應(yīng)用吉布斯隨機(jī)場模擬北京大興土壤質(zhì)地分布圖.pdf

1、中國農(nóng)業(yè)大學(xué)碩士學(xué)位論文應(yīng)用吉布斯隨機(jī)場模擬北京大興土壤質(zhì)地分布圖姓名：孫啟超申請學(xué)位級別：碩士專業(yè)：土地利用與信息技術(shù)指導(dǎo)教師：任理20060601ABSTRACTDistributionofsoiltextureintherealworldisverycomplexAtpresentthemaindescriptivetoolsusedaregeostatisticsandMarkovChain(MarkovRandomFielnt

2、hemethodofthispaperbelongstothelaterThepaperappliesGibbsRandomFieldwhichisallimagerestorationmodelintroducedbyTommyNorbergeta1(2002)tosimulatethedistributionofsoiltextureinDaxing。BeijingThemodelmainlyIlsesthemaximumlikel

3、ihoodtechniqueoftransitionprobabilitymatrixforMarkovChaininordertopredictunobservedpartofthemapFirstabout90dataofsoiltexturedistributeinDaringissurveyed，inordertocheckthecapabilityofthemodelasthesameproportionofthesurvey

4、eddataisknew，wemadesomesimulationsbasedonexactgeologicalmapfromTommyNorbergeta1(2002)WemadeaBemodlithinningofthegeologicalmapandtheprobabilityfordatatobekeptisabout90／0theseveralrestorationsshowedthattheexactofthepredict

5、ionsisnotgoodeomparedwithmentionedinthepaperfromTommyNorbergeta1(2002)Second，wemadethesimulationofthemapofsoiltextureinDaxingBeiiingIntheend，tounderstanddeop睜thecapabilj口ofthemodeltoprediction，wedidtwokindsofexperiments：

6、changeedtheformofenergy矗ⅡK_tionandincreasedtheproportionofthesurveyeddatasuchas30％，45％，60％75％Thestudyshowedthatbothofthemhavenoobviousimprovementandthatthemodelcouldcapturethemaindistributionfeatureofthemapalongwiththein