论文部分内容阅读
以乌龙江流域龙祥岛地区为研究区,以两期不同水位的Landsat-8 OLI影像与高分一号(GF-1)PMS1影像为数据源,经过面向对象的分割技术,以CART(Classification And Regression Tree,CART)决策树方法挖掘不同地类的特征指数阈值,提取出该地区滩涂湿地并根据水位值进行变化检测分析.研究表明:GF-1 PMS1传感器8 m分辨率影像可以与Landsat-8 OLI传感器15 m分辨率影像结合进行变化检测,总体分类精度分别达到0.89和0.92;CART决策树算法与面向对象的分类方法相结合可以获得很好的分类效果,对于湿地滩涂信息提取具有高效快捷的优势,这种方法随着多元遥感的发展将展现巨大的应用潜力;水位变化对湿地滩涂的影响可以通过遥感影像进行检测分析,两期影像T1与T2两时刻水位高程分别为3.34 m与3.56 m,水位变化0.22 m,T1时刻河流水面面积为1 811.09 hm~2,T2时刻河流水面面积为2 092.34hm~2,两期变化增加的水面面积为281.25 hm~2,未变化面积为1 778.47 hm~2,占T1时刻总面积的98.20%,T1时刻内陆滩涂转化为T2时刻河流水面的面积为244.91 hm~2,占T1时刻内陆滩涂总面积的34.60%.
Taking Longxiang Island of Wulongjiang River Basin as the research area, Landsat-8 OLI image and GF-1 PMS1 image of two different water levels were used as data sources. After object-oriented segmentation and CART ( Classification And Regression Tree (CART) was used to mine the characteristic index thresholds of different land types and to extract the wetland of the beach and to detect and analyze the change of the wetland in accordance with the water level value.The results show that the 8 m resolution image of GF-1 PMS1 sensor can be compared with Landsat -8 OLI sensor with 15 m resolution images, the overall classification accuracy is 0.89 and 0.92, respectively. Combining the CART decision tree algorithm and the object-oriented classification method, good classification results can be obtained. For the wetland tidal flat information extraction, This method can show great potential for application with the development of multivariate remote sensing. The influence of water level changes on wetland tidal flat can be detected and analyzed by remote sensing images. The water level elevation at two times of T1 and T2 are 3.34 m and 3.56 m, the water level changes 0.22 m, the river surface area at T1 is 1811.09 hm ~ 2, the river surface area at T2 is 2 092.34hm ~ 2, The area of the surface is 281.25 hm ~ 2, and the area of unaltered area is 1778.47 hm ~ 2, accounting for 98.20% of the total area of T1. At T1, the surface area of river surface is 244.91 hm ~ 2, The total land area of 34.60%.