ABSTRACT Background: This study aimed to evaluate UAV images of Pinus taeda L. stands for classifying trees attacked by Sapajus nigritus in Southern Brazil. UAV images were acquired on March 2018, using a DJI Phantom Pro 4 over 18.73 hectares. We evaluated different band compositions and vegetation indices. Using photo interpretation based on the color of the crown and field measurements, the trees were manually labeled as not attacked, dead, and attacked. The classified trees were divided into training (75%) and validation (25%), considering three tree crown diameters (0.5, 1, and 1.5 m) and three region-oriented classification algorithms. The classification accuracy was assessed by overall accuracy and the kappa index. Results: A total of 3,773 trees were manually detected, of which 39% were attacked, 5% died and 56% were not attacked. The results also indicated that the best-chosen diameter was 0.5 meters, the best classifier algorithm was the SVM, and the highest accuracy was represented by the composition of the ExG index associated with the original spectral bands. Conclusion: We argue that the attacks can be monitored using UAV images and such results provide insights for forest management initiatives.

Resumo Este estudo realizou um mapeamento das áreas de pomares de macieira, considerando área plantada, localização, altitude, declividade e o uso de cobertura antigranizo, no município de São Joaquim (Planalto Sul Catarinense, Brasil). Foram empregadas imagens espectrais da plataforma orbital Sentinel-2 adquiridas em agosto de 2018, e o modelo digital de elevação aprimorado do Shuttle Radar Topography Mission(SRTM). Em aplicativo GIS configurado com o sistema de referência SIRGAS 2000,4 e projeção cartográfica UTM, foram adicionadas as imagens da constelação Sentinel-2 e os modelos digitais de elevação advindos da missão SRTM refinados mais recentemente com dados do sensor Phased Array type L-band Synthetic Aperture Radar (PALSAR). Todas as imagens foram reamostradas para a resolução espacial de 10 m. Os resultados foram validados com base em imagens disponíveis do Google Earth. Os resultados mostram que São Joaquim conta com área plantada de 7.974,80 ha de macieira, e destes 12% com sistema de cobertura antigranizo. A característica da região é de pequenos produtores em áreas de um a cinco ha. Um Percentual superior a 50% dos pomares encontra-se no intervalo entre 1.200 e 1.400 m de altitude. Com 45% dos pomares localizados em áreas com declividade entre 8 e 20%. Em relação à localização, os pomares estão concentrados principalmente no raio de até 20 km de distância em relação ao centro urbano de São Joaquim, onde se encontra quase a totalidade das indústrias e cooperativas para o seu acondicionamento, beneficiamento e logística. Este estudo demonstrou que dados orbitais do Sentinel-2 podem quantificar efetivamente a distribuição dos pomares de macieira, sendo uma alternativa viável e eficaz no levantamento das informações para o monitoramento agrícola. Desse modo, viabiliza o planejamento eficiente da produção de maçã, como a assistência técnica, a comercialização com os produtores e o escoamento da produção.

Abstract This study carried out a mapping procedure focusing on apple orchards considering the planted area, spatial location, altitude range, slope interval, and presence of anti-hail nets in the city of São Joaquim (Southern Santa Catarina Plateau, Brazil). Spectral images from the Sentinel-2 orbital platform acquired in August 2018 and an enhanced digital elevation model from the Shuttle Radar Topography Mission (SRTM) were used. In a GIS application configured with the SIRGAS 2000,4 reference system and UTM cartographic projection, Sentinel-2 constellation images and digital elevation models from the SRTM mission and more recently refined with sensor data Phased Array type L-band Synthetic Aperture Radar (PALSAR) were added. All images were resampled to a spatial resolution of 10m. The results were validated based on high spatial resolution images available from Google Earth. The results show that São Joaquim has a planted area of 7,974.80 ha, and only 12% use an anti-hail coverage system. The majority of the orchards range from one to five ha and belong to small producers. More than 50% of the orchards are between 1,200 and 1,400 m in altitude, with 45% of orchards located in areas with slopes between 8 to 20%. Interestingly, most of the orchards are concentrated in a radius of up to 20km from the urban center of São Joaquim, where industries and cooperatives are located for packaging, processing, and logistics. This study demonstrated that orbital data from Sentinel-2 can effectively quantify the distribution of apple orchards, being a viable and effective alternative for collecting information for agricultural monitoring. In this way, it enables efficient planning of apple production, such as technical assistance, marketing with producers, and production flow.

Abstract: This study investigates the potential use of close-range and low-cost terrestrial RGB imaging sensor for fruit detection in a high-density apple orchard of Fuji Suprema apple fruits (Malus domestica Borkh). The study area is a typical orchard located in a small holder farm in Santa Catarina’s Southern plateau (Brazil). Small holder farms in that state are responsible for more than 50% of Brazil’s apple fruit production. Traditional digital image processing approaches such as RGB color space conversion (e.g., rgb, HSV, CIE L*a*b*, OHTA[I 1 , I 2 , I 3 ]) were applied over several terrestrial RGB images to highlight information presented in the original dataset. Band combinations (e.g., rgb-r, HSV-h, Lab-a, I” 2 , I” 3 ) were also generated as additional parameters (C1, C2 and C3) for the fruit detection. After, optimal image binarization and segmentation, parameters were chosen to detect the fruits efficiently and the results were compared to both visual and in-situ fruit counting. Results show that some bands and combinations allowed hits above 75%, of which the following variables stood out as good predictors: rgb-r, Lab-a, I” 2 , I” 3 , and the combinations C2 and C3. The best band combination resulted from the use of Lab-a band and have identical results of commission, omission, and accuracy, being 5%, 25% and 75%, respectively. Fruit detection rate for Lab-a showed a 0.73 coefficient of determination (R2), and fruit recognition accuracy rate showed 0.96 R2. The proposed approach provides results with great applicability for small holder farms and may support local harvest prediction.

SUMMARY: This study aimed at modeling competition indices for individual trees of Araucaria angustifolia in natural forests. A total of 397 objetive trees were sampled in Lages and São Francisco de Paula municipalities. Through a non-linear regression analysis, 22 distance-independent and distance-dependent competition indices were modeled based on the diameter at breast height (d) and relative height (hrel: ratio between the total height of the objetive tree and the maximum potential height estimated for the same diameter). The adjusted equations of the indices according to d of the objetive tree achieved values of R² ≤ 0.546 (Syx% ≥ 15.0). The independent variable hrel improved the estimates, explaining up to 63 % of the variation in the data (Syx% ≥ 14.6). In general, competition indices - which consider in the calculation only the variables d of the objetive tree, d of the competitors, its horizontal distance and eventually the height variable- showed superior performance. Therefore, they are indicated for use in growth modeling of A. angustifolia, due to superior facility of application in natural forests. The developed equations showed to be good representations of the degree of competition at the level of the individual tree of A. angustifolia for the two evaluated sites, measuring only the diameter or including the height of the central tree for more accurate precision. The expansion of research to other sites along the distribution of the species will confirm the application of these equations to their full extent. In addition, this work contributes to the inclusion of competition indices in modeling of individual trees in uneven-aged mixed forests.

RESUMEN: El presente estudio modeló índices de competencia para árboles individuales de Araucaria angustifolia en bosque nativo. Fueron muestreados 397 árboles objetivo en los municipios de Lages y São Francisco de Paula. Mediante análisis de regresión no linear fueron modelados 22 índices de competencia independientes y dependientes de la distancia en función del diámetro a la altura del pecho y altura relativa (razón entre la altura total del árbol objetivo y la máxima altura potencial estimada para el mismo diámetro). Las ecuaciones ajustadas de los índices en función del diámetro del árbol objetivo alcanzaron valores de R² ≤ 0,546 (Syx% ≥ 15,0). La altura relativa mejoró las estimaciones, explicando hasta el 63 % de la variación de datos (Syx% ≥ 14,6). En general, los índices de competencia que consideraron solo los diámetros del árbol objetivo y de los competidores, sus distancias horizontales, y eventualmente la altura, presentaron desempeño superior, siendo así indicados para el modelado del crecimiento de A. Angustifolia, debido a la mayor facilidad de aplicación en bosque nativo. Las ecuaciones desarrolladas representaron bien el grado de competencia a nivel del árbol individual de A. angustifolia para los dos sitios evaluados, midiendo solo el diámetro, o incluyendo también la altura del árbol objetivo para una mayor precisión. La expansión de la investigación a otros sitios a lo largo de la distribución de las especies permitirá confirmar la aplicación de estas ecuaciones en toda su amplitud. Además, este trabajo contribuye a la inclusión de índices de competencia en el modelado de árboles individuales en bosques mixtos disetáneos.

ABSTRACT This study aimed to diagnose the environmental infractions that occurred in the northern region of Santa Catarina State (Southern Brazil). The environmental infractions were classified by category and by violation activities concerning the hydrographic regions of Canoinhas Plateau and Baixada Norte from 2014 to 2018 to assess the growth and causes of a diagnosis regarding the environmental damages occurrences. Results indicated that the major illegal environmental practices registered in both region are related to environmental violations regarding the damage caused in native vegetation, and these damages are caused without authorization of the competent environmental authority. In addition, it was also important to note the activity of maintaining birds in captivity and of functioning of potentially pollutant activities or establishments that do not have a license of the competent regional environmental agencies. Therefore, efforts from the Environmental Policy should be directed to implement a forest monitoring system using remote sensing technology apart from starting a stronger cooperation with municipal authorities from Santa Catarina State by promoting Environmental Monitoring Programs (EMP).

ABSTRACT Accurate forest inventory is of great economic importance to optimize the entire supply chain management in pulp and paper companies. The aim of this study was to estimate stand dominate and mean heights (HD and HM) and tree density (TD) of Pinus taeda plantations located in South Brazil using in-situ measurements, airborne Light Detection and Ranging (LiDAR) data and the non- k-nearest neighbor (k-NN) imputation. Forest inventory attributes and LiDAR derived metrics were calculated at 53 regular sample plots and we used imputation models to retrieve the forest attributes at plot and landscape-levels. The best LiDAR-derived metrics to predict HD, HM and TD were H99TH, HSD, SKE and HMIN. The Imputation model using the selected metrics was more effective for retrieving height than tree density. The model coefficients of determination (adj.R2) and a root mean squared difference (RMSD) for HD, HM and TD were 0.90, 0.94, 0.38m and 6.99, 5.70, 12.92%, respectively. Our results show that LiDAR and k-NN imputation can be used to predict stand heights with high accuracy in Pinus taeda. However, furthers studies need to be realized to improve the accuracy prediction of TD and to evaluate and compare the cost of acquisition and processing of LiDAR data against the conventional inventory procedures.

ABSTRACT The aim of this study was to evaluate the performance of four ground filtering algorithms to generate digital terrain models (DTMs) from airborne light detection and ranging (LiDAR) data. The study area is a forest environment located in Washington state, USA with distinct classes of land use and land cover (e.g., shrubland, grassland, bare soil, and three forest types according to tree density and silvicultural interventions: closed-canopy forest, intermediate-canopy forest, and open-canopy forest). The following four ground filtering algorithms were assessed: Weighted Linear Least Squares (WLS), Multi-scale Curvature Classification (MCC), Progressive Morphological Filter (PMF), and Progressive Triangulated Irregular Network (PTIN). The four algorithms performed well across the land cover, with the PMF yielding the least number of points classified as ground. Statistical differences between the pairs of DTMs were small, except for the PMF due to the highest errors. Because the forestry sector requires constant updating of topographical maps, open-source ground filtering algorithms, such as WLS and MCC, performed very well on planted forest environments. However, the performance of such filters should also be evaluated over complex native forest environments.

Resumo: A classificação remota dos diferentes estádios sucessionais da vegetação ainda constitui um desafio devido à similaridade espectral destas classes. Este artigo tem o objetivo de avaliar o desempenho de imagens Landsat-8 e RapidEye para a classificação do estádio sucessional da vegetação em um fragmento de Floresta Ombrófila Mista, localizado no Parque Nacional de São Joaquim- SC. Para isto, três grupos de variáveis gerados a partir de cada imagem foram avaliados, sendo: (1) composto somente pelas bandas espectrais puras; (2) composto pelas métricas texturais GLCM geradas a partir das bandas espectrais; e (3) composto pelas variáveis dos dois grupos anteriores, além de dois índices de vegetação no caso da imagem Landsat-8, e três índices para a RapidEye. Cada grupo foi testado com os classificadores florestas randômicas (Random Forest- RF), máquinas de vetor de suporte (Support Vector Machine - SVM) e máxima verossimilhança (Maxver). Todos os experimentos alcançaram resultados satisfatórios, com índice Kappa variando de 0,66 a 0,88 e acurácia de usuário e produtor superiores a 50%. O melhor resultado alcançado foi com a imagem Landsat-8, grupo 3, associado ao algoritmo RF. A medida de importância das variáveis obtida com o algoritmo RF mostrou que as métricas texturais média, contraste e dissimilaridade destacaram-se na classificação para ambas as imagens

Abstract: The remote classification of the different vegetation successional stages still represents a challenging task in face of the similar spectral response of such classes. This paper is committed to evaluate the performance of Landsat-8 and RapidEye images in the classification of successional stages within a patch of Mixed Ombrophilous Forest located in São Joaquim National Park, Santa Catarina State, south of Brazil. Three variables dataset extracted from each image were analyzed, namely; (1) one solely consisting of the spectral bands themselves; (2) a second one comprising GLCM-based texture measures derived from the spectral bands; and (3) a third one containing these two datasets and additionally two vegetation indices obtained from the Landsat 8 image and three vegetation indices from the RapidEye image. Each dataset was subject to three classifiers: random forest (RF), support vector machine (SVM), and maximum likelihood estimation (MLE or maxver). All conducted experiments achieved satisfactory results, with Kappa coefficients ranging from 0.66 to 0.88, and both user´s and producer´s accuracies lying over 50%. The best result was attained with the Landsat 8 image using the third dataset and the RF classifier. The analysis of the variables relevance with this classifier showed that the texture measures mean, contrast and dissimilarity were decisive for the successful classification of both images

RESUMO: Estudou-se os parâmetros das Leis Federais no4.771 e 12.651 para a delimitação de PPA de nascentes e cursos d'água em 68 imóveis rurais dos municípios da Região Serrana do Estado de Santa Catarina. Elaboraram-se mapas temáticos da cobertura da terra com as áreas de PPA, a partir da interpretação de imagens aéreas ortorretificadas. Os resultados mostram que o percentual de PPA, a recompor em área rural consolidada, está diretamente relacionado com a área do imóvel e com o tamanho do módulo fiscal do município onde se encontra. Comparando-se com a Lei no4.771, houve redução significativa nas PPA a serem recompostas em áreas rurais consolidadas. Em pequenas propriedades rurais, a redução média variou de 54,6 a 81,8%. Estima-se que 122.372 hectares de terras da região serrana (7,6% da área), caracterizadas como PPA, em área rural consolidada, podem continuar sendo exploradas economicamente.

ABSTRACT: This research aimed to study the parameters of the current and past Brazilian Forest Act based on the Federal Laws No. 4,771 and 12,651 for the determination of the permanent preservation areas (PPA). Water springs and streams for 68 rural properties located at the Santa Catarina State Southern Plateau were considered. Thematic land use/land cover (LULC) as well as PPA maps from the visual interpretation of ortho rectified aerial images were elaborated. The PPA percentage for recovering over consolidated rural areas is directly proportional to both fiscal module size and area of the rural property. Comparing to the Law No. 4,771, there was a significant reduction in the PPA to be recovered on consolidated rural areas. In small rural properties there was an average reduction ranging from 54.6% to 81.8%. A total of 122,372 hectares (7.6% of the total area) characterized as PPA in consolidated rural areas can now be used economically.

Neste estudo foi abordada a segmentação de imagens de alta resolução e a utilização da técnica da mineração de dados. O objetivo foi encontrar correlação entre resposta espectral, espacial, de contexto e de textura, e variáveis dendrométricas obtidas de inventários florestais em áreas de estágio inicial de sucessão da Floresta Ombrófila Densa em três locais do Estado de Santa Catarina. Os dados de campo foram levantados em seis Unidades Amostrais (UA) com 1.600 m² cada. No processamento digital utilizaram-se três imagens de alta resolução espacial (0,39 m) obtidas pelo sensor SAAPI, com três bandas no visível, três no infravermelho próximo, modelo numérico de terreno e de superfície. Dados extraídos do produto digital (atributos) foram utilizados na etapa da mineração dos dados, que selecionou atributos relevantes e descartou aqueles de menor peso. Verificou-se que tanto no estrato arbóreo quanto na regeneração existe heterogeneidade de variáveis, como número de indivíduos (N), diâmetro a altura do peito (DAP) e área basal (AB). Mesmo assim, foram encontradas correlações significativas entre atributos das imagens e os dados de campo. A correlação de maior magnitude absoluta de N foi com as médias da banda 1 (-0,64), 3 (-0,62) e IR1 (0,63), para DAP foi a razão das bandas IR3 (0,56) e 2 (0,55) e para AB foi o menor valor de pixel das bandas 1 (-0,64) e IR3 (-0,60), todas altamente significativas (p<0,01). Esses resultados configuram pontos de partida para futuras investigações a respeito da construção de um estimador de parâmetros biofísicos da vegetação.

High-resolution image segmentation and data mining techniques were used in order to find statistical correlation between spatial, spectral, context and texture attributes, considering field data from early stages of secondary forests in the three study sites of Rain Forest in Santa Catarina State, Brazil. Field data were collected in six sampling plots with 1,600 m² each. Digital processing utilized three high resolution images (0.39 m) obtained by SAAPI sensor with three visible bands, three near infrared bands, one terrain and one surface model. Extracted data from the digital product (attributes) were used for data mining purposes, selecting the ones with relevance and dismissing those of lesser weight. It was found that both tree and regeneration layers are heterogeneous considering number of individuals (N), diameter at breast height (DBH) and basal area (AB). However, significant correlations between forest attributes and field data were found. The higher correlation of N is with the average of the band 1 (-0.64), 3 (-0.62) and IR1 (0.63) for DBH was the ratio of bands IR3 (0.56) and 2 (0.55) and the variable AB was the smallest pixel value of band 1 (-0.64) and IR3 (-0.60). All correlation coefficients were highly significant (p <0.01). These results constitute the starting point for future research on the relationship between groups of variables in order to estimate biophysical parameters

Resumo: O objetivo deste trabalho foi avaliar o desempenho dos classificadores digitais SVM e K-NN para a classificação orientada a objeto em imagens Landsat-8, aplicados ao mapeamento de uso e cobertura do solo da Alta Bacia do Rio Piracicaba-Jaguari, MG. A etapa de pré-processamento contou com a conversão radiométrica e a minimização dos efeitos atmosféricos. Em seguida, foi feita a fusão das bandas multiespectrais (30 m) com a banda pancromática (15 m). Com base em composições RGB e inspeções de campo, definiram-se 15 classes de uso e cobertura do solo. Para a segmentação de bordas, aplicaram-se os limiares 10 e 60 para as configurações de segmentação e união no aplicativo ENVI. A classificação foi feita usando SVM e K-NN. Ambos os classificadores apresentaram elevados valores de índice Kappa (k): 0,92 para SVM e 0,86 para K-NN, significativamente diferentes entre si a 95% de probabilidade. Uma significativa melhoria foi observada para SVM, na classificação correta de diferentes tipologias florestais. A classificação orientada a objetos é amplamente aplicada em imagens de alta resolução espacial; no entanto, os resultados obtidos no presente trabalho mostram a robustez do método também para imagens de média resolução espacial.

Abstract: The objective of this work was to evaluate the performance of SVM and K-NN digital classifiers for the object-based classification on Landsat-8 images, applied to mapping of land use and land cover of Alta Bacia do Rio Piracicaba-Jaguari, in the state of Minas Gerais, Brazil. The pre-processing step consisted of using radiometric conversion and atmospheric correction. Then the multispectral bands (30 m) were merged with the panchromatic band (15 m). Based on RGP compositions and field inspection, 15 land-use and land-cover classes were defined. For edge segmentation, the bounds were set to 10 and 60 for segmentation configuring and merging in the ENVI software. Classification was done using SVM and K-NN. Both classifiers showed high values for the Kappa index (k): 0.92 for SVM and 0.86 for K-NN, significantly different from each other at 95% probability. A major improvement was observed for SVM by the correct classification of different forest types. The object-based classification is largely applied on high-resolution spatial images; however, the results of the present work show the robustness of the method also for medium-resolution spatial images.

Composições de 16 dias de índices de vegetação do sensor MODerate resolution Imaging Spectroradiometer (MODIS), com resolução espacial de 1km, a bordo dos satélites TERRA e AQUA, foram usadas para caracterizar a dinâmica sazonal em 2004 de cinco fitofisionomias de Cerrado e analisar a sua separabilidade espectral. Os índices Normalized Difference Vegetation Index (NDVI) e Enhanced Vegetation Index (EVI), calculados a partir dos dados dos sensores de ambas as plataformas e de uma base comum de pixels, foram comparados entre si. Os resultados indicaram que: (a) dentre as fitofisionomias estudadas, a Floresta Estacional decídua apresentou uma dinâmica sazonal muito marcante em função da perda de folhas da estação chuvosa para a seca (substancial redução nos índices) e do rápido verdejamento com o início da precipitação no final de outubro (rápido incremento de NDVI e EVI); (b) o NDVI mostrou maior variabilidade entre as classes de vegetação do que o EVI apenas na estação seca; (c) a discriminação entre as fitofisionomias melhorou da estação chuvosa para a seca; (d) o NDVI foi mais eficiente do que o EVI para discriminar as classes de vegetação na estação seca, ocorrendo o contrário na estação chuvosa; e (e) na maioria das datas selecionadas para estudo, não houve diferenças estatisticamente significativas entre os índices de vegetação gerados de ambas as plataformas, apesar das variações na qualidade dos pixels selecionados para as composições de 16 dias e na geometria de iluminação e de visada.

MODerate-resolution Imaging Spectroradiometer (MODIS) 16-day vegetation index composites with 1 km of spatial resolution from TERRA and AQUA satellites were used to characterize the seasonal dynamics of five Brazilian savanna physiognomies and to analyze their spectral separability in 2004. The Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI), using data from both platforms and from a common set of pixels, were compared to each other. The results showed that: (a) among the physiognomies under study, Dry Forest (Floresta Estacional decídua) presented a marked seasonal dynamics as a result of the leaf fall from the rainy to the dry season (strong decrease in the indices) and of the fast green up of vegetation with precipitation at the end of October (strong and rapid increase in NDVI and EVI values); (b) NDVI showed greater variability between the vegetation classes than EVI only in the dry season; (c) the discrimination between the physiognomies improved from the rainy to the dry season; (d) the NDVI was more efficient than EVI to separate the vegetation classes in the dry season, but the contrary was observed in the rainy season; and (e) for the majority of the dates under analysis, in spite of the variations in the quality of the pixels selected to compose the vegetation index MODIS product and in the Sun-view geometry, no statistically significant differences between the indices generated from both platforms were observed.