摘要
将分形几何学用于低对比度前视红外图像的分割,分析研究了自然场景与人造物体的分形特性随尺度的不同变化,提出并构造一种新的多尺度分形特征矢量,通过K-均值算法在特征空间中对其进行类聚。
Fractal geometry was used for the segmentation of low contrast FLIR image.The variation of fractal feature in different scales of natural scenes and man made objects was analyzed.A new multiscale fractal feature vector was proposed and constructed.The accurate segmentation of tank target was realized by clustering the multiscale fractal feature vector in the feature space.
出处
《红外与毫米波学报》
SCIE
EI
CAS
CSCD
北大核心
1996年第3期183-188,共6页
Journal of Infrared and Millimeter Waves
基金
国家自然科学基金
关键词
前视红外
图像分割
红外成像
图像检测
FLIR,fractal,multiscale analysis,multiresolution pyramid decomposition,image segmentation.