Across Date Species Detection Using Airborne Imaging Spectroscopy

Abstract

Imaging spectroscopy is a promising tool for airborne tree species recognition in hyper-diverse tropical canopies. However, its widespread application is limited by the signal sensitivity to acquisition parameters, which may require new training data in every new area of application. This study explores how various pre-processing steps may improve species discrimination and species recognition under different operational settings. In the first experiment, a classifier was trained and applied on imaging spectroscopy data acquired on a single date, while in a second experiment, the classifier was trained on data from one date and applied to species identification on data from a different date. A radiative transfer model based on atmospheric compensation was applied with special focus on the automatic retrieval of aerosol amounts. The impact of spatial or spectral filtering and normalisation was explored as an alternative to atmospheric correction. A pixel-wise classification was performed with a linear discriminant analysis trained on individual tree crowns identified at the species level. Tree species were then identified at the crown scale based on a majority vote rule. Atmospheric corrections did not outperform simple statistical processing (i.e., filtering and normalisation) when training and testing sets were taken from the same flight date. However, atmospheric corrections became necessary for reliable species recognition when different dates were considered. Shadow masking improved species classification results in all cases. Single date classification rate was 83.9% for 1297 crowns of 20 tropical species. The loss of mean accuracy observed when using training data from one date to identify species at another date in the same area was limited to 10% when atmospheric correction was applied.

A Convolutional Neural Network Classifier Identifies Tree Species in Mixed-Conifer Forest from Hyperspectral Imagery

“…Our results were achieved under ideal circumstances, contained in a single strip of imagery with a uniform solar illumination angle, using isolated, tall canopies in a temperate forest ecosystem with seven species. Atmospheric correction radiative transfer models can be applied to help mitigate issues caused by multi-date imagery collection for the purpose of species identification [7]. Other efforts to classify individual tree species from hyperspectral imagery using machine learning have yielded accuracies that vary widely depending on the forest type, the remote sensing platform, and the classification methods used.…”

“…Automated species mapping of forest trees using remote sensing data has long been a goal of remote sensing and forest ecologists [1][2][3][4][5][6][7][8][9][10][11]. Conducting remote inventories of forest species composition using an imaging platform instead of field surveys would save time, money, and support analysis of species composition over vast spatial extents [12][13][14][15].…”

“…Performing this classification of adjacent flight lines where no training data exists is the logical next step for this research. Detecting species across dates using hyperspectral imagery can be improved when a radiative transfer model based on atmospheric compensation is applied, but accounting for atmospheric corrections remains a challenge for species classification across dates and conditions, and having spatially extensive training data is important in such cases [7].…”

A Comparative Assessment of the Performance of Individual Tree Crowns Delineation Algorithms from ALS Data in Tropical Forests

“…ALS data can also be merged with other remote sensing techniques to take advantage of different sensor characteristics. For instance, ITC delineation from ALS can be merged with hyperspectral data to identify tree species from the spectral information at the crown level, which gives better results than at the pixel level [15][16][17][18][19]. However, this implies a good ITC delineation, especially of the upper canopy crowns for which the hyperspectral information is available.…”

Determinação das Áreas Atingidas por Queimadas em Bacias Hidrográficas por meio do Índice De Queimada (NBR), Estudo de Caso da Bacia do Rio Piracicaba-MG

“…Apesar da disponibilidade de dados de sensoriamento remoto e eficiência das técnicas de aplicação é recomendado o pré tratamento das cenas a serem utilizadas (Chen et al, 2005;Nia et al, 2015) por meio da realização de correção radiométrica para minimizar inconsistência radiométrica entre os alvos e divergências entre as características do sensor, condições atmosféricas e angulação solar e de visada (Chen et al, 2005;Laybros et al, 2019), visto que os efeitos atmosféricos modificam as medidas radiométricas do sensor (Bernardo et al, 2016;Keukelaere et al, 2018). Essa técnica consiste na conversão de número digital (ND) ou níveis de cinza de cada pixel em radiância espectral (Chen et al, 2005;Thompson et al, 2018), utiliza os coeficientes de redimensionamento radiométrico fornecidos no arquivo metadados disponibilizados simultaneamente à disponibilização das bandas (arquivo MTL.txt no caso das imagens LandSat 8) e permite minimizar os erros causados pela atmosfera, ângulo solar e geometria de visada (Chen et al, 2005).…”