一种多尺度高光谱影像小波分形维特征计算方法

提出了一种基于高光谱曲线小波分形测度的高光谱影像多尺度分形维特征分析方法。对高光谱影像的光谱响应曲线的小波域高频和低频系数统计特性、分形特征进行了分析,提出以小波低频分形维表征原始光谱曲线分形特征,以小波系数高频分形维表征高光谱细节特征方法,设计了基于高光谱曲线小波分形维的多尺度特征计算算法,实验结果表明,小波分形维值可有效表征丰富的光谱特征,可用于高光谱影像特征提取和分类。

Multi-scale Spectral Feature Analysis Based on Wavelet Fractal Measurement of Hyperspectral Images

In this paper,a multi-scale spectral feature analysis based on wavelet fractal measurement is proposed for hyperspectral image.Statistics and fractal characteristic of the high-frequency and low-frequency coefficients from spectral curve wavelet decomposition are discussed.The fractal dimension of low frequency coefficients is proposed to represent the fractal feature of original spectrum curve and the fractal dimension of high-frequency coefficients can be taken as spectrum detail feature.A multi-scale spectrum feature calculation algorithm of hyperspectral image based on wavelet fractal measurement is designed.Experiments of proposed multi-spectrum feature analysis of hyperspectral images shown,that the wavelet fractal dimension can effectively represent the rich spectrum characteristics,which can be used for the feature extraction and classification of hyperspectral images.