کنفرانس مهندسی برق ایران

صفحه اصلی / سی و یکمین کنفرانس بین المللی مهندسی برق

Kernel-Based Band Selection for Hyperspectral Image Classification

نویسندگان :

Mehdi Kamandar¹

1- دانشگاه تحصیلات تکمیلی صنعتی و فناوری پیشرفته

کلمات کلیدی :

Hyperspectral image classification،support vector machine،anisotropic Gaussian kernel،kernel-based band selection،l1 regularization،minibatch proximal gradient ascent

چکیده :

Hyperspectral image classification is a challenging task due to problems such as complex nonlinear structure of data, noisy and redundant spectral bands, and small-sample-size problem. The nonlinear support vector machine (SVM) is an excellent classifier to tackle these challenges. Selecting the most appropriate kernel and tuning its parameters for the data has a vital role in classification performance of the SVM. In this paper, an anisotropic Gaussian kernel is used. Anisotropic Gaussian kernel has a special scaling factor for each band. The scaling factors are tuned by maximizing a classifier independent separability measure with l1 regularization. The separability measure uses within-class and between-class dispersion information in the feature space. Some of the scaling factors will be zero after separability maximization due to l1 regularization, that is equivalent to removing corresponding bands from classification processes. Therefore, an embedded band selection is done to remove the destructive effect of irrelevant and redundant bands. The results show on average 5% classification accuracy improvement compared to an isotropic Gaussian kernel for classifying Indiana Pin Site hyperspectral image.