V.M. Marshall, M.M. Lewis and B. Ostendorfhttp://www.journals.elsevier.com/journal-of-arid-environments

Hyperspectral sensing, measuring reflectance over visible to shortwave infrared wavelengths, has enabled the classification and mapping of vegetation at a range of taxonomic scales, often down to the species level. Classification with hyperspectral measurements, acquired by narrow band spectroradiometers or imaging sensors, has generally required some form of spectral feature selection to reduce the dimensionality of the data to a level suitable for the construction of a classification model. Despite the large number of hyperspectral plant classification studies, an in-depth review of feature selection methods and resultant waveband selections has not yet been performed. Here, we present a review of the last 22 years of hyperspectral vegetation classification literature that evaluates the overall waveband selection frequency, waveband selection frequency variation by taxonomic, structural, or functional group, and the influence of feature selection choice by comparing such methods as stepwise discriminant analysis (SDA), support vector machines (SVM), and random forests (RF). This review determined that all characteristics of hyperspectral plant studies influence the wavebands selected for classification. This includes the taxonomic, structural, and functional groups of the target samples, the methods, and scale at which hyperspectral measurements are recorded, as well as the feature selection method used. Furthermore, these influences do not appear to be consistent. Moreover, the considerable variability in waveband selection caused by the feature selectors effectively masks the analysis of any variability between studies related to plant groupings. Additionally, questions are raised about the suitability of SDA as a feature selection method, with it producing waveband selections at odds with the other feature selectors. Caution is recommended when choosing a feature selector for hyperspectral plant classification: We recommend multiple methods being performed. The resultant sets of selected spectral features can either be evaluated individually by multiple classification models or combined as an ensemble for evaluation by a single classifier. Additionally, we suggest caution when relying upon waveband recommendations from the literature to guide waveband selections or classifications for new plant discrimination applications, as such recommendations appear to be weakly generalizable between studies.

Native vegetation around the world is under threat from historical and ongoing clearance, overgrazing, invasive species, increasing soil and water salinity, altered fire regimes, poor land management and other factors, resulting in a degradation of natural ecosystem services. Consequently, maintaining and improving native vegetation condition is a target frequently adopted by natural resource managers and government agencies world-wide. Adequate monitoring of vegetation condition remains a prerequisite for environmental decision-making and for tracking progress towards management goals. Throughout we consider vegetation condition to include the compositional, structural and functional attributes of vegetation relative to undisturbed vegetation of the same type.Site-based methods have long been used to assess compositional, structural and functional attributes as indicators of vegetation condition, and these methods continue to be used widely today. With developing technologies, remote sensing methods are being employed increasingly for monitoring a range of remotely detectable properties of vegetation, and there is now a growing demand to explicitly integrate the two approaches for mapping and monitoring vegetation condition across a range of scales.Here we review the attributes of vegetation identified as important for monitoring vegetation condition, those indicators that are best measured using traditional site-based methods and those that are more readily detectable using remote sensing methods, including their application in operational programmes within Australia. Further to this we review recent literature on the integration of the two approaches for monitoring indicators of vegetation condition.We find that remote sensing methods have the advantage of offering broad scale automated and repeatable methods for monitoring indicators of vegetation condition, but when combined with detailed ecological site-based data, together can improve monitoring for answering ecological questions across a range of scales. Further work, however, is required to effectively integrate the two approaches for mapping and monitoring vegetation condition