基于非下采样Shearlet和FCM变化检测的入侵植物遥感监测

Remote sensing monitoring of invasive plants based on change detection of Nonsubsampled Shearlet transform and FCM

尝试利用变化检测技术从遥感卫星影像上提取入侵植物蔓延的变化区域,实现对入侵植物的遥感监测。以预处理后的福建罗源湾2000年和2006年的Landsat TM/ETM+分布影像为数据源。首先采用ENVI软件对预处理过的入侵植物互花米草分布的遥感图像进行配准;然后对配准后的两时相遥感图像采用基于非下采样Shearlet变换和模糊C均值聚类(FCM)的算法进行变化检测,实现入侵植物蔓延区域的位置和边界的检测。实验结果表明:该算法可以自动识别入侵植物生长扩散的的面积和边界,能准确地对入侵植物进行遥感监测。

Try to use change detection techniques to extract the spread region of invasive plants from the remote sensing satellite images, in order to achieve remote sensing monitoring of invasive plants. The pretreated Landsat TM / ETM ＋ remote - sening data of 2000 and 2006 of Luoyuan Bay is a source data of this paper. First of all, the pretreated remote sensing images of the distribution of invasive plants are registered by ENVI software. Then change detection based on nonsubsampled Shearlet transform and fuzzy C - means clustering algorithm is implemented between two o＇ clock registered remote sensing images to detect the location and boundaries of the spread area of invasive plants. The experimental results show that： the area and boundary of growth and diffusion of invasive plants can be automatically identified by the proposed algorithm. The invasive plants can be accurately monitored by remote sensing.