics and analysis method implement space-time grid process flow, and by depth learning technology, abundant digging utilization multi-source is different
The rule and connection of the change in time and space of prime number evidence, using ground station monitoring data as valuable true value sample, inverting obtains essence
Thin PM2.5 spatial and temporal distributions.This method has the accurate advantage of strong operability, inversion accuracy, and not with big data
Stopping pregnancy is raw and accumulation, this method have broader exploration and application value, be readily put into practical.
It should be understood that the part that this specification does not elaborate belongs to the prior art.
It should be understood that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered to this
The limitation of invention patent protection range, those skilled in the art under the inspiration of the present invention, are not departing from power of the present invention
Benefit requires to make replacement or deformation under protected ambit, fall within the scope of protection of the present invention, this hair
It is bright range is claimed to be determined by the appended claims.
Claims (6)
1. the PM2.5 deep learning inversion method of a kind of combination remotely-sensed data and social perception data, which is characterized in that including with
Lower step:
Step 1, ground station PM2.5 data, remotely-sensed data, social perception data and auxiliary data are pre-processed;
Step 2, characteristic variable is carried out to multi-source data and is mentioned using ground spatial statistics and analysis method, remote sensing information process means
It takes and calculates, specific implementation includes following sub-step:
Step 2.1, for ground station PM2.5 data, temporal-spatial interpolating initial fields concentration value is calculated according to space-time autocorrelation performance,
Obtain PMS, PMT;
Step 2.2, for remotely-sensed data and auxiliary data, aerosol optical depth, vegetation index, temperature, wind speed, opposite is chosen
Humidity, air pressure, precipitation, atmospheric boundary layer height parameter by image projecting conversion, resampling, cut process, obtain corresponding
Aerosol optical depth variables A OD, vegetation index variable NDVI, temperature variables Temp, wind speed variable WS, relative humidity variable
RH, pressure variation PS, precipitation variable Pre, atmospheric boundary layer height variable PBLH;
Step 2.3, for social perception data, spatial statistics are learned with taking and analysis method extracts feature relevant to PM2.5
Variable obtains traffic index variable Tindex, real-time population is registered quantitative variation RTLD, point of interest number variable POIs, road network
Density variables Road;
Step 3, time-space registration is carried out to multi-source data using grid mode, has the grid of ground station true value will be as training
Sample generates the multi-source data collection of space-time uniformity;
Step 4, it will be input in deep learning model and be trained after the Grid square collection for having website PM2.5 true value normalization,
Inverting is carried out to unknown grid PM2.5 concentration by the model after being verified;
Step 5, fine PM2.5 spatial and temporal distributions are carried out to inversion result to chart.
2. the PM2.5 deep learning inversion method of combination remotely-sensed data according to claim 1 and social perception data,
It is characterized in that: being pre-processed described in step 1, including reject exceptional value in website PM2.5 data;Remote sensing image data format is turned
Change unified tiff image file format into;Social perception data is formatted using Python, ArcGIS, data sieve
After choosing, geographical coordinate conversion, vectoring operations are then carried out.
3. the PM2.5 deep learning inversion method of combination remotely-sensed data according to claim 1 and social perception data,
Be characterized in that: the calculation of PMS and PMT is as follows in step 2.1,
Wherein n, m are respectively spatial neighbor station number, time interval number, and w is weight, d, Δ t be respectively space length, when
Between be spaced.
4. the PM2.5 deep learning inversion method of combination remotely-sensed data according to claim 1 and social perception data,
It is characterized in that: original traffic exponent data progress cuclear density being analyzed by ArcGIS batch processing tool in step 2.3 and is handed over
Logical index variable Tindex, road mileage variable R oad;Real-time population is obtained by spatial interpolation methods to register quantitative variation
RTLD;Point of interest number variable POIs is extracted by buffer zone analysis.
5. the PM2.5 deep learning inversion method of combination remotely-sensed data according to claim 1 and social perception data,
Be characterized in that: the data input of deep learning model described in step 4 is using maximin normalized, model framework
Are as follows:
PM2.5=f (PM2.5 space-time initial fields, remote sensing variable, society's perception variable, auxiliary data variable) (3).
6. the PM2.5 deep learning inversion method of combination remotely-sensed data according to claim 1 and social perception data,
Be characterized in that: auxiliary data described in step 1 is meteorological data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910451339.6A CN110287455A (en) | 2019-05-28 | 2019-05-28 | A kind of PM2.5 deep learning inversion method of combination remotely-sensed data and social perception data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910451339.6A CN110287455A (en) | 2019-05-28 | 2019-05-28 | A kind of PM2.5 deep learning inversion method of combination remotely-sensed data and social perception data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110287455A true CN110287455A (en) | 2019-09-27 |
Family
ID=68002610
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910451339.6A Pending CN110287455A (en) | 2019-05-28 | 2019-05-28 | A kind of PM2.5 deep learning inversion method of combination remotely-sensed data and social perception data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110287455A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111723525A (en) * | 2020-06-23 | 2020-09-29 | 南通大学 | PM2.5 inversion method based on multi-source data and neural network model |
CN111738600A (en) * | 2020-06-23 | 2020-10-02 | 南通大学 | Urban road air quality evaluation method based on high-precision PM2.5 inversion result |
CN112905560A (en) * | 2021-02-02 | 2021-06-04 | 中国科学院地理科学与资源研究所 | Air pollution prediction method based on multi-source time-space big data deep fusion |
CN113344149A (en) * | 2021-08-06 | 2021-09-03 | 浙江大学 | PM2.5 hourly prediction method based on neural network |
CN114926749A (en) * | 2022-07-22 | 2022-08-19 | 山东大学 | Near-surface atmospheric pollutant inversion method and system based on remote sensing image |
CN114996624A (en) * | 2022-04-06 | 2022-09-02 | 武汉大学 | Remote sensing PM2.5 and NO based on multitask deep learning 2 Collaborative inversion method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106442236A (en) * | 2015-07-30 | 2017-02-22 | 中国科学院遥感与数字地球研究所 | Ground PM2.5 inversion method and system based on satellite remote sensing |
CN108241779A (en) * | 2017-12-29 | 2018-07-03 | 武汉大学 | Ground PM2.5 Density feature vectors space filter value modeling method based on remotely-sensed data |
CN108645768A (en) * | 2018-05-16 | 2018-10-12 | 长江师范学院 | A kind of PM2.5 remote sense monitoring systems, monitoring model method for building up and monitoring method |
-
2019
- 2019-05-28 CN CN201910451339.6A patent/CN110287455A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106442236A (en) * | 2015-07-30 | 2017-02-22 | 中国科学院遥感与数字地球研究所 | Ground PM2.5 inversion method and system based on satellite remote sensing |
CN108241779A (en) * | 2017-12-29 | 2018-07-03 | 武汉大学 | Ground PM2.5 Density feature vectors space filter value modeling method based on remotely-sensed data |
CN108645768A (en) * | 2018-05-16 | 2018-10-12 | 长江师范学院 | A kind of PM2.5 remote sense monitoring systems, monitoring model method for building up and monitoring method |
Non-Patent Citations (2)
Title |
---|
TONGWEN LI等: "Estimating ground-level PM2.5 by fusing satellite and station observations: A geo-intelligent deep learning approach", 《RESEARCHGATE》 * |
XIN FANG等: "Satellite-based ground PM2.5 estimation using timely structure adaptive modeling", 《RESEARCHGATE》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111723525A (en) * | 2020-06-23 | 2020-09-29 | 南通大学 | PM2.5 inversion method based on multi-source data and neural network model |
CN111738600A (en) * | 2020-06-23 | 2020-10-02 | 南通大学 | Urban road air quality evaluation method based on high-precision PM2.5 inversion result |
CN111723525B (en) * | 2020-06-23 | 2023-10-31 | 南通大学 | PM2.5 inversion method based on multi-source data and neural network model |
CN112905560A (en) * | 2021-02-02 | 2021-06-04 | 中国科学院地理科学与资源研究所 | Air pollution prediction method based on multi-source time-space big data deep fusion |
CN113344149A (en) * | 2021-08-06 | 2021-09-03 | 浙江大学 | PM2.5 hourly prediction method based on neural network |
CN114996624A (en) * | 2022-04-06 | 2022-09-02 | 武汉大学 | Remote sensing PM2.5 and NO based on multitask deep learning 2 Collaborative inversion method |
CN114996624B (en) * | 2022-04-06 | 2024-04-05 | 武汉大学 | Remote sensing PM2.5 and NO based on multitasking deep learning 2 Collaborative inversion method |
CN114926749A (en) * | 2022-07-22 | 2022-08-19 | 山东大学 | Near-surface atmospheric pollutant inversion method and system based on remote sensing image |
CN114926749B (en) * | 2022-07-22 | 2022-11-04 | 山东大学 | Near-surface atmospheric pollutant inversion method and system based on remote sensing image |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110287455A (en) | A kind of PM2.5 deep learning inversion method of combination remotely-sensed data and social perception data | |
US11333796B2 (en) | Spatial autocorrelation machine learning-based downscaling method and system of satellite precipitation data | |
CN109508360B (en) | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton | |
Shahzad et al. | TecDEM: A MATLAB based toolbox for tectonic geomorphology, Part 2: Surface dynamics and basin analysis | |
CN104361611B (en) | Group sparsity robust PCA-based moving object detecting method | |
CN108459318A (en) | Potential landslide EARLY RECOGNITION method based on remote sensing technology | |
CN108416985B (en) | Geological disaster monitoring and early warning system and method based on image recognition | |
CN113222283B (en) | Mountain torrent forecasting and early warning method and system based on digital twinning | |
CN110929607A (en) | Remote sensing identification method and system for urban building construction progress | |
CN111210483B (en) | Simulated satellite cloud picture generation method based on generation of countermeasure network and numerical mode product | |
CN111414954B (en) | Rock image retrieval method and system | |
CN107392252A (en) | Computer deep learning characteristics of image and the method for quantifying perceptibility | |
CN111563408B (en) | High-resolution image landslide automatic detection method with multi-level perception characteristics and progressive self-learning | |
CN116595121B (en) | Data display monitoring system based on remote sensing technology | |
CN115761513A (en) | Intelligent remote sensing identification method for mountain large landslide based on semi-supervised deep learning | |
CN117078627B (en) | Method and system for monitoring and safely evaluating defects of dam body of silt dam | |
CN108764527B (en) | Screening method for soil organic carbon library time-space dynamic prediction optimal environment variables | |
CN115439753A (en) | Steep river bank identification method and system based on DEM | |
CN114898089B (en) | Functional area extraction and classification method fusing high-resolution images and POI data | |
CN117933095B (en) | Earth surface emissivity real-time inversion and assimilation method based on machine learning | |
CN117409168B (en) | Flood forecasting and flood simulation method and system for real-time dynamic rendering | |
CN113671599A (en) | Global climate mode-based login cyclone identification method | |
CN116401879B (en) | Method for simulating downstream evolution of breaking tailing sand | |
CN115952743A (en) | Multi-source precipitation data collaborative downscaling method and system coupled with random forest and HASM | |
Socaci et al. | XNow: A deep learning technique for nowcasting based on radar products’ values prediction |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190927 |