【摘要】 随着数字化调查技术的发展,国土资源管理对土地利用动态遥感监测提出了更高的要求,目前主要采用的多光谱数据由于受光谱分辨率限制以及“同谱异物,同物异谱”现象的影响,难以满足管理需要。成像光谱数据具有较高光谱分辨率,在类别细分方面具有一定的优势,在当前土地利用动态遥感监测中具有一定的应用潜质。该文针对成像光谱数据特点,探索了与成像光谱数据相适应的土地利用动态遥感监测方法,提出了异常光谱检测法,该方法在试验区应用中取得了良好效果。更多还原

【Abstract】 Land use change detecting is very important in land resource management. With the development of digital land survey technology, higher accuracy of land use and land cover classification is needed, especially of main land use type such as arable land, forest, orchard etc. However, normal multi-spectral remote sensing data can not meet the requirement. Many experiments show that high spectral resolution remote sensing data may solve this problem. In this paper, a study on extraction method of land use change type based on spectral feature using high spectral resolution remote sensing data is conducted. Firstly, a comparison among different image fusion and detecting methods are made. Then an analysis on these methods is performed to find which is suitable for land change detecting and which is unsuitable. Based on the analysis and special feature of high spectral resolution remote sensing data, a method called anomaly spectral feature detecting is proposed,which based on the assumption that spectral feature will be anomaly in land use type changed area when fusion different temporal remote sensing data. The method has applied to experiment site successfully. The result shows that high spectral resolution remote sensing data combined with high spatial resolution data can improve land use change accuracy to 90% and the error is less than 10%, which is meet needs of present land resource management in China.更多还原