Nonlinear Classification of Multispectral Imagery Using Representation-Based Classifiers

Abstract

This paper investigates representation-based classification for multispectral imagery. Due to small spectral dimension, the performance of classification may be limited, and, in general, it is difficult to discriminate different classes with multispectral imagery. Nonlinear band generation method with explicit functions is proposed to use which can provide additional spectral information for multispectral image classification. Specifically, we propose the simple band ratio function, which can yield better performance than the nonlinear kernel method with implicit mapping function. Two representation-based classifiers-i.e., sparse representation classifier (SRC) and nearest regularized subspace (NRS) method-are evaluated on the nonlinearly generated datasets. Experimental results demonstrate that this dimensionality-expansion approach can outperform the traditional kernel method in terms of high classification accuracy and low computational cost when classifying multispectral imagery.

Robust Capsule Network Based on Maximum Correntropy Criterion for Hyperspectral Image Classification

“…To overcome this problem, linear discriminant analysis (LDA) [14] was proposed, which is further improved by local Fisher's discriminant analysis (LFDA) [15]. Considering the large spatial variability of spectral signature introduced by light-scattering mechanisms, the hyperspectral data usually presents a nonlinear characteristic [16], [17]. In this case, the aforementioned FE methods based on linear transformation may not be adequate for subsequent analysis.…”

Multiple feature regularized kernel for hyperspectral imagery classification

“…Hyperspectral images (HSIs) contain hundreds of contiguous spectral bands ranging from visible to infrared bands, which can provide rich spectral information [1]. Compared with multispectral images with only a few spectral bands, they offer an advantage in terms of classification and detection [2]. Due to the rich spectral bands, HSIs have a variety of applications including environmental pollution control, agriculture precision farming, mineral exploration, etc [3][4][5][6].…”

Hyperspectral image clustering via sparse dictionary‐based anchored regression

“…Two main categories of methods for HSI classification can be divided into supervised and unsupervised methods depending on whether or not labelled information is available. For supervised classification, the popular choice of classifiers is support vector machines [1], maximum‐likelihood classifier [2], neural network classifier [3, 4], and representation‐based classifier [5–8]. For unsupervised classification, the existing major clustering algorithms for HSIs include density‐based clustering methods, graph‐based models method, and subspace‐based models method.…”