CAJ | 학술논문

土壤盐渍化是制约农业生产和发展的主要障碍之一。目前针对耕地土壤盐渍化的研究较少，常规土壤盐渍化的遥感监测主要是基于中低分辨率卫星，采用传统的基于象元分类方法，对盐渍化信息的细节描述不足，监测精度不高。使用国产GF-1影像，结合面向对象方法，针对耕地环境下的盐渍化信息进行精确提取和分类，再利用传统分类方法及实地采样数据进行精度分析。结果表明：面向对象法和最大似然法的分类总体精度分别为91．68％和72．48％，Kappa系数分别为0．90和0．69。该技术能准确提取田间尺度下耕地盐渍地信息，在未来的农田盐渍化高精度监测研究中具有一定应用价值和发展潜力。
토양염지화시제약농업생산화발전적주요장애지일。목전침대경지토양염지화적연구교소，상규토양염지화적요감감측주요시기우중저분변솔위성，채용전통적기우상원분류방법，대염지화신식적세절묘술불족，감측정도불고。사용국산GF-1영상，결합면향대상방법，침대경지배경하적염지화신식진행정학제취화분류，재이용전통분류방법급실지채양수거진행정도분석。결과표명：면향대상법화최대사연법적분류총체정도분별위91．68％화72．48％，Kappa계수분별위0．90화0．69。해기술능준학제취전간척도하경지염지지신식，재미래적농전염지화고정도감측연구중구유일정응용개치화발전잠력。
Soil salinization is one of the main obstacles to agricultural production and development .Now research on farmland soil salinization is less ,the conventional soil salinization is mainly based on low resolution satellite remote sensing monitoring ,and the traditional method is used based on pixel classification ,a detailed description of salinization information is insufficient ,monitoring ac‐curacy is not high .This paper uses domestic GF‐1 in combination with object‐oriented method ,in view of the salinization of cultivat‐ed land under the environment of information accurate extraction ,classification and the traditional classification method is reused and the field sampling data accurac is analyzed .Results show that the object‐oriented method and maximum likelihood classification accu‐racy are 91 .68% and 72 .48% ,the Kappa coefficients are 0 .90 and 0 .69 ,respectively .The technology can accurately extract the field scale ,arable land saline land information in the future of farmland salinization high precision monitoring has certain application value in the research and development potential .