This paper presents a method for hyperspectral image classification that uses support vector data description (SVDD) with the Gaussian kernel function. SVDD has been a popular machine learning technique for single-class classification, but selecting the proper Gaussian kernel bandwidth to achieve the best classification performance is always a challenging problem. This paper proposes a new automatic, unsupervised Gaussian kernel bandwidth selection approach which is used with a multiclass SVDD classification scheme. The performance of the multiclass SVDD classification scheme is evaluated on three frequently used hyperspectral data sets, and preliminary results show that the proposed method can achieve better performance than published results on these data sets.
Proc. SPIE. 8048, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XVII
KEYWORDS: Target detection, Signal to noise ratio, Hyperspectral imaging, Detection and tracking algorithms, Sensors, Reflectivity, Transform theory, Image classification, Heads up displays, Signal detection
This paper describes a novel approach for the detection and classification of man-made objects using discriminating
features derived from higher-order spectra (HOS), defined in terms of higher-order moments of hyperspectral-signals.
Many existing hyperspectral analysis techniques are based on linearity assumptions. However, recent research suggests
that significant nonlinearity arises due to multipath scatter, as well as spatially varying atmospheric water vapor
concentrations. Higher-order spectra characterize subtle complex nonlinear dependencies in spectral phenomenology of
objects in hyperspectral data and are insensitive to additive Gaussian noise. By exploiting these HOS properties, we have
devised a robust method for classifying man-made objects from hyerspectral signatures despite the presence of strong
background noise, confusers with spectrally similar signatures and variable signal-to-noise ratios. We tested
classification performance hyperspectral imagery collected from several different sensor platforms and compared our
algorithm with conventional classifiers based on linear models. Our experimental results demonstrate that our HOS
algorithm produces significant reductions in false alarms. Furthermore, when HOS-based features were combined with
standard features derived from spectral properties, the overall classification accuracy is substantially improved.
Proc. SPIE. 5089, Detection and Remediation Technologies for Mines and Minelike Targets VIII
KEYWORDS: Statistical analysis, Detection and tracking algorithms, Sensors, Metals, Algorithm development, Signal detection, Ground penetrating radar, Land mines, General packet radio service, Data fusion
Numerous detection algorithms, using various sensor modalities, have been developed for the detection of mines in cluttered and noisy backgrounds. The performance for each detection algorithm is typically reported in terms of the Receiver Operating Characteristic (ROC), which is a plot of the probability of detection versus false alarm as a function of the threshold setting on the output decision variable of each algorithm. In this paper we present multi-sensor decision fusion algorithms that combine the local decisions of existing detection algorithms for different sensors. This offers, in certain situations, an expedient, attractive and much simpler alternative to "starting over" with the redesign of a new algorithm which fuses multiple sensors at the data level. The goal in our multi-sensor decision fusion approach is to exploit complimentary strengths of existing multi-sensor algorithms so as to achieve performance (ROC) that exceeds the performance of any sensor algorithm operating in isolation. Our approach to multi-sensor decision fusion is based on optimal signal detection theory, using the likelihood ratio. We consider the optimal fusion of local decisions for two sensors, GPR (ground penetrating radar) and MD (metal detector). A new robust algorithm for decision fusion is presented that addresses the problem that the statistics of the training data is not likely to exactly match the statistics of the test data. ROC's are presented and compared for real data.