摘要
针对从全色航空影像中进行植被区域提取的随机性和复杂性,阐述了一种基于小波的Laws纹理测度进行植被提取的新算法,它的特点是先用小波变换将图像变换到不同的尺度层上,然后再在多尺度层上提取Laws纹理测度,形成植被区域的特征.与传统的植被提取方法比,它用到了不同频率上纹理的Laws信息,从而更准确的刻画了植被区域的纹理特征,试验结果表明:基于小波变换的Laws纹理测度对植被有较好的分割效果.
Aiming at the randomicity and complexity of Vegetation area extraction from black and white aero-photo, this paper points out a new algorithms that extraction of the vegetation area in wavelet pyramid-structured and Laws texture measure, its characteristic is transform the image to different scale layer firstly, then extracts the Laws texture measure from different scale layer, forming the characteristic of vegetation area. Compare with the traditional method, it use the features of different scale layer, so it can describe the characteristic of vegetation texture more accurately. Experiences show that the method based on Wavelet and Laws texture measure is reasonable and feasible to the vegetation area.
出处
《大学数学》
北大核心
2005年第2期5-9,共5页
College Mathematics
