摘要
利用多时相的遥感图像和基于知识的分层分类方法对珠江口内伶仃洋湿地的时空变化进行了分析,并结合土地利用转换矩阵分析了近20 a的变化。通过精度分析,4个时相的分类精度大于85%,kappa系数大于0.8,达到较高分类精度。研究表明,1988-2004年间,几类典型湿地(河流、水库、鱼塘、滩涂和红树林)的面积减少了12265.68 hm2,由1988年湿地面积占研究区整个面积的43.35%下降至2004年的39.98%。有91.92%的滩涂转为其他用地,其中39.18%(1760.49 hm2)转为鱼塘,19.45%(873.81 hm2)转为农田,16.53%(742.77 hm2)转为河流,5.14%(230.94 hm2)转为红树林。有80.43%的鱼塘转为其他用地,其中25.67%转为建设用地,22.83%转为城市用地,21.26%转为林地。有62.21%的红树林专为其他用地,其中26.03%(123.12 hm2)转为鱼塘,14.92%(70.56 hm2)转为建设用地,9.46%(44.73 hm2)转为农田。
The Pearl River Estuary wetland in the developed region was selected as the studying object in this thesis. The integrated approach including the remote sensing and geography information system （GIS） , the field survey, and collecting history data was adopted to analyze the spatio-temporal evolution in recent twenty years. In the wetland classification of east and west shore of inner Lingding ocean, taking ERDAS software as the flat roof, gradation and classification method based-on knowledge were adopted. The main conclusions were summed up as following： （1） It is a great advantage to study the wetland landscape evolution using remote sensing and GIS in the regional scale of the Pearl River Estuary. The gradation and classification method based-on knowledge can overcome the shortage of human-computer alternation translation with eyes, realize the goal of classification for great region and the more temporal images, and also fulfill the landscape mapping. In this paper, the total classification resolution reaches 85 percent, the coefficient of kappa exceeds 0. 8. （2） The results indicated that the areas of wetlands reduced 12270 hm^2 during the period of 1988 to 2004. The proportion of wetland accounting for the research area descended from 43.4 percent in 1988 to 40 percent in 2004. The land use conversion matrix analysis indicated that 91.92 percent tidal flat, 80.43 percent dike-pond, 62.21 percent mangrove forest, 54.63 percent reservoir, and 13.06 percent river transferred into other land utilization from 1988 to 2004.
出处
《中山大学学报（自然科学版）》
CAS
CSCD
北大核心
2007年第2期105-109,共5页
Acta Scientiarum Naturalium Universitatis Sunyatseni
基金
"985工程"GIS与遥感的地学应用科技创新平台(Ⅱ类)基金资助项目(105203200400006)
中山大学理工科青年教师科研启动基金资助项目
关键词
湿地资源
河口湿地
遥感
珠江口
wetland landscape
estuary wetland
remote sensing
Pearl River estuary
