机载DEM特征点提取及压缩方法研究

摘要

To make full use of the physiognomy characteristics and improve compression efficiency when compressing onboard Digital Elevation Model ( DEM), a new compressing method based on Radial Basis Function (RBF) neural network was proposed. Denoising process was carried out for the original DEM data to avoid that the noises be mistaken for feature points. Then an algorithm for extracting feature points including terrain ridge line and valleys was given. Finally,the RBF neural network was trained to implement DEM compression by using the extracted feature points. The experimental results demonstrate the effectiveness of presented method,which can enhance DEM compression effect for onboard use.%为了在机载数字高程(Diotal Elevation Model,DEM)数据压缩时充分利用地貌特征,进一步提高压缩效率,提出通过提取DEM特征点并使用径向基函数(Radial Basis Function,RBF)神经网络进行数据压缩的方法.首先对原始DEM数据进行去噪预处理,避免噪点成为特征点,给出了提取特征点及山脊线和山谷线具体算法；进而以地形特征点作为学习样本点采用RBF神经网络完成数据压缩任务.仿真结果表明,所提方法极大提高了DEM压缩效率,压缩效果满足机载要求.