Tree Species Classification in a Highly Diverse Subtropical Forest Integrating UAV-Based Photogrammetric Point Cloud and Hyperspectral Data

Camile Sothe,Antonio Maria Garcia Tommaselli,Cláudia Maria De Almeida,Marcos Benedito Schimalski,Veraldo Liesenberg,Carla Luciane Lima,Gabriela Takahashi Miyoshi,Michele Dalponte

doi:10.3390/rs11111338

Abstract

The use of remote sensing data for tree species classification in tropical forests is still a challenging task, due to their high floristic and spectral diversity. In this sense, novel sensors on board of unmanned aerial vehicle (UAV) platforms are a rapidly evolving technology that provides new possibilities for tropical tree species mapping. Besides the acquisition of high spatial and spectral resolution images, UAV-hyperspectral cameras operating in frame format enable to produce 3D hyperspectral point clouds. This study investigated the use of UAV-acquired hyperspectral images and UAV-photogrammetric point cloud (PPC) for classification of 12 major tree species in a subtropical forest fragment in Southern Brazil. Different datasets containing hyperspectral visible/near-infrared (VNIR) bands, PPC features, canopy height model (CHM), and other features extracted from hyperspectral data (i.e., texture, vegetation indices-VIs, and minimum noise fraction-MNF) were tested using a support vector machine (SVM) classifier. The results showed that the use of VNIR hyperspectral bands alone reached an overall accuracy (OA) of 57% (Kappa index of 0.53). Adding PPC features to the VNIR hyperspectral bands increased the OA by 11%. The best result was achieved combining VNIR bands, PPC features, CHM, and VIs (OA of 72.4% and Kappa index of 0.70). When only the CHM was added to VNIR bands, the OA increased by 4.2%. Among the hyperspectral features, besides all the VNIR bands and the two VIs (NDVI and PSSR), the first four MNF features and the textural mean of 565 and 679 nm spectral bands were pointed out as more important to discriminate the tree species according to Jeffries–Matusita (JM) distance. The SVM method proved to be a good classifier for the tree species recognition task, even in the presence of a high number of classes and a small dataset.

Highlights

Tropical forests are among the most complex ecosystems on Earth, hosting an overwhelming proportion of global tree diversity: approximately 53,000 tree species in contrast to only 124 across temperate Europe [1], they play a crucial role in biodiversity conservation and in ecological dynamics at global scale [2].According to Viana and Tabanez [3], the Atlantic Rain Forest (Mata Atlântica) is one of the most endangered tropical biomes in the world, as it is reduced to less than 16% of its original area [4]
The first work involving the investigation of unmanned aerial vehicle (UAV)-based photogrammetry and hyperspectral imaging in individual tree detection and tree species classification was made by Nevalainen et al [23], in which they tested features extracted from UAV hyperspectral data and from photogrammetric point cloud (PPC) to classify tree species in a boreal forest, achieving 95% of overall accuracy (OA)
This study investigated the ability of UAV-based photogrammetry and hyperspectral imaging for tree species classification in a subtropical forest area

Summary

Introduction

Tropical forests are among the most complex ecosystems on Earth, hosting an overwhelming proportion of global tree diversity: approximately 53,000 tree species in contrast to only 124 across temperate Europe [1], they play a crucial role in biodiversity conservation and in ecological dynamics at global scale [2].According to Viana and Tabanez [3], the Atlantic Rain Forest (Mata Atlântica) is one of the most endangered tropical biomes in the world, as it is reduced to less than 16% of its original area [4]. 1544 plant species [6] and 380 animal species [7] are endangered in this biome, and they represent the equivalent of 60% of the threatened species for both flora and fauna in Brazil [8] Despite these threats, the Atlantic Rain Forest and its associated ecosystems (sandbanks and mangroves) are still rich in terms of biodiversity, containing high rates of endemism and species diversity, even greater than that observed in the Amazon Forest [9]. Due to the importance of this biome, it is crucial to build a reliable tree species mapping system for several applications, such as resource management, biodiversity assessment, ecosystem services assessment and conservation [10] In this respect, remote sensing data represents an efficient and potentially economical way of inventorying forest resources and mapping tree species [11,12]. Most of the aforementioned studies reported some common issues that hamper tree species classification in tropical forests, such as high number of species with similar spectral responses, irregularly stratified canopy, overlap between canopies leading to the absence of clear boundaries between individual trees, and the presence of dominant and minority classes, resulting in an imbalanced training data set in which only a small number of samples are available for the less frequently occurring tree species

Objectives

Methods

Results

Discussion

Conclusion