Comparison of Hyperspectral Techniques for Urban Tree Diversity Classification

Charlotte Brabant,Alban Thomas,Emilien Alvarez-Vanhard,Gwénaël Morin,Kim Thanh Nguyen,Achour Laribi,Thomas Houet

doi:10.3390/rs11111269

Abstract

This research aims to assess the capabilities of Very High Spatial Resolution (VHSR) hyperspectral satellite data in order to discriminate urban tree diversity. Four dimension reduction methods and two classifiers are tested, using two learning methods and applied with four in situ sample datasets. An airborne HySpex image (408 bands/2 m) was acquired in July 2015 from which prototypal spaceborne hyperspectral images (named HYPXIM) at 4 m and 8 m and a multispectral Sentinel2 image at 10 m have been simulated for the purpose of this study. A comparison is made using these methods and datasets. The influence of dimension reduction methods is assessed on hyperspectral (HySpex and HYPXIM) and Sentinel2 datasets. The influence of conventional classifiers (Support Vector Machine –SVM– and Random Forest –RF–) and learning methods is evaluated on all image datasets (reduced and non-reduced hyperspectral and Sentinel2 datasets). Results show that HYPXIM 4 m and HySpex 2 m reduced by Minimum Noise Fraction (MNF) provide the greatest classification of 14 species using the SVM with an overall accuracy of 78.4% (±1.5) and a kappa index of agreement of 0.7. More generally, the learning methods have a stronger influence than classifiers, or even than dimensional reduction methods, on urban tree diversity classification. Prototypal HYPXIM images appear to present a great compromise (192 spectral bands/4 m resolution) for urban vegetation applications compared to HySpex or Sentinel2 images.

Highlights

Urbanization reflects important environmental transformations, such as a change in the energy balance [1], ecosystem loss or fragmentation, air, soil and water contamination, or even a loss of farming land, as well as an increase in the need for water and a reduction of biodiversity [2,3,4,5]
The feature selection method using Principal component analysis (PCA) provides a selection of 19 vegetation index for HySpex 2 m, 19 for HYPXIM 4 m (84.5%) and 17 for HYPXIM 8 m (85.89%)
The first 10 vegetation indices (VI) exhibiting the highest F-values are related to chlorophyllEmR ReaEnViInEWg that this biophysical property has the most discriminating p1o0teonf 1ti9al

Summary

Introduction

Urbanization reflects important environmental transformations, such as a change in the energy balance [1], ecosystem loss or fragmentation, air, soil and water contamination, or even a loss of farming land, as well as an increase in the need for water and a reduction of biodiversity [2,3,4,5]. Urban areas are extremely sensitive to global change as they have higher temperatures and atmospheric concentrations of pollutants or fine particles than rural areas [6] In this context, urban vegetation has several interests: it provides many services such as reducing the urban heat island by providing shade, the evapotranspiration and the photosynthesis of the trees [7,8,9] and improving air quality or even urban biodiverse habitats [8]. Urban vegetation presents drawbacks such as the exacerbation of allergies due to pollen, and the provision of habitats for certain bird species considered harmful by the population These services and disservices depend on the plant species, their location and structure [6,10,11,12]. Regarding these advantages and drawbacks, mapping urban tree diversity is an important issue; diversity is defined as the greatest number of species or families that can be distinguished

Objectives

Methods

Results

Discussion

Conclusion