摘要
干涉图像滤波一直是InSAR图像处理的关键步骤。针对经典滤波方法的不足,提出一种基于噪声分布模型的滤波新方法。该方法以干涉图像噪声分布模型为基础,首先对图像做噪声和有用信息判断,若判断为噪声进行滤除,否则保留信息。最后选取实际SAR数据进行滤波实验,与经典的滤波方法进行对比分析,并进行定量评价。实验结果表明,基于噪声分布模型的滤波方法,在去除残差点和边缘保持指数方面明显优于经典方法,在等效视数、相位标准差、斑点指数和平滑指数方面与经典方法相当。因此,本文提出的滤波方法在去除噪声的同时,最大程度的保持了图像信息的完整性,无疑是一种有效的新滤波方法。
Interference image filter has been a key step in the InSAR image processing. Aiming at the shortcoming of the classical filter methods, a new method is proposed based on the noise distribution model. Firstly, separating noises and useful information, and if it is a noise filter it out, otherwise keep it. At last, the actual SAR data is used in the filtering experiment and quantitative evaluation at the same time. The experiment results show that removing residual points pow- erfully and maintaining the image edge detail are good by the new method. And, other indexes （ENL, RMSE, SIR, FI） of this method are almost as same effect as those of classical method. Therefore, When using this method to removing noises, and the integrity of the image information is maintained in the greatest degree at the same time and it is proved that it is an effective new filter method.
出处
《北京测绘》
2014年第2期1-4,16,共5页
Beijing Surveying and Mapping
