论文部分内容阅读
边缘提取是图像处理与图像分析的基础。本文基于小波变换 ,研究了 SAR图像边缘提取的方法 ,从而为合成孔径雷达 (SAR)图像辅助导航系统奠定基础。由于 SAR图像受固有斑点噪声的影响 ,采用小波变换与固定阈值提取 SAR图像边缘特征时 ,将有大量虚假边缘产生 ,特别是在图像的亮区。为此 ,本文提出了采用具有恒定误警率的阈值 ,以消除斑点噪声的影响。同时 ,分别对两幅不同的图进行仿真 ,仿真结果表明 ,斑点噪声的影响被大大抑制 ,且由于阈值是由小波变换系数直接构成 ,能保证辅助导航对实时性的要求。
Edge extraction is the basis of image processing and image analysis. In this paper, the method of edge detection of SAR images is studied based on wavelet transform, which lays the foundation for Synthetic Aperture Radar (SAR) image assistant navigation system. Due to the inherent speckle noise of SAR images, when using the wavelet transform and fixed threshold to extract the edge features of SAR images, there will be a large number of false edges, especially in the bright areas of the image. To this end, this paper proposes a threshold with a constant false alarm rate to eliminate the effects of speckle noise. At the same time, two different graphs are simulated respectively. The simulation results show that the influence of speckle noise is greatly restrained, and because the threshold is directly constituted by the wavelet transform coefficients, it can guarantee the real-time requirements of the auxiliary navigation.