CN109409441A - Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest - Google Patents

Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest Download PDF

Info

Publication number
CN109409441A
CN109409441A CN201811365431.2A CN201811365431A CN109409441A CN 109409441 A CN109409441 A CN 109409441A CN 201811365431 A CN201811365431 A CN 201811365431A CN 109409441 A CN109409441 A CN 109409441A
Authority
CN
China
Prior art keywords
sample
chlorophyll
concentration
random forest
value
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
Application number
CN201811365431.2A
Other languages
Chinese (zh)
Inventor
苏华
张明慧
季博文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fuzhou University
Original Assignee
Fuzhou University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Fuzhou University filed Critical Fuzhou University
Priority to CN201811365431.2A priority Critical patent/CN109409441A/en
Publication of CN109409441A publication Critical patent/CN109409441A/en
Pending legal-status Critical Current

Links

Classifications

    • 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
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • 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
    • G06F18/211—Selection of the most significant subset of features
    • 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
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N2021/1793—Remote sensing
    • G01N2021/1797—Remote sensing in landscape, e.g. crops

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)
  • Image Processing (AREA)

Abstract

The present invention relates to a kind of based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest, utilize time series remote sensing image data, data are observed in conjunction with timing buoy, with " it is sparse that Time Continuous makes up space " for modeling strategy, using improved random forest --- double weight random forest methods, it establishes intelligent remote sensing estimation model and remote-sensing inversion is carried out to nearshore waters chlorophyll-a concentration, good inversion result is obtained, can intuitively, accurately show a wide range of chlorophyll-a concentration spatial distribution of offshore.The invention can monitor for coastal waters chlorophyll-a concentration and provide a kind of macroscopical, continuous, effective method, can make up the deficiency of traditional chlorophyll a concentration monitor, practical value with higher.

Description

Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest
Technical field
The present invention relates to remote sensing information process and application field, and in particular to a kind of based on the offshore water for improving random forest Body chlorophyll-a concentration remote sensing inversion method.
Background technique
Chlorophyll-a concentration be can direct remote-sensing inversion one of important water quality parameter, be commonly used to evaluation coastal waters Eutrophic extent.The main means of traditional water quality monitoring are that sampling and lab analysis, this method have in precision on the spot There is certain accuracy, there is part and typically represent meaning, but the surface water quality situation on monitoring section can only be obtained, no It can reflect the overall change in time and space of entire water ecological setting, and time-consuming, laborious, at high cost, can not achieve real-time monitoring.With Traditional water quality monitoring is compared, and remote sensing water quality monitoring has the significant advantages such as quick, macroscopic view, dynamic, can be used to carry out a wide range of Water quality monitoring activity, can also with the occurrence and development process of dynamically track water pollution event, have the advantages that can not replace. How using limited buoy observation data and timing remote sensing image, reliable chlorophyll-a concentration timing remote-sensing inversion mould is constructed Type, it is accurate to understand the distribution of offshore case Ⅱ waters chlorophyll-a concentration and its spatial and temporal variation, have to offshore monitoring water environment important Meaning.
Offshore case Ⅱ waters chlorophyll-a concentration remote sensing inversion method can substantially be divided into empirical statistics method, Semi-empirical Analysis Method Method and radiative transfer model.The method is simple and quick for empirical statistics, but region versatility is poor.Semiempirical model is in empirical statistics It joined inherent optical properties on the basis of method, so that model is had more applicability, and improve inversion accuracy.Radiative transfer model Although precision with higher and preferable versatility, this model is absorption based on water body each component, known to scattering signatures It is assumed that a large amount of earlier measurement and analysis work must be done to corresponding waters, by measurement means and measure data precision Limitation, using effect are ideal not enough.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of coastal waters chlorophyll a based on improvement random forest is dense Remote sensing inversion method is spent, realizes the remote sensing monitoring of offshore water environment (such as red tide).
To achieve the above object, the present invention adopts the following technical scheme:
A kind of coastal waters chlorophyll-a concentration remote sensing inversion method based on improvement random forest, comprising the following steps:
Step S1: reflectance and actual measurement buoy data are obtained；
Step S2: according to actual measurement buoy data, making to survey chlorophyll-a concentration and reflectance correspond, and Establish eigenmatrix and label matrix；
Step S3: according to obtained eigenmatrix and label matrix, feature selecting is carried out using RReliefF algorithm；
Step S4: it is sampled using bootstrap sampling method, obtains multiple sample data sets, establish back respectively Return decision tree；
Step S5: assigning decision tree weight using weighted voting method, in conjunction with base learner, establishes double weights Random Forest model；
Step S6: using eigenmatrix as double weight Random Forest model input datas, it is green that coastal waters leaf is extracted in inverting Plain a concentration.
Further, the step S1 specifically includes the following steps:
Step S11: acquisition time sequence remote sensing image, and geometric correction is carried out to original remote sensing image, and according to research Region is cut；
Step S12: radiation calibration is carried out to the image after geometric correction and cutting；
Step S13: atmospheric correction is carried out to the image after radiation calibration, obtains reflectance；
Step S14: announcing data according to official, obtains actual measurement buoy data, and rejects actual measurement buoy data outliers.
Further, in step S2 specifically: according to buoy point latitude and longitude coordinates, by buoy point and remote sensing image picture Element is corresponding, extracts the remote sensing image pixel value of corresponding point, establishes eigenmatrix, and enabling actual measurement buoy data is label square Battle array.
Further, in the step S3, the RReliefF algorithm that uses specifically:
Step S31: sample R is selected from the sample set that size is mi, selected from remaining m-1 sample from the sample away from From k nearest sample；
Step S32: sample R is calculatediIn value P0Under the conditions of weight sets ndC, using following formula:
Wherein, P0Indicate sample RiPredicted value P value, Pi(1≤i≤k) is that i-th sample takes in k sample Value, PmaxAnd PminThe maximum value and minimum value of P respectively in sample；
Step S33: it calculates in the sample RiWeight sets n under the conditions of approximate variables AdA[A],
Using following formula:
Wherein, A0Indicate sample RiApproximate variables A value, Ai(1≤i≤k) is close in i-th of sample in k sample Like the value of variables A, AmaxAnd AminIt is the maximum value and minimum value of approximate variables A in sample respectively；
Step S34: the sample Ri predictive variable P is calculated0With weight sets n under the conditions of approximate variables AdC&dA[A], under Formula:
Step S35: step S31, step S32, step S33, step S34 are repeated, Repeated m -1 time, is selected every time different M n is obtained in sampledC, m ndA[A], m ndC&dA[A]；
Step S36: N is calculated separatelydC、NdA[A]、NdC&dA[A]；
Wherein, NdCIt is m ndCThe sum of, NdA[A] is m ndAThe sum of [A], NdC&dA[A] is m ndC&dAThe sum of [A]；
Step S37: the weighted value W [A] of approximate variables A is calculated:
Further, it in step S4, is sampled using bootstrap sampling method, establishes multiple regression trees The following steps are included:
Step S41: the given sampling set comprising m sample is sampled using bootstrap sampling method, is sampled The T sampling sets containing m training sample out；
Step S42: a regression tree base learner is trained based on each sampling set；
Step S43: step S41, S42 is repeated, T decision tree is established.
Further, in step S5, assign decision tree weight using weighted voting method, establish double weights with Machine forest model the following steps are included:
Step S51: T decision tree base learner is combined, using following formula:
Wherein, hiIt (x) is base learner hiOutput on feature x, ωiFor the weight of individual base learner, usual ωi >=0,
Compared with the prior art, the invention has the following beneficial effects:
The present invention, using the modeling strategy of " it is sparse that Time Continuous makes up space ", is made up based on time series data Traditional water quality monitoring can not provide continuous space, macroscopic view, the limitation of a wide range of parameter information distribution, make in modeling process Feature selecting is carried out with RreliefF algorithm, the combination of base learner is carried out using weighted voting method, avoids tradition Random forests algorithm enables the problem of regression tree effect difference when feature is sampled because sampling characterization ability difference, can be offshore water Environment is continuous, a wide range of monitor provides technical support.
Detailed description of the invention
Fig. 1 is method flow schematic diagram of the invention.
Fig. 2 is the precision test figure of the embodiment of the present invention.
Fig. 3 is the result display diagram of the embodiment of the present invention.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and embodiments.
Fig. 1 is please referred to, the present invention provides a kind of based on the coastal waters chlorophyll-a concentration remote-sensing inversion for improving random forest Method, comprising the following steps:
Step S1: obtain survey region time series original remote sensing image (the present embodiment obtain the image time be 2017 In May, 2018 in May-year), the pre- places such as geometric correction, region cutting, radiation calibration, atmospheric correction are carried out to remote sensing image data Reason obtains reflectance；The actual measurement buoy data of acquisition are arranged, according to the publication of Fujian Province's marine forecasting platform Water quality information screens actual measurement exceptional value.
Specific step is as follows:
Step S11: sequential images (MODIS 500m LB1) the time required to collection research region, to original remote sensing image into Row geometric correction cuts the image after geometric correction according to survey region；
Step S12: radiation calibration is carried out to the image after geometric correction and cutting；
Step S13: selection FLAASH atmospheric correction method carries out atmospheric correction to the image after radiation calibration, obtains Reflectance is taken, and converts tiff format for image；
Step S14: the oceanographic observation data announced according to Fujian Province's marine forecasting platform are rejected abnormal in actual measurement buoy data High or abnormal low numerical value.
Step S2: grid coordinate corresponding with tiff image is converted by actual measurement buoy dump point coordinate data, makes to survey Chlorophyll-a concentration and image spectral reflectivity correspond, and using respective pixel values as input feature vector, establish eigenmatrix X, right Station data should be surveyed as label, establish label matrix Y.
Step S3: feature selecting is carried out using RReliefF algorithm, k arest neighbors of sample is found to update weight, subtracts The light influence of noise, and without mutually independent strong assumption between feature, it can preferably identify strong dependence feature.
Specific step is as follows by the step S3:
Step S31: sample R is selected from the sample set that size is mi, selected from remaining m-1 sample from the sample away from From k nearest sample；
Step S32: sample R is calculatediIn predictive variable value P0Under the conditions of weight sets ndC, using following formula:
Wherein, P0Indicate sample RiPredictive variable P value, Pi(1≤i≤k) is the pre- of i-th sample in k sample Survey variable, PmaxAnd PminThe maximum value and minimum value of predictive variable respectively in sample；
Step S33: it calculates in the sample RiWeight sets n under the conditions of approximate variables AdA[A], using following formula:
Wherein, A0Indicate sample RiApproximate variables A value, Ai(1≤i≤k) is close in i-th of sample in k sample Like the value of variables A, AmaxAnd AminIt is the maximum value and minimum value of approximate variables A in sample respectively；
Step S34: the sample R is calculatediPredictive variable P0With weight sets n under the conditions of approximate variables AdC&dA[A], under Formula:
Step S35: repeating step S31, step S32, step S33, step S34, and Repeated m -1 time in total selects not every time Same sample, is obtained m ndC, m ndA[A], m ndC&dA[A]；
Step S36: N is calculated separatelydC、NdA[A]、NdC&dA[A]；
Wherein, NdCIt is m ndCThe sum of, NdA[A] is m ndAThe sum of [A], NdC&dA[A] is m ndC&dAThe sum of [A]；
Step S37: the weighted value W [A] of approximate variables A is calculated:
Step S4: being sampled using bootstrap sampling method, i.e., the given data set comprising m sample, I It is first random take out one and be put into sampling set, then the sample is put back to initial data set, so that the sample is still when sampling next time It is possible that selected, in this way, operating by m stochastical sampling, we obtain the sampling set containing m sample, and initial training is concentrated Some samples repeatedly occur in sampling set, and some then never occurs.We sample out T containing m sample according to the above method Data set establishes regression tree based on each sampling set respectively.
Specific step is as follows:
Step S41: the given sampling set comprising m sample is sampled using bootstrap sampling method, is sampled The T sampling sets containing m training sample out；
Step S42: a regression tree base learner is trained based on each sampling set, process is as follows:
Step S43: step S41, S42 is repeated, T decision tree is established.
Step S5: and the combination strategy of traditional random forest simple linear weighted regression is different, is assigned using different algorithms Decision tree weight is given, in conjunction with base learner, establishes double weight Random Forest models.T decision tree base learner is combined, Using following formula:
Wherein, hiIt (x) is base learner hiOutput on feature x, ωiFor the weight of individual base learner, usual ωi >=0,
Step S6: the original remote sensing image of survey region is pre-processed according to step S1, obtains image feature matrix Afterwards, make it as mode input data, coastal waters chlorophyll-a concentration is extracted in inverting.
The foregoing is merely presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with Modification, is all covered by the present invention.

Claims (6)

1. it is a kind of based on improve random forest coastal waters chlorophyll-a concentration remote sensing inversion method, which is characterized in that including with Lower step:
Step S1: reflectance and actual measurement buoy data are obtained；
Step S2: according to actual measurement buoy data, make to survey chlorophyll-a concentration and reflectance corresponds, and establish Eigenmatrix and label matrix；
Step S3: according to obtained eigenmatrix and label matrix, feature selecting is carried out using RReliefF algorithm；
Step S4: being sampled using bootstrap sampling method, obtain multiple sample data sets, is established to return respectively and be determined Plan tree；
Step S5: assigning decision tree weight using weighted voting method, and in conjunction with base learner, it is random to establish double weights Forest model；
Step S6: using eigenmatrix as double weight Random Forest model input datas, it is dense that coastal waters chlorophyll a is extracted in inverting Degree.
2. according to claim 1 a kind of based on the coastal waters chlorophyll-a concentration remote-sensing inversion side for improving random forest Method, it is characterised in that: the step S1 specifically includes the following steps:
Step S11: acquisition time sequence remote sensing image, and geometric correction is carried out to original remote sensing image, and according to survey region It is cut；
Step S12: radiation calibration is carried out to the image after geometric correction and cutting；
Step S13: atmospheric correction is carried out to the image after radiation calibration, obtains reflectance；
Step S14: announcing data according to official, obtains actual measurement buoy data, and rejects actual measurement buoy data outliers.
3. according to claim 1 a kind of based on the coastal waters chlorophyll-a concentration remote-sensing inversion side for improving random forest Method, it is characterised in that: in step S2 specifically: according to buoy point latitude and longitude coordinates, by buoy point and remote sensing image pixel It is corresponding, the remote sensing image pixel value of corresponding point is extracted, eigenmatrix is established, and enabling actual measurement buoy data is label matrix.
4. according to claim 1 a kind of based on the coastal waters chlorophyll-a concentration remote-sensing inversion side for improving random forest Method, it is characterised in that: in the step S3, the RReliefF algorithm that uses specifically:
Step S31: sample R is selected from the sample set that size is mi, selected with a distance from the sample most from remaining m-1 sample K close sample；
Step S32: sample R is calculatediIn value P0Under the conditions of weight sets ndC, using following formula:
Wherein, P0Indicate sample RiPredicted value P value, Pi(1≤i≤k) is the value of i-th of sample in k sample, Pmax And PminThe maximum value and minimum value of P respectively in sample；
Step S33: it calculates in the sample RiWeight sets n under the conditions of approximate variables AdA[A],
Using following formula:
Wherein, A0Indicate sample RiApproximate variables A value, Ai(1≤i≤k) is approximate in i-th of sample in k sample become Measure the value of A, AmaxAnd AminIt is the maximum value and minimum value of approximate variables A in sample respectively；
Step S34: the sample Ri predictive variable P is calculated0With weight sets n under the conditions of approximate variables AdC&dA[A], using following formula:
Step S35: repeating step S31, step S32, step S33, step S34, Repeated m -1 time, select different samples every time, M n is obtaineddC, m ndA[A], m ndC&dA[A]；
Step S36: N is calculated separatelydC、NdA[A]、NdC&dA[A]；
Wherein, NdCIt is m ndCThe sum of, NdA[A] is m ndAThe sum of [A], NdC&dA[A] is m ndC&dAThe sum of [A]；
Step S37: the weighted value W [A] of approximate variables A is calculated:
5. according to claim 1 a kind of based on the coastal waters chlorophyll-a concentration remote-sensing inversion side for improving random forest Method, it is characterised in that: in step S4, sampled using bootstrap sampling method, establish multiple regression tree packets Include following steps:
Step S41: the given sampling set comprising m sample is sampled using bootstrap sampling method, samples out T A sampling set containing m training sample；
Step S42: a regression tree base learner is trained based on each sampling set；
Step S43: step S41, S42 is repeated, T decision tree is established.
6. according to claim 1 a kind of based on the coastal waters chlorophyll-a concentration remote-sensing inversion side for improving random forest Method, it is characterised in that: in step S5, assign decision tree weight using weighted voting method, it is gloomy at random to establish double weights Woods model the following steps are included:
Step S51: T decision tree base learner is combined, using following formula:
Wherein, hiIt (x) is base learner hiOutput on feature x, ωiFor the weight of individual base learner, usual ωi>=0,
CN201811365431.2A 2018-11-16 2018-11-16 Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest Pending CN109409441A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811365431.2A CN109409441A (en) 2018-11-16 2018-11-16 Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811365431.2A CN109409441A (en) 2018-11-16 2018-11-16 Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest

Publications (1)

Publication Number Publication Date
CN109409441A true CN109409441A (en) 2019-03-01

Family

ID=65473846

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811365431.2A Pending CN109409441A (en) 2018-11-16 2018-11-16 Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest

Country Status (1)

Country Link
CN (1) CN109409441A (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109781626A (en) * 2019-03-11 2019-05-21 王祥 A kind of offshore based on spectrum analysis uphangs husky water body green tide remote sensing recognition method
CN110851789A (en) * 2019-09-30 2020-02-28 广州地理研究所 Island reef shallow sea water depth prediction method based on extreme gradient lifting
CN110865040A (en) * 2019-11-29 2020-03-06 深圳航天智慧城市系统技术研究院有限公司 Sky-ground integrated hyperspectral water quality monitoring and analyzing method
CN110909949A (en) * 2019-11-29 2020-03-24 山东大学 Near-shore sea area chlorophyll a concentration prediction method based on clustering-regression algorithm
CN110927065A (en) * 2019-11-02 2020-03-27 生态环境部卫星环境应用中心 Remote sensing assisted lake and reservoir chl-a concentration spatial interpolation method optimization method and device
CN111291621A (en) * 2020-01-14 2020-06-16 中国科学院南京地理与湖泊研究所 Method for quantitatively evaluating influence of offshore aquaculture pond on offshore Chl-a concentration
CN111504915A (en) * 2020-04-27 2020-08-07 中国科学技术大学先进技术研究院 Method, device and equipment for inverting chlorophyll concentration of water body and storage medium
CN112070234A (en) * 2020-09-04 2020-12-11 中国科学院南京地理与湖泊研究所 Ground-based remote sensing machine learning algorithm for chlorophyll and phycocyanin in water body under complex scene
CN116338819A (en) * 2023-03-27 2023-06-27 北京智科远达数据技术有限公司 Water dissolved oxygen concentration prediction system
CN116337819A (en) * 2023-03-27 2023-06-27 北京智科远达数据技术有限公司 Inversion method of water body chemical oxygen demand concentration
CN117434010A (en) * 2020-10-27 2024-01-23 淮阴师范学院 Inland lake water body pheophytin concentration remote sensing inversion model and method based on Decission Tree algorithm

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101852722A (en) * 2010-05-20 2010-10-06 北京航空航天大学 Method for evaluating remote sensing inversion accuracy of chlorophyll a in water body
CN104318270A (en) * 2014-11-21 2015-01-28 东北林业大学 Land cover classification method based on MODIS time series data
US20150213389A1 (en) * 2014-01-29 2015-07-30 Adobe Systems Incorporated Determining and analyzing key performance indicators
CN107271382A (en) * 2017-06-02 2017-10-20 西北农林科技大学 A kind of different growing rape leaf SPAD value remote sensing estimation methods

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101852722A (en) * 2010-05-20 2010-10-06 北京航空航天大学 Method for evaluating remote sensing inversion accuracy of chlorophyll a in water body
US20150213389A1 (en) * 2014-01-29 2015-07-30 Adobe Systems Incorporated Determining and analyzing key performance indicators
CN104318270A (en) * 2014-11-21 2015-01-28 东北林业大学 Land cover classification method based on MODIS time series data
CN107271382A (en) * 2017-06-02 2017-10-20 西北农林科技大学 A kind of different growing rape leaf SPAD value remote sensing estimation methods

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
MOHAMED BADER-EL-DEN: "Self-adaptive heterogeneous random forest", 《AICCSA》 *
ONDŘEJ ZAPLETAL: "SROVNÁNÍ SPRÁVNOSTI KLASIFIKACE POMOCÍ TRADIČNÍCH MODELŮ A META-MODELŮ", 《DSPACE.VUTBR.CZ》 *
VRUSHALI Y. KULKARNI等: "Weighted Hybrid Decision Tree Model for Random Forest Classifier", 《JOURNAL OF THE INSTITUTION OF ENGINEERS》 *
余绍黔: "《服务外包校企合作对高校教师队伍建设的影响因素及对策研究》", 30 June 2018 *
张明慧 等: "MODIS时序影像的福建近岸叶绿素a浓度反演", 《环境科学学报》 *
郭豪: "双权重随机森林预测算法及其并行化研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109781626A (en) * 2019-03-11 2019-05-21 王祥 A kind of offshore based on spectrum analysis uphangs husky water body green tide remote sensing recognition method
CN109781626B (en) * 2019-03-11 2021-07-06 王祥 Near-shore high-suspended sand water body green tide remote sensing identification method based on spectral analysis
CN110851789A (en) * 2019-09-30 2020-02-28 广州地理研究所 Island reef shallow sea water depth prediction method based on extreme gradient lifting
CN110927065A (en) * 2019-11-02 2020-03-27 生态环境部卫星环境应用中心 Remote sensing assisted lake and reservoir chl-a concentration spatial interpolation method optimization method and device
CN110909949A (en) * 2019-11-29 2020-03-24 山东大学 Near-shore sea area chlorophyll a concentration prediction method based on clustering-regression algorithm
CN110865040A (en) * 2019-11-29 2020-03-06 深圳航天智慧城市系统技术研究院有限公司 Sky-ground integrated hyperspectral water quality monitoring and analyzing method
CN110909949B (en) * 2019-11-29 2023-08-25 山东大学 Method for predicting chlorophyll a concentration of offshore area based on clustering-regression algorithm
CN111291621A (en) * 2020-01-14 2020-06-16 中国科学院南京地理与湖泊研究所 Method for quantitatively evaluating influence of offshore aquaculture pond on offshore Chl-a concentration
CN111291621B (en) * 2020-01-14 2023-05-16 中国科学院南京地理与湖泊研究所 Method for quantitatively evaluating influence of offshore aquaculture pond on offshore Chl-a concentration
CN111504915A (en) * 2020-04-27 2020-08-07 中国科学技术大学先进技术研究院 Method, device and equipment for inverting chlorophyll concentration of water body and storage medium
CN112070234A (en) * 2020-09-04 2020-12-11 中国科学院南京地理与湖泊研究所 Ground-based remote sensing machine learning algorithm for chlorophyll and phycocyanin in water body under complex scene
CN112070234B (en) * 2020-09-04 2024-01-30 中国科学院南京地理与湖泊研究所 Water chlorophyll and phycocyanin land-based remote sensing machine learning algorithm under complex scene
CN117434010A (en) * 2020-10-27 2024-01-23 淮阴师范学院 Inland lake water body pheophytin concentration remote sensing inversion model and method based on Decission Tree algorithm
CN116338819A (en) * 2023-03-27 2023-06-27 北京智科远达数据技术有限公司 Water dissolved oxygen concentration prediction system
CN116337819A (en) * 2023-03-27 2023-06-27 北京智科远达数据技术有限公司 Inversion method of water body chemical oxygen demand concentration
CN116337819B (en) * 2023-03-27 2023-10-27 北京智科远达数据技术有限公司 Inversion method of water body chemical oxygen demand concentration

Similar Documents

Publication Publication Date Title
CN109409441A (en) Based on the coastal waters chlorophyll-a concentration remote sensing inversion method for improving random forest
CN105608293B (en) Merge Biomass retrieval method and system on the woodland of spectrum and textural characteristics
EP2287638B1 (en) Method and device for evaluation of solar radiation intensity
CN111968080B (en) Method for detecting quality of inside and outside of Feicheng peaches based on hyperspectral and deep learning
CN110865040A (en) Sky-ground integrated hyperspectral water quality monitoring and analyzing method
Fournier et al. Hemispherical photography in forest science: Theory, methods, applications
Blanchette et al. Predicting wood fiber attributes using local-scale metrics from terrestrial LiDAR data: A case study of Newfoundland conifer species
CN110927082A (en) Winter wheat yield prediction method based on unmanned aerial vehicle imaging hyperspectral remote sensing
CN109726705B (en) Mangrove forest information extraction method and device and electronic equipment
CN107423537A (en) A kind of method of the surface temperature NO emissions reduction based on adaptive threshold
CN102855485B (en) The automatic testing method of one grow wheat heading
CN115877555B (en) Microscope illumination simulation system
CN109490306A (en) A kind of pork freshness detection method based on color and smell data fusion
CN104915982A (en) Canopy layer illumination distribution prediction model construction method and illumination distribution detection method
CN105913460A (en) Skin color detection method and device
CN109191386A (en) A kind of quick Gamma bearing calibration and device based on BPNN
CN110321774A (en) Crops evaluation methods for disaster condition, device, equipment and computer readable storage medium
CN117111092A (en) High-spatial-resolution remote sensing water quality detection method based on machine learning
CN117348831B (en) Picture adjustment method and system for liquid crystal display screen
AU2010333914A1 (en) Method and apparatus for predicting information about trees in images
CN114140691A (en) Method for predicting crop lodging angle and form based on visible light image of unmanned aerial vehicle
CN113256733A (en) Camera spectral sensitivity reconstruction method based on confidence voting convolutional neural network
CN117035066A (en) Ground surface temperature downscaling method coupling geographic weighting and random forest
CN108240806B (en) Method for selecting high-spectrum water depth inversion wave band
KR102576427B1 (en) Real-time Rainfall Prediction Device using Cloud Images, and Rainfall Prediction Method using the same, and a computer-readable storage medium

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

Application publication date: 20190301

RJ01 Rejection of invention patent application after publication