Dissertation > Agricultural Sciences > Agriculture as the foundation of science > Agricultural physics > The application of remote sensing in agriculture

Estimation of Crop’s Chlorophyll Content with Hyperspectral Remote Sensing

Author QiaoZhenMin
Tutor XingLiXin
School Jilin University
Course Cartography and Geographic Information Systems
Keywords Hyperion Hyperspectral Remote Sensing Chlorophyll Estimation Crop’sClassification
CLC S127
Type Master's thesis
Year 2012
Downloads 249
Quotes 2
Download Dissertation

Chlorophyll is an important crop photosynthesis pigment that directly indicatesthe crop’s healthes status, the photosynthesis ability and the crop’sharvest.Accuratly,rapidly getting the specific crop’s chlorophyll content regional canprovide effective data source for agricultural decision-making.Corn as a summercrop,is not like the wheat and rice that can easly distinguish with other crops.Coin isselected as a crop study so that ensure the method of Estimation of Crop’sChlorophyll Content with Hyperspectral Remote Sensing can applies to all crops.Using satellites and ground synchronous observation method, preprocessHyperion image and then extract corn thematic information, at the same timecalculate the chlorophyll estimation model, The ground measured data was gettedfrom”Standard Spectral Knowledge Base”.With the model do the regional cornchlorophyll content map to the corn thematic information map.The experimentconfirmed,the method of hyperspectral remote sensing of crop chlorophyll contentestimation and mapping can quickly map the specific crop chlorophyll content andcan quickly,accuratly reflect the crop’s growing space distribution.The conclusionsare as follows based on the above studies:1、With the hyperion data can extract the crop thematic information,so that getthe corn thematic information.The pre-processing of hyperion data should be donebefore corn thematic information extracting,which include poor quality bandsremoving, projection transforming and data subsetting, bad lines and stripesrepairing,Smile effect eliminating and atmospheric correcting.The step of bandsremoving can remove the poor quality bands and water vapor absorption bands.Theprojection transforming and data subsetting can remove the cloud cover area andensure the coordinate system of image coordinates and ground truth points’consistency. The absolute radiation values calculation can get the hyperion data’s truereflection values.The bad lines repairation refers to detectors’ not working properly orthe calibration problem that cause the entire column data has no spectralinformation.The stripes are the lines whoes value is very low but zero.Smile effect isthe Phenomenon that the pixels’ wavelengths offsets from the central location to both sides.The Smile effect can be detected by MNF transformation method.The firstprincipal component has too small contrast and with a clear vertical brightnessgradient.Atmospheric correction refers to the impact of factors such as atmosphericand illumination’s elimination so that obtain the true reflectance value.The betterquality hyperspectral reflectance data can be getted with the above pretreatment.The hyperion reflectance data should be done with the dimensionality reduction,endmember extraction, classification and classification accuracyassessment.Hyperspectral data contained lots of redundant information,which wasdone by MNF transformation method,the eigenvalue of MNF’s principal componentsshows that the first15principal components can be used for endmember extraction,and the PPI was show in the n-dimensional visualization space,which was rotated anddisplayed so that the endmember can be selected.The thematic information extractionwas done with the spectral angle mapper method and the endmembers as the trainingsamples. Accuracy assessment was done with the ground points, four of six groundpoins was in the classify results which shows a good result.2、The chlorophyll content estimation model was established and the corn’schlorophyll mapping was done with the model. correlation between chlorophyllcontent and the true ground spectrum, its first-order differential form and spectralindex of corn was calculated.The index of PRI has high correlations with chlorophylla, chlorophyll b and chlorophyll a+b,which was0.429,0.563and0.4733.The PRIwas used to establish the chlorophyll content estimation model.The corn chlorophyllcontent mapping was done with thechlorophyll a, chlorophyll b and chlorophylla+b’s estimation model.The corn was selected as the study’s crop,which is the summer crop and wasmixed cropping with other crops so that ensuring the method of estimation of crop’schlorophyll content with Hyperspectral Remote Sensing can applies to all crops.Thestudy can be a reference of the estimation of crop chlorophyll content as well asphysical and chemical parameters estimation and mapping later

Related Dissertations
More Dissertations