Abstract The extraction of information from point cloud is usually done after the application of classification methods based on the geometric characteristics of the objects. However, the classification of photogrammetric point clouds can be carried out using radiometric information combined with geometric information to minimize possible classification issues. With this in mind, this work proposes an approach to the classification of photogrammetric point cloud, generated by correspondence of aerial images acquired by Remotely Piloted Aircraft System (RPAS). The proposed approach for classifying photogrammetric point clouds consists of a pixel-supervised classification method, based on a decision tree. To achieve this, three data sets were used, one to define which attributes allow discrimination between the classes and the definition of the thresholds. Initially, several attributes were extracted based on a training sample. The average and standard deviation values for the attributes of each class extracted were used to guide the decision tree definition. The defined decision tree was applied to the other two point clouds to validate the approach and for thematic accuracy assessment. The quantitative analyses of the classifications based on kappa coefficient of agreement, applied to both validation areas, reached values higher than 0.938.

O cultivo da cana-de-açúcar no Brasil, embora assistido por técnicas modernas de plantio, é alvo constante de parasitas do sistema radicular. Por registrar seletivamente o fluxo espectral da radiação eletromagnética refletida pela vegetação, o sensoriamento remoto tornou-se uma poderosa ferramenta na detecção das plantas infectadas por patógenos do solo. Com o objetivo de caracterizar espectralmente a cana-de-açúcar sadia e infectada por nematoides e pela larva do besouro Migdolus fryanus, foram tomadas medidas radiométricas in situ e geradas curvas hiperespectrais de plantas sadias e infectadas. Técnicas específicas de análise espectral, como a determinação da posição da borda do vermelho limítrofe (Red Edge Position Determination - REPD) e diferentes índices espectrais foram avaliados para discriminar as três ocorrências. As curvas de reflectância mostraram diferenças em magnitude principalmente nos comprimentos de onda do vermelho e infravermelho próximo e, assim como a determinação do REP e os índices de clorofila b, NDVI, MCARI e TCARI, permitiram distinguir apenas entre plantas sadias e infectadas. As razões espectrais sensíveis aos pigmentos clorofila a e carotenoides, porém, discriminaram as três ocorrências, inclusive plantas infectadas por nematoides e Migdolus fryanus. A melhor discriminação foi obtida com o índice de carotenoides, um pigmento fortemente relacionado com estresse da planta

Despite all technological support used in sugarcane cultivation in Brazil, this planting is still subject to the action of root pathogens. Due to its ability to selectively record the spectral flux of electromagnetic radiation reflected by vegetation, remote sensing is a powerful tool in detecting plants infected by pests and parasites. The objective of this work was to characterize the spectral response of healthy sugarcane and its plants infected by nematodes and Migdolus fryanus beetle larva. In situ radiometric measurements were performed and specific spectral analysis techniques, such as Red Edge Position Determination (REPD) and different vegetation indices were evaluated to discriminate these occurrences. The reflectance curves generated showed differences in magnitude mainly at red and near-infrared wavelengths and, as well as REPD, chlorophyll b index, NDVI, MCARI and TCARI, were able to discriminate only healthy and infected plants. The spectral ratios sensitive to chlorophyll a and carotenoids pigments, however, discriminated the three occurrences, including nematode-infected plants and Migdolus fryanus. The best discrimination was obtained by the index of carotenoids, a pigment whose content is related to plant stress.