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.
The natural regeneration process is essential for forest maintenance since it is critical for establishing new tree individuals. This study aimed to improve the understanding of the regenerative component dynamics of Araucaria Forests in Southern Brazil. We investigated the effects of climate, light, tree component structure and anthropogenic disturbance on tree species regeneration. Regenerating communities from six different fragments in forest remnants of the "Planalto Sul Catarinense" region was evaluated in permanent plots two years after the first inventory. The following demographic rates were determined: recruitment, mortality, net change in the number of individuals and the changes to both upper and lower height classes. The following variables were measured in each fragment: altitude, climatic variables, light environment, tree component density and cattle presence. Association between dynamics rates, regenerating species abundance and explanatory variables was verified by the fourth-corner and RLQ methods. A total of 4,379 and 5,268 individuals were sampled for the first and second inventories, respectively, with recruitment rate (21 % yr -1 ) higher than mortality rate (13 % yr -1 ). The dynamics pattern of the fragment with greater presence of cattle stood out for the intense height increase of regenerating species caused by the presence of fast growth and light-demanding species.Natural regeneration of forest remnants under study is facing a structuring process. The main conclusions of this study were: i) climate and altitude play a relevant role in defining floristic identity and ii) chronic disturbances may influence the definition of ecological strategies.
Resumo Objetivou-se investigar como a heterogeneidade ambiental influencia as variações espaciais dos atributos funcionais e a diversidade funcional do componente arbóreo em uma floresta nebular no sul do Brasil. Foram selecionadas as 19 espécies arbóreas mais abundantes identificadas em um inventário realizado em um fragmento florestal. As variáveis ambientais utilizadas foram obtidas de um trabalho prévio realizado na mesma área. Dessas espécies, foram obtidas a densidade da madeira, área foliar, área foliar específica, altura máxima potencial, deciduidade e síndrome de dispersão. Foram determinados os valores médios dos atributos ponderados para a comunidade e de diversidade funcional. Os dados ambientais foram ordenados por meio de Análise de Componentes Principais e modelos lineares foram ajustados relacionando os Componentes Principais significativos e os valores de atributos funcionais e de diversidade funcional. Enquanto os locais com maior fertilidade do solo favoreceram espécies com estratégias aquisitivas, representadas pela menor densidade da madeira e maior área foliar, e apresentaram maior diversidade funcional; os ambientes menos férteis favoreceram estratégias conservativas, representadas pela maior densidade da madeira e menor área foliar, e apresentaram uma menor diversidade funcional. Conclui-se que a floresta estudada apresentou o particionamento de habitats, em função de variações edáficas, por espécies com estratégias ecológicas distintas.
RESUMOObjetivou-se verificar as interações entre a configuração espacial da paisagem, a organização florístico-estrutural e as taxas demográficas do componente arbóreo em um sistema de fragmentos e corredores de Floresta com Araucárias em Lages, Santa Catarina. Para isso, foi elaborado um modelo conceitual das possíveis inter-relações, que foi avaliado pela técnica de Modelagem de Equações Estruturais. No ano de 2010, foram obtidos as métricas da paisagem (área, distância do vizinho mais próximo e relação borda e o interior da floresta) e os dados do primeiro inventário florestal. Foram alocadas parcelas permanentes em cinco fragmentos e corredor florestal, onde todos os indivíduos arbóreos com CAP (circunferência a altura do peito, medida a 1,30 do solo) igual ou superior a 15,7 cm foram identificados e mensurados. Em 2014 foi realizado o segundo inventário, com a inclusão de indivíduos recrutas, contagem de mortos e sobreviventes, e foram calculadas as taxas demográficas. Os dados foram analisados por meio da Análise de Componentes Principais (PCA), Análise de Coordenadas Principais (PCoA), Modelagem de Equações Estruturais e Modelos Lineares Generalizados (GLM). Os resultados demonstraram que a estrutura da paisagem (PCA 1) exerceu influência apenas sobre a organização florístico-estrutural do componente arbóreo, indicada pela distribuição preferencial de espécies arbóreas em função da intensidade da fragmentação. Por sua vez, as taxas demográficas (taxas de ganho e perda em área basal e de mortalidade) foram influenciadas por aspectos estruturais da vegetação (abundância e área basal). Conclui-se que existem variações florístico-estruturais associadas à configuração espacial dos fragmentos na paisagem e que as taxas demográficas apresentam relação com o estágio sucessional da floresta, sintetizado pelas variáveis estruturais de área basal e abundância. Palavras-chave: modelagem de equações estruturais; fragmentação florestal; Floresta Ombrófila Mista. ABSTRACTThe aim of this study was to investigate the interactions among the landscape spatial configuration, the floristic-structural organization and demographic rates of the tree component of a system of araucaria forest fragments, in Lages, Santa Catarina state. To do so, we developed a conceptual model of inter-relationship that was evaluated by Structural Equation Modeling. In 2010, the landscape metrics (area, distance from the nearest neighbor and edge-core ratio) and first vegetation inventory data were obtained. Permanent plots
We aimed to investigate the taxonomic and functional variations of tree component of Araucaria Forest (AF) areas located along an altitudinal gradient (700, 900 and 1,600 m asl), in the southern region of Brazil. The functional traits determined were leaf area, specific leaf area, wood density, maximum potential height and dispersal syndromes and deciduousness. The data were analyzed through a functional and taxonomic dissimilarity dendrograms, community-weighted mean trait values, parametric and nonparametric tests, and Principal Component Analysis. The largest floristic-structural similarity was observed between the lower altitude areas (700 and 900 m asl), whose Bray-Curtis distance was 0.63. The area at 700 m asl was characterized by a predominance of deciduous and semi-deciduous species, with a high number of selfand wind-dispersed species, whereas the area at 1,600 m asl exhibited a predominance of animal-dispersed and evergreen species. It was also observed that there were significant variations for leaf traits, basic wood density and maximum potential height. Over all altitudinal gradient, the ordinations indicated that there was no evidence of functional differentiation among dispersal and deciduousness groups. In conclusion, the evaluated Araucaria Forest areas presented high floristic-functional variation of the tree component along the altitudinal gradient.
Abstract. The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced feature extraction and classification methods. Different from the traditional feature extraction methods, that highly depend on user’s knowledge, the convolutional neural network (CNN)-based method can automatically learn and extract the spatial-related features layer by layer. However, in order to capture significant features of the data, the CNN classifier requires a large number of training samples, which are hardly available when dealing with tree species in tropical forests. This study investigated the following topics concerning the classification of 14 tree species in a subtropical forest area of Southern Brazil: i) the performance of the CNN method associated with a previous step to increase and balance the sample set (data augmentation) for tree species classification as compared to the conventional machine learning methods support vector machine (SVM) and random forest (RF) using the original training data; ii) the performance of the SVM and RF classifiers when associated with a data augmentation step and spatial features extracted from a CNN. Results showed that the CNN classifier outperformed the conventional SVM and RF classifiers, reaching an overall accuracy (OA) of 84.37% and Kappa of 0.82. The SVM and RF had a poor accuracy with the original spectral bands (OA 62.67% and 59.24%) but presented an increase between 14% and 21% in OA when associated with a data augmentation and spatial features extracted from a CNN.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
334 Leonard St
Brooklyn, NY 11211
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.