Font Size: a A A

Recognition And Extraction Of Agricultural Crops On County Scale Using Remote Sensing Technology

Posted on:2015-06-30Degree:MasterType:Thesis
Country:ChinaCandidate:S T BanFull Text:PDF
GTID:2283330434960239Subject:Land Resource and Spatial Information Technology
Abstract/Summary:PDF Full Text Request
This study chose Fufeng County, Shaanxi Province as the study area. The goal is toclassify the different agricultural crops and make their distribution map in the study area fromthe summer of2013to the spring of2014. Multiple data was used in the process ofclassification, including remote sensing images acquired by Landsat8OLI, ground spectraldata of crops and digital elevation model (DEM) of Fufeng County. The results are as follows:(1)From the summer of2013to the spring of2014, the main agricultural crops in thestudy area were winter wheat, summer maize, orchard and oilseed rape. The area of winterwheat was the largest, which was284.11km2, the area of summer maize was215.17km2, thearea of orchard was166.83km2, the area of oilseed rape was11.19km2.(2) The best time to extract oilseed rape from the remote sensing image was April, whenthe oilseed rape was blooming. The best band combination for extracting oilseed rape was5(R),3(G),2(B) combination. The optimal extracting method was Target Detection method.(3) The extraction of summer maize need to do the change detection statistics of thelandcover between summer when the maize was still green and fall when the maize washarvested and the land became bare. The optimal extracting method was Change Detectionmethod..(4) Winter was the best time to extract winter wheat from the remote sensing image. Twodifferent methods were used to do the extraction. First, the winter wheat can be extracted fromthe satellite image of January1,2014using Target Detection method. Second, the winterwheat area can be detected by doing change detection statistics of the landcover betweenOctober22,2013when the wheat land was bare and January1,2014when the wheat landwas green. The results showed that the second method was more accurate.(5) The best time to extract orchard from the remote sensing image was October, whenthe autumn crop were harvested while the winter crops were still under the ground.Classification rules were made according to the spectral and geographic features of orchard.These rules were used by the Decision Tree Classification method to extract orchard.(6) The precision analysis of the extraction results showed: the extraction accuracy oforchard was99.90%, the extraction accuracy of oilseed was88.10%, the extraction accuracy of summer maize was98.80%, the extraction accuracy of winter wheat by Target Detectionmethod was96.44%, the extraction accuracy of winter wheat by Change Detection methodwas98.80%. All the extraction results were precise enough to be used as the reference inagricultural production planning.
Keywords/Search Tags:Remote sensing, GIS, Landsat8, Agricultural crops, Recognition
PDF Full Text Request
Related items