Analysis of hyperspectral signatures by double Weibull functions

作者：Schmidt Kai^*

来源：Photogrammetrie Fernerkundung Geoinformation, 2011, (5): 349-359.

DOI：10.1127/1432-8364/2011/0092

• Hyperspectral signatures
• Weibull function
• Parameter estimation
• Discriminant analysis

Analysis of hyperspectral signatures by double Weibull functions. It is considered as an advantage for the analysis of hyperspectral signatures if the complete information of a spectral sensor is taken into account. Therefore a new algorithm based on additive double Weibull functions is introduced. The model is applicable to the wavelength range from visible light (VIS) to near infrared (NIR) and up to short wave infrared (SWIR). It includes sufficient accuracy (R(2) = 0.99) and reduces the complex sensor information to a few model parameters. The resulting parameter vector characterises the object and is open for statistical analysis. As an example the diagnosis of plant diseases is used for the introduction of a two step procedure: In the first step the model is fitted to the signature data, taken from healthy and infected plants. In the second step the resulting parameter vectors are analysed by a discriminant analysis. The introduced procedure contains a high discriminatory power and is suitable for object classification and the distinction of virtually similar spectral profiles.

• 出版日期2011