Ciências Florestais

URI Permanente desta comunidade

http://repositorio.ufes.br/handle/10/17465

Programa de Pós-Graduação em Ciências Florestais

Centro: CCAE
Telefone: (28) 3558 2528
URL do programa: http://www.cienciasflorestais.ufes.br/pos-graduacao/PPGCFL

Navegar

Agora exibindo 1 - 16 de 16

Avaliação de diâmetro e altura total de plantio misto com laser scanner portátil
(Universidade Federal do Espírito Santo, 2022-08-29) Oliveira, Klisman; Silva, Gilson Fernandes da; https://orcid.org/0000000178536284; http://lattes.cnpq.br/8643263800313625; https://orcid.org/0000000180519387; http://lattes.cnpq.br/1741069350944065; Torres, Carlos Moreira Miquelino Eleto; https://orcid.org/0000-0003-0255-2637; http://lattes.cnpq.br/2087212860072636; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607
The use of remote sensing equipment has been growing in the forestry scenario, associated with time savings and reduction of field work that can bring to projects, one of which is Light Detection and Ranging (LiDAR). The objective of the present work was to estimate the total height and diameter at breast height of different species in a mixed forest stand with Light Detection and Ranging data. A forest inventory was carried out to collect diameter at breast height and total height, from twelve different species of the Atlantic Forest, being collected ten class centers by species, totaling 120 measured arboreal individuals. The collections took place by means of a Portable Laser Scanner (PLS), the walking was carried out with the equipment in the eight areas studied, ranging from two to ten years of age. Through the walking, it was possible to generate the 3D point cloud of the plantations, from where the segmentation and cleaning of the studied trees was carried out, thus generating a point cloud for each isolated tree. Through the TreeLS and LiDR packages of the R software, it was possible to extract the metrics of diameter at breast height and total height of the point clouds of each tree. At the end of the segmentation, 92 trees of eleven different species were isolated, among the problems found for the isolation of the trees, the competition of the younger plantations with weeds and invasive trees. Another problem found was the competition between crowns, taking into account the intertwining between crowns and obstruction of the lasers by obstacles, which generated occlusions and noise, which made it impossible to isolate the trees. For the diameter metric, species in general the RMSE reached a value of 38.17%, the species Schizolobium parahyba obtained the best statistics reaching an RMSE of 13.14% and Bias 40.65%. The best statistics that Schizolobium parahyba arrived at can be associated with the similarity of growth with commercial species, from which the metrics in the used packages are usually processed. For the total height data obtained with PLS, the general RMSE (%) was 15.14 %, while its bias (%) was 17.73 %, in which four studied species obtained RMSE (%) below 10%, being Ceiba speciosa, Inga edulis, Tibouchina granulosa and Cedrela fissilis, being C. fissilis the only species with RMSE (%) below 5%. The species in general showed overestimated results of diameters at breast height and difficulty in segmenting the trees, whereas the total height metrics were obtained with greater precision, and similar to the statistics found in the literature.
Avaliação de estágios sucessionais de florestas estacionais semideciduais com uso de dados hiperespectrais e LiDAR obtidos a partir de aeronave remotamente pilotada
(Universidade Federal do Espírito Santo, 2025-06-03) Pinon, Tobias Baruc Moreira; Almeida, André Quintão de ; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Effgen, Emanuel Maretto ; https://orcid.org/0000-0002-9031-6337; http://lattes.cnpq.br/0205196565849611; Mendonça, Adriano Ribeiro de ; https://orcid.org/0000-0003-3307-8579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000-0001-9200-1024; http://lattes.cnpq.br/8571909054406808; Almeida, Catherine Torres de ; https://orcid.org/0000-0002-8140-2903; http://lattes.cnpq.br/5534145837431294; Fernandes, Milton Marques ; https://orcid.org/0000-0002-9394-0020; http://lattes.cnpq.br/2151263512584100; Martins Neto, Rorai Pereira; https://orcid.org/0000-0001-5318-2627; http://lattes.cnpq.br/4925375972651580; Silva, Gilson Fernandes da ; https://orcid.org/0000-0001-7853-6284; http://lattes.cnpq.br/8643263800313625
The Atlantic Forest in the state of Espírito Santo has undergone intense degradation, highlighting the urgent need for rapid and accurate methods for its monitoring and conservation. Brazilian Resolution Conama No. 29/1994 establishes criteria for classifying secondary vegetation into successional stages, which determine the potential for forest use. However, this classification, when carried out in the field, is heavily reliant on the expertise of the technical team, due to factors such as training, subjective criteria, and the lack of adequate instruments—potentially compromising the reliability of the results. In this context, the objective of this study was to classify successional stages of vegetation using data acquired by hyperspectral and LiDAR sensors mounted on a Remotely Piloted Aircraft (RPA). The research was conducted in regenerating pasturelands and forest fragments located in southern Espírito Santo, where dendrometric variables such as diameter at breast height (DBH) and total tree height were collected within 30 × 30 m plots. These field measurements were related to hyperspectral (with and without shadow) and LiDAR-derived metrics to estimate dendrometric parameters—mean diameter (D), mean height (H), and basal area (G)—using regression models. Model accuracy was evaluated using the root mean square error (RMSE), adjusted coefficient of determination (adjusted R²), and histograms of percentage error. Successional stage classification was performed using a rule-based method under two scenarios: one with three stages (initial, intermediate, and advanced), and another including the regenerating pasture class. In addition, an unsupervised classification was conducted using hierarchical clustering based on the estimated dendrometric variables and structural and spectral metrics, resulting in five groups: three successional stages and two pasture categories (open and dense shrublands). A principal component analysis (PCA) was also applied. The variables D and H were estimated with higher accuracy using combined data (adjusted R² = 88% and 90%, respectively), while G performed best with LiDAR data alone (adjusted R² = 92%). Shadow pixel removal slightly improved model performance, although its impact on predictive quality was limited. The rule-based classification with three categories achieved an overall accuracy of 88% (Kappa = 0.81), decreasing to 68% (Kappa = 0.59) with the inclusion of the regenerating pasture class. The unsupervised classification using the estimated variables for five classes (open and dense shrublands, and successional stages) reached an accuracy of 64% (Kappa = 0.55). Conversely, the classification based solely on hyperspectral metrics showed high agreement with field-defined stages (92%), whereas LiDAR metrics presented lower correspondence. Multivariate analysis revealed that spectral and structural metrics adequately represent the successional gradient. The integration of hyperspectral and LiDAR data proved effective for the automated mapping of large and inaccessible areas, providing a promising tool to complement forest inventories and reduce subjectivity in the application of legal criteria
Caracterização espectral e fusão de dados lidar e hiperespectrais coletados por drone para estimar a biomassa acima do solo de florestas secundárias da mata atlântica
(Universidade Federal do Espírito Santo, 2025-03-28) Rodrigues, Nívea Maria Mafra; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Silva, Gilson Fernandes da; https://orcid.org/0000-0001-7853-6284; http://lattes.cnpq.br/8643263800313625; https://orcid.org/0000-0002-3750-0813; http://lattes.cnpq.br/1359706450652133; Almeida, Catherine Torres de; https://orcid.org/0000-0002-8140-2903; http://lattes.cnpq.br/5534145837431294; Gonçalves, Fábio Guimarães; http://lattes.cnpq.br/http://lattes.cnpq.br/1116245566543036 ; Martins Neto, Rorai Pereira; https://orcid.org/0000-0001-5318-2627; http://lattes.cnpq.br/4925375972651580; Gorgens, Eric Bastos; https://orcid.org/0000-0003-2517-0279; http://lattes.cnpq.br/2266409430041146
Tropical forests play a fundamental role in the global carbon cycle, biodiversity conservation, soil and water preservation, and provide a wide range of ecosystem services. Therefore, improving tropical forest monitoring using data collected by a remotely piloted aircraft (RPA) is crucial to ensuring these services. In this context, this study aimed to evaluate the use of hyperspectral data collected by an RPA to characterize the vegetation of secondary forest fragments of the Atlantic Forest at different successional stages. Additionally, another objective was to combine LiDAR and hyperspectral data to enhance the estimation of aboveground biomass (AGB) and to spatialize these estimates in the studied areas. To achieve this, all tree individuals (D > 5 cm) were identified and inventoried in 30 field plots (30 × 30 m each) across five forest remnants located in the southern region of Espírito Santo state. Aerial point clouds and hyperspectral image cubes were generated for all analyzed fragments simultaneously with the field forest inventory. Subsequently, traditional metrics and metrics derived from the Fourier transform of canopy height were estimated from the point clouds, along with spectral information, including reflectance values and vegetation indices, for each plot. The successional stages of the analyzed secondary forest fragments could be distinguished using hyperspectral data collected by RPA. In the context of secondary tropical forests, characterized by high structural variability and different successional stages, the integration of LiDAR and hyperspectral data resulted in minimal improvements in AGB estimation accuracy. In some cases, data fusion did not improve the results compared to models based solely on LiDAR, indicating that spectral information did not significantly contribute to enhancing AGB estimates.
Classificação do estágio sucessional da vegetação em áreas da mata atlântica com a utilização de nuvem de pontos 3D aérea e terrestre
(Universidade Federal do Espírito Santo, 2022-08-22) Cabral, Ricardo Pinheiro; Silva, Gilson Fernandes da; https://orcid.org/0000000178536284; http://lattes.cnpq.br/8643263800313625; https://orcid.org/0000000246312603; http://lattes.cnpq.br/6008950574275011; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Bonilla-Bedoya, Santiago
The definition of strategies for conducting, recovering and restoring deforested and or degraded areas is due to the stage of ecological succession it is in. Usually, the classification of the successional stage is carried out in the field, from forest inventory campaigns. However, these campaigns are considerably costly and require a high execution time, and still have limited spatial coverage. Currently, forest inventories are being enhanced from the use of three-dimensional data obtained by remote sensing, with emphasis on metrics derived from Light Detection And Ranging (LiDAR) and Digital Aerial Photogrammetry (DAP). Therefore, the main objective of this work was to estimate some parameters of forest interest and to classify the stage of ecological succession of areas with degraded vegetation of the Atlantic Forest Biome with the use of 3D DAP data obtained by Remotely Piloted Aircraft (ARP). An analysis of the costs of traditional and improved inventory was also performed. Initially, the field estimation of the values of total height (h), diameter at 1,30 m of soil (dbh) and basal area (ba) of the individuals of 40 inventory plots (30 x 30 m each) was estimated in the field. In the same plots, 3D point clouds were generated by DAP-RPA and Portable LiDAR (PLS). Next, regression models were adjusted and validated to estimate the values of mean h and dbh, and ab from the traditional metrics based on the heights of the DAP and LiDAR point cloud. Finally, maps of the successional stage of the vegetation were generated with spatial resolution of 30 m. The maps were created based on pre-established intervals of mean h and dbh and ab, according to CONAMA resolution 29/94. Expenses were considered in the cost analysis for equipment acquisition, collection, processing and data analysis. Finally, the costs of inventories were compared. The estimation models based on DAP showed performance similar to models adjusted with LiDAR, with Values of R² ranging from 88,3% to 94% and RMSE (%) between 11,11 and 28,46 for DAP and R² between 83,6% and 96,4% and RMSE (%) between 8,58 and 33,63 for LiDAR. The maps of the succession stages estimated by DAP were compatible with the succession classes estimated in the 40 field plots. The cost of acquiring the equipment used in the dap was twice as high as the traditional method and ten times lower than the enhanced inventory performed with LiDAR. Analyzing the costs of data acquisition and processing per hectare, DAP presented a cost of only R$83,40, against R$16449,16 and R$26268,33 for LiDAR and the traditional method, respectively. DAP metrics can be used to estimate the values of average height, diameter to breast height and basal area of vegetation analyzed with accuracy similar to those performed with LiDAR. In addition to presenting the lowest cost, the estimates made by DAP-RPA allowed the classification of successional stages in the secondary forest areas of Atlantic vegetation analyzed.
Estimação de área basal, volume e biomassa em um fragmento de Caatinga Hiperxerófila densa no alto sertão sergipano com base em dados MSI/Sentinel-2
(Universidade Federal do Espírito Santo, 2018-10-26) Fernandes, Márcia Rodrigues de Moura; Almeida, André Quintão de; Silva, Gilson Fernandes da; Gonçalves, Fabio Guimarães; Binoti, Daniel Henrique Breda; Mendonça, Adriano Ribeiro de
The aim of this study was to estimate the basal area, the wood of volume and the aerial biomass of the Caatinga vegetation of the semi-arid region of Sergipe based on MSI/Sentinel-2 sensor data. In order to reach this objective, the dendrometric variables were measured: the diameter at the height of 1.30 m of the soil (DBH) and the total height (H), obtained by means of systematic sampling, with fixed square plots of 30 mx 30 m (900 m2 ), totaling 40 plots. The independent variables were extracted from the spectral bands in the spectral windows 3 x 3, 5 x 5, 7 x 7 and 9 x 9 pixels, and calculated the ratio of bands, vegetation indices, image fractionvegetation and texture metrics based on co-occurrence matrix. The variables derived from Sentinel-2 were examined for their accuracy in the estimation of the variables basal area (m2 ), wood of volume (m3 ) and aerial biomass (Mg) using multiple linear (MLR) regression analysis and Artificial Neural Networks (ANN). The statistics coefficient of determination (R 2 ), root mean square error (RMSE and RMSE%) and bias (B%) were used in the evaluation of the estimates generated by the models. The results of this study demonstrated that the texture metrics, calculated in window sizes 5 x 5 and 7 x 7 pixels, were more accurate. The best statistics were in the estimation of the basal area that presented a R 2 = 0.9591, RQME = 0.63 m2 ha-1 (10.19%) and bias = -0.39% in the validation of the MLR; and R 2 = 0.9782, RQME = 0.68 m2 ha-1 (10.85%) and bias = -0.80% in ANN validation. In the end, it was concluded that the use of independent variables from the MSI sensor in the analysis MLR and ANN estimate basal area, wood of volume and aerial biomass presented as an effective and accurate method, emphasizing the importance of the texture of the image in the prediction of these variables in the studied area.
Estimação de áreas seccionais de troncos de árvores individuais por meio de dados coletados remotamente
(Universidade Federal do Espírito Santo, 2025-04-09) Lavagnoli, Gabriel Lessa da Silva; Silva, Gilson Fernandes da; https://orcid.org/0000-0001-7853-6284; http://lattes.cnpq.br/8643263800313625; https://orcid.org/0000-0002-9007-1990; http://lattes.cnpq.br/9310315398167707; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Soares, Carlos Pedro Boechat; https://orcid.org/0000-0001-6475-3376; http://lattes.cnpq.br/0959425632265455; Cosenza, Diogo Nepomuceno; https://orcid.org/0000-0001-8495-8002; http://lattes.cnpq.br/0496006405127895; Mendonça, Adriano Ribeiro de; https://orcid.org/0000-0003-3307-8579; http://lattes.cnpq.br/9110967421921927
This thesis investigates the accuracy in the measurement of tree trunk cross-sectional areas, highlighting the practical importance of this variable for forest inventories and its implications for volume and biomass estimates. The work is structured into two complementary studies. The first study evaluates the impacts of convexity and isoperimetric deficits on traditional measurement methods, such as calipers and diameter tapes, comparing them to a photographic method developed by the author, which calculates areas and estimates contours through pixel counting. The results showed that traditional methods exhibit significant systematic errors, arising from the incorrect assumption of perfect circularity of cross-sections, whereas the photographic method demonstrated high precision, with mean relative errors below 0.1%. The second study proposes a computational methodology for estimating cross-sectional areas from point clouds obtained using a GeoSLAM LIDAR sensor, comparing the measurements with those obtained from a high-precision infrared scanner (EinScan). The research involved the analysis of 56 eucalyptus trees, comprising more than 1,000 cross-sections. Additional simulations of traditional methods were also conducted for direct comparison. It was observed that traditional techniques, once again, tended to overestimate the areas (with a mean bias of approximately 2.8%), while the LiDAR-based method showed the opposite trend, with a mean bias of -8.12%. However, after applying a specific mathematical correction, the LiDAR estimates achieved excellent accuracy, with a relative root mean square error (RMSE) of 2.4%, a mean relative bias close to zero, and a mean absolute relative error (MAE) of 1.65%, demonstrating great potential for practical applications after appropriate adjustments.
Estimação de características dendométricas para floresta estacional semidecidual submontana com o uso de dados OLI e SRTM
(Universidade Federal do Espírito Santo, 2018-02-20) Gonçalves, Anny Francielly Ataide; Almeida, André Quintão de; Silva, Gilson Fernandes da; Binoti, Daniel Henrique Breda; Mendonça, Adriano Ribeiro de
Brazil's forestry policy predict that all states of the federation should update the forest inventory. Linked to this, it is necessary to use techniques, such as remote sensing, that make it possible to obtain accurate information and reduce costs in the development of this activity. The objective of this study was to evaluate the use of Landsat 8 OLI sensor data and SRTM data in equations for estimation of variables basal area and volume of wood for a fragment of Submontane Semidecidual Seasonal Forest belonging to the Private Reserve of the Natural Heritage (RPPN) Cafundó, located in the municipality of Cachoeiro do Itapemirim, ES. The forest inventory was realized out in 25 plots of 1,000 m² (20 m x 50 m) and the estimates of basal area and volume of wood with bark were obtained by means of allometric equations. Subsequently, these estimates were related to the variables derived from the remote sensing, through the regression analysis. In the regression analysis, the dependent variables were the basal area and volume of bark wood, and the independent variables were the OLI sensor spectral bands, the ratio between bands, vegetation indices and relief characteristics extracted from the SRTM, tested for different spectral windows. The technique of selection of explanatory variables used was the exhaustive search and the statistical evaluation of the regression made use of the 2 R , RMSE (%), residue dispersion and Leave-one-out ( 2 CV R and RMSEcv) cross-validation. For the studied variables, it was observed that the 3 x 3 pixel spectral window was the most related to the data of basal area and volume of wood, and relief variables extracted from the SRTM presented good performance when combined with the spectral variables of the sensor OLI. For the basal area, the equation that best fit the data presented 2 R of 0,6554, 2 CV R of 0,6244, RMSE (%) of 14,53% and RMSEcv (%) of 18,15%. In relation to volume, the equation presented 2 R of 0,6039, 2 CV R of 0,5380, RMSE (%) of 23,03% and RMSEcv (%) of 30,30%. The estimation of the basal area and volume of wood for the Submontane Semidecidual Seasonal Forest fragment using spectral data presented satisfactory results, emphasizing the importance of topography in the prediction of these variables in the studied area.
Inventário de uma floresta de produção com a utilização de imagens MSI/Sentinel-2 e fotogrametria aérea digital
(Universidade Federal do Espírito Santo, 2021-04-30) Carvalho, Rachel Clemente; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Mendonça, Adriano Ribeiro de; https://orcid.org/0000000333078579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000000167085369; http://lattes.cnpq.br/4085851196278587; Gonçalves, Fábio Guimarães; https://orcid.org/0000-0001-6925-3012; http://lattes.cnpq.br/1116245566543036; Silva, Gilson Fernandes da; https://orcid.org/0000000178536284; http://lattes.cnpq.br/8643263800313625
In the forestry sector, knowledge of forest productivity is obtained through forest inventories. However, the sampling techniques traditionally applied to forest inventories have a high demand for time and high cost of execution. Therefore, it is necessary to evaluate the use of alternative techniques to obtain this data, such as remote sensing applications. In this context, the use of remote sensing allows the acquisition of data in large areas quickly and at a reduced cost. This work had as main objective to estimate attributes of forest interest of a commercial planting of eucalyptus by orbital image (IO) and digital aerial photogrammetry (DAP) and to compare with the results obtained by the traditional forest inventory. As a secondary objective, an evaluation of DAP products was carried out based on planting attributes collected in the field. For the inventory based on the IO, spectral bands of an image from the MSI/Sentinel-2 sensor were selected and various vegetation indices were calculated. The individual bands and the vegetation indexes were used as predictive variables for the modeling. To obtain the DAP data, a flight was performed with an unmanned aerial vehicle (UAV) for the generation of a three-dimensional point cloud by the SfM algorithm and also a digital terrain model (DTM) for its normalization. The quality of FAD's DTM was evaluated by comparing the values of dominant height of each plot with the metrics representative of the maximum height of the normalized point cloud. Traditional height-based metrics were extracted for each plot, which were used as predictor variables. Multiple linear regression (MLR) and artificial neural networks (ANN) performed the basal area (G) and volume (V) estimation process. For the modeling, three data sources were considered, IO, DAP and the combination of IO and DAP. At the plot level, to estimate the G from IO and DAP data, the lowest values of RMSE in the validation occurred in the ANN modeling, being 13.22% and 13.36%, respectively. For the combination of IO and DAP, the MLR presented a lower RMSE in the validation (RMSE = 12.46%). The same happened for V, with the lowest values of RMSE in the validation with data from IO (15.05%) and DAP (16.58%), obtained in ANN modeling, for the combination of IO and DAP, the lower RMSE was obtained by MLR (14.14%). When performing the modeling for the entire area, it was possible to observe that the ANN presented greater capacity for generalization, with results closer to those obtained in the traditional forest inventory for all data sources. All the averages of G and V were close to the values obtained in the inventory, with a maximum of 3.2% difference. As in the plot level results, the combination of IO and DAP generated more accurate results for the whole area, with a difference of 0.3% for G and 0.4% for V, in relation to the inventory. The results obtained in this study indicate that IO and DAP data can be used for the inventory of G and V in eucalyptus plantations, with results compatible with those obtained in the traditional forest inventory.
Metodologia para detecção de colheita de eucalyptus no Espírito Santo: uma abordagem com sentinel-2
(Universidade Federal do Espírito Santo, 2023-10-31) Marques, Leon Muller; Mendonça, Adriano Ribeiro de; https://orcid.org/0000000333078579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000000277073190; http://lattes.cnpq.br/9801928324329659; Almeida, André Quintão de; Vieira, Carlos Antonio Oliveira; Santos, Jeangelis Silva
The public sector requires tools enabling systematic monitoring of cultivated forest areas. The use of remote sensing techniques facilitates large-scale surveillance by environmental agencies. The primary aim of this study was to create a methodology for identifying Eucalyptus stands' harvesting in the state of Espírito Santo. To conduct this monitoring, the Eucalyptus stands' database provided by a company was utilized. This database was constructed from visual analysis of Sentinel-2 images from the year 2020. The accuracy of the map was validated using the Kappa coefficient. Reference data for harvesting was also obtained through visual interpretation by analyzing a fiveyear history of Sentinel-2 images. Each year, the best images from each quarter were selected, identifying changes in spectral response from forest to exposed soil, indicating harvesting.In constructing the harvest detection algorithm, land cover classes provided by the European Space Agency (ESA) between 2019 and 2020 were used, employing the Scene Classification (SCL). This product offers land cover information for each Sentinel-2 image collected every five days. Subsequently, classes of interest (cloud, soil, and vegetation) were filtered. To acquire an image with fewer cloud cover instances, three temporal categories (monthly, bimonthly, and quarterly) were analyzed, generated by aggregating weekly images. SCL classes were intersected with Eucalyptus stands, determining the temporal category with the least cloud cover in the stand database.Once the best temporal category was defined, harvest detection relied on accumulating pixels classified as soil in each new image composition, evaluating the percentage of soil in the stand. Experimentations were conducted to define the best soil percentage threshold to consider a stand as harvested. To calculate algorithm accuracy, performance evaluations were done with two distinct strategies. Strategy 1 directly compared the algorithm-identified harvest month with the reference date for each Eucalyptus stand. Strategy 2 assessed algorithm accuracy by varying one month before and after the reference date. Subsequently, the impact of slope and stand size on algorithm accuracy was analyzed, calculating errors and successes for each slope class and stand size.The Eucalyptus map obtained a Kappa of 0.851 and an overall accuracy of 94%. The quarterly temporal category proved most effective in minimizing cloud effects, as no stand exhibited over 20% cloud coverage. Strategy 2 was the most efficient, achieving an algorithm accuracy of 84.5% with a 25% soil threshold. It was observed that higher slopes corresponded to lower accuracy, while stand size showed a direct relationship: larger size led to higher accuracy.The developed algorithm represents an advancement in applying new monitoring and oversight methods for Eucalyptus plantations, and it can be adopted by the public sector.
Modelagem de riscos de incêndios florestais e otimização da alocação das estruturas de combate por meio de técnicas de inteligência artificial
(Universidade Federal do Espírito Santo, 2023-08-31) Silva, Jeferson Pereira Martins; Silva, Gilson Fernandes da; https://orcid.org/0000000178536284; http://lattes.cnpq.br/8643263800313625; https://orcid.org/0000000315521127; http://lattes.cnpq.br/6748966859692740; Barros Junior, Antonio Almeida de; https://orcid.org/0000-0002-2449-7221; http://lattes.cnpq.br/5104467305835940; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Pezzopane, Jose Eduardo Macedo; https://orcid.org/0000000300244016; http://lattes.cnpq.br/3640768649683482; Silva, Evandro Ferreira da
This study presents an approach to wildfire management integrating a WebGIS system and artificial intelligence. To train the deep learning model, data related to vegetation, topography, anthropogenic factors, and historical fire records for the year 2008 in Andalusia, Spain were collected. The dataset was duly normalized and split into 70% for training, 10% for validation, and 20% for testing. Various algorithms and activation functions were evaluated, with the combination of Adam and Relu standing out, recording an accuracy of 0.86 during training. Based on this model, a risk map was generated. By applying the K-means method to this map, high-risk areas were identified, and central points for the installation of firefighting infrastructures were suggested. To validate the model's efficacy, the suggested positions were compared with the actual locations of firefighting aircraft in Andalusia, Spain. With 31 clusters and a risk threshold of 0.75, the proximity between the proposed coordinates and the actual ones was notable, reinforcing the practical potential of the approach proposed in this study.
Modelagem de variáveis biométricas em diferentes classes de vegetação por meio de dados sintéticos do satélite landsat
(Universidade Federal do Espírito Santo, 2023-09-27) Ferreira, Larissa Garcia; Mendonça, Adriano Ribeiro de; https://orcid.org/0000000333078579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000000226340631; http://lattes.cnpq.br/6247239452984557; Almeida, André Quintão de; Silva, Gilson Fernandes da; https://orcid.org/0000000178536284; http://lattes.cnpq.br/8643263800313625
The use of passive remote sensing is an alternative to forest inventory to estimate dendrometric variables. Therefore, the main objective of this work was to evaluate the accuracy of estimates of average diameter, average total height and stem biomass at different stages of succession of the Atlantic Forest biome based on synthetic images from the Landsat satellite. Remote sensing data were obtained from synthetic images using the Continuous Change Detection and Classification (CCDC) algorithm. Spectral bands and NDVI were used as explanatory variables for modeling. Graphs of observed versus estimated variables, the adjusted coefficient of determination (𝑅̅2%) %), and the root mean square error (RMSE%) were used to evaluate the models. Model selection was performed using the F test to verify the significance of model parameters and analysis of variance was used to compare nested models. Furthermore, k-fold cross validation was performed repeated 1000 times with k equal to 10 for the selected model for each variable analyzed. The selected variables were the annual percentiles (10 to 100) of spectral bands 𝐵2, 𝐵3, 𝐵4, 𝐵5, 𝐵6e 𝐵7. The near-infrared (𝐵5) and mid-infrared 1 (𝐵5) bands were selected for all estimation equations for mean diameter, mean total height and bole biomass at different successional stages. Estimates of average diameter and average total height showed good accuracy in cross-validation. The bole biomass estimates were of low accuracy and are not recommended for estimating the biomass of the different successional stages. The use of metrics extracted from synthetic images obtained from Landsat satellite images, together with traditional forest inventory data, made it possible to estimate the biometric variables of the study area.
Modelo Espacial para Prevenção de Incêndios Florestais em Unidade de Conservação de Restinga
(Universidade Federal do Espírito Santo, 2024-02-19) Biazatti, Leonardo Duarte; Fiedler, Nilton Cesar; https://orcid.org/0000-0002-3895-661X; http://lattes.cnpq.br/8699171075880935; https://orcid.org/0000-0002-9345-8592; http://lattes.cnpq.br/9329666073648836; Simões, Danilo; https://orcid.org/0000-0001-8009-2598; http://lattes.cnpq.br/4290623857436137; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Dias, Patrícia Borges; https://orcid.org/0000000252278341; http://lattes.cnpq.br/1194831380343570
The Restingas face constant threats from anthropogenic pressure, especially forest fires, which act as agents of degradation in these environments, especially for the purposes of real estate speculation. Despite advances in understanding the behavior of fires in the most varied types of vegetation, the occurrence of these events in sandbanks has been little studied. This research aimed to develop a model for predicting, preventing, detecting and helping to fight forest fires, using geotechnological tools, in the Paulo César Vinha State Park (PEPCV) and the Setiba Environmental Protection Area (APA), located in the municipality of Guarapari, which spatially and structurally represent a large part of the Atlantic Forest sandbank formations in the south of the state of Espírito Santo. Through the application of geoprocessing techniques using geotechnological tools such as Fuzzy logic, Hierarchical Process Analysis (HPA) logic, least-cost path and network analysis on the variables of land use and occupation, altitude, slope, relief orientation, proximity to roads, urban areas, and watercourses, rainfall, air temperature and land surface temperature and heat zones, the areas most at risk of forest fires were delimited, as were the priority sites for the construction of firebreaks and roads, the optimum points for installing video monitoring towers and water collection points. It was found that the area is at very high risk of fires, as 55.17% of the territory is in this risk range. The sites classified as priorities for the construction of firebreaks, because they are close to urban areas and areas of higher risk, represented 65.41% of the study area, so there was a need to install them around these sites. The road allocation model indicated 30 possible routes for opening complementary roads, always going through already anthropized sites. The model for installing video surveillance towers determined six possible allocation scenarios and, considering the 96.55% visualization rate achieved and the cost of implementation, the third scenario (2A) was the best. On the other hand, the optimization for the allocation of water collection points established sixteen possible scenarios, with the eleventh (3C) being the most promising as it showed 80% coverage of the demand points, using only eleven points, which makes it cost-effective to install. Given these results, it was possible to conclude that the high level of risk of forest fires in the PEPCV and Setiba APA highlights the urgency of implementing preventive measures aimed at reducing the risk and mitigating the possible consequences of fires in the region. In this context, this study is a source of information for designing predictive and preventive strategies for the PEPCV, as well as providing support when fighting fires. Finally, it should be noted that the methodology used is applicable, replicable and adaptable to any conservation unit.
Novo algoritmo para detecção automática e mensuração de alturas de árvores individuais em plantios florestais usando scanner a laser móvel
(Universidade Federal do Espírito Santo, 2025-03-31) Silva, Valeria Alves da; Silva, Gilson Fernandes da; https://orcid.org/0000-0001-7853-6284; http://lattes.cnpq.br/8643263800313625; https://orcid.org/; http://lattes.cnpq.br/; Mendonça, Adriano Ribeiro de; https://orcid.org/0000-0003-3307-8579; http://lattes.cnpq.br/9110967421921927; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Soares, Carlos Pedro Boechat; http://lattes.cnpq.br/0959425632265455; Cosenza, Diogo Nepomuceno; https://orcid.org/0000-0001-8495-8002; http://lattes.cnpq.br/0496006405127895
Measuring the total height (H) of trees in forest plantations is crucial for several reasons. Tree height is a key variable in calculating tree volume and biomass. Accurate height measurements, combined with diameter at breast height (D) data, allow for accurate estimates of individual tree volume and, by extension, total volume and biomass of the plantation. This is essential for assessing timber yield, carbon sequestration potential, and overall forest productivity. The height distribution within a plantation reflects stand density and structure. Accurate measurements of tree height are vital for the economic assessment of the plantation. Knowing volume and biomass allows for more accurate estimates of timber value and potential plantation revenue. In the context of climate change, accurate estimates of forest biomass (influenced by height) are necessary for carbon accounting and monitoring carbon sequestration efforts. In summary, accurate measurement of tree height (H) is not merely a component of forest inventory; it is an integral part of the overall forestry inventory. is a fundamental parameter that underpins many crucial aspects of forest management, from economic assessment to ecological assessments and sustainable resource planning. This work presents an algorithm for measuring the total height (H) of trees in forest plantations efficiently and accurately using Mobile Laser Scanner (MLS) data. In the first part of the paper, the algorithm focuses on detecting individual tree trunks in plantations. It employs DBSCAN and RANSAC methods for accurate detection, achieving 100% accuracy under ideal conditions and approximately 96% in challenging scenarios. The efficiency of the algorithm was evaluated on various computer configurations. In the second part of the paper, the algorithm measures the total height (H) of trees found in the plantation by its trunk identification method. The algorithm achieved significant accuracy, particularly in the challenge of accurately measuring shorter measurements obtained using a tape measure and a total station, demonstrating superior accuracy compared to the algorithms used in the experiments (TreeLS and 3DFin), especially for shorter trees. The study also analyzes the error distribution for each method, and the proposed algorithm stands out by presenting a more normal and less skewed error distribution than the other algorithms. Although slightly slower than 3DFin, its improved accuracy makes it a valuable tool for forest inventory. Areas for future improvements include processing speed and handling for processing low-density point clouds.
Predição e projeção do crescimento e da produção de plantios de eucalipto por meio de imagens multiespectrais de média resolução espacial
(Universidade Federal do Espírito Santo, 2020-02-11) Santos, Jeangelis Silva; Mendonça, Adriano Ribeiro de; https://orcid.org/0000000333078579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000000347857573; http://lattes.cnpq.br/8339532503141256 ; Gonçalves, Fabio Guimarães; http://lattes.cnpq.br/1116245566543036 ; Carvalho, Samuel de Padua Chaves e; https://orcid.org/0000-0002-5590-9049; http://lattes.cnpq.br/6176482316661283; Almeida, André Quintão de; Silva, Gilson Fernandes da; https://orcid.org/0000000178536284; http://lattes.cnpq.br/8643263800313625
The efficient management and planning of forest areas depends directly on the acquisition of accurate information about the stands. Information about the development of forests can be previously obtained by growth and yield models. However, the adjustment of these models requires data from continuous forest inventories, which are complex and costly activities. One of the alternatives that can reduce the costs of the forest inventory is the use of remote sensing tools. Therefore, the objective of this work was to propose a methodology for using medium spatial resolution multiespectral data for the prediction and projection of growth and yield and to determine the technical age of harvesting of eucalyptus forests, aiming at reducing the number of plots measured in the forest inventory. For this purpose, two databases were used: one containing information on age and volume per hectare of 40 permanent plots measured between 2006 and 2011, with ages varying from two to seven years, and other containing time series of Tasseled Cap (TC) metrics extracted from ETM+/Landsat 7 imagery, smoothed by the Savitzky-Golay filter. To assess the possibility of reducing the number of plots measured in the continuous forest inventory when using remote sensing data, three scenarios were proposed, with different sampling intensities: 1) one plot every 28 ha; 2) one plot every 42 ha, and; 3) one plot every 83 ha. The estimation was performed by artificial neural networks and, in the prediction, the input variables were the age of the stand and the metrics of the Tasseled Cap transformation (brightness, greenness and wetness). For the projection, the variables were the current and future age and the current volume, obtained by the prediction for the first year of the continuous forest inventory. The prediction and projection were applied wall-to-wall, and the projection maps were used to calculate the mean and current annual increment and to determine the technical age of harvest. In the wall-to-wall prediction, the RMSE values ranged from 7.92% in scenario 1 to 10.67% in scenario 3. As for the projection, the RMSE varied from 9.68% in scenario 2 to 11.75% in scenario 3. In general, there was no major discrepancy between the accuracy measures in the three scenarios. In addition, all the scenarios analyzed for prediction and projection presented estimated values within the confidence interval of the forest inventory. The mean and current monthly increment values projected by the different scenarios analyzed did not differ from that obtained by the continuous forest inventory, with the growth curve inflection and forest maturity points being very close. Therefore, it can be concluded that the use of remotely sensed data allowed to accurately estimate the prediction and projection of growth and production of eucalyptus forests. In addition, by applying the methodology presented here, it is possible to significantly reduce the sampling intensity by up to one plot every 83 ha, with accuracy compatible with the methodology traditionally used in the continuous forest inventory.
Sistema de Detecção Remota do Desmatamento Florestal no Estado do Espírito Santo
(Universidade Federal do Espírito Santo, 2024-03-28) Leite, Igor Vieira; Almeida, André Quintão de; https://orcid.org/0000-0002-5063-1762; http://lattes.cnpq.br/5929672339693607; Mendonça, Adriano Ribeiro; https://orcid.org/0000-0003-3307-8579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000-0003-2910-0251; http://lattes.cnpq.br/2626092800028236; Silva, Jeferson Pereira Martins; https://orcid.org/; http://lattes.cnpq.br/6748966859692740; Moreira, Taís Rizzo; https://orcid.org/0000-0001-5536-6286; http://lattes.cnpq.br/6717864186103246
Deforestation and forest degradation pose a significant threat to biodiversity and environmental balance. The state of Espírito Santo, although known for its rich biodiversity, faces challenges in monitoring these environmental impacts. Orbital remote sensing and the development of robust platforms for geospatial data processing, such as Google Earth Engine (GEE), emerge as alternatives to overcome these monitoring challenges. In this context, this study aimed to evaluate a system with high spatial and temporal resolution to monitor forest degradation and deforestation in the Atlantic Forest of Espírito Santo, using Sentinel-2 images. The native forest base map was derived from the annual cover map data of MapBiomas, combined with Sentinel-2 images in the Scene Classification Layer (SCL) band for soil, vegetation, and cloud classes. To validate the data obtained by FlorESat, a confusion matrix was calculated. A pixellevel concordance analysis between classes was performed using the alert database compiled by RAD MapBiomas, which contains 403 polygons from 2019 to 2022. FlorESat's deforestation mapping indicated that the total deforested area for the same years was 1,780.14 ha. In the validation, the proposed system showed mapping accuracy, precision, and specificity of 93.3%, 94.7%, and 93.8% respectively, for 52 randomly delineated polygons in forest areas in ES when compared to photointerpretation. Additionally, it was found that 59.85% of the pixels identified by the FlorESat tool matched directly with the alert database issued and validated in the field. Of the non-coincident pixels, 65.11% were covered by clouds and 34.89% were mapped as forests, highlighting a limitation of orbital data. However, when considering the scenario where cloud pixels are considered as deforested areas, the percentage of concordance was 85.99% between the two datasets.
Uso da fotogrametria aérea digital via imagens coletadas por drone no inventário quantitativo de uma floresta urbana
(Universidade Federal do Espírito Santo, 2023-10-30) Souza, Laís Goncalves Pires de; Mendonca, Adriano Ribeiro de; https://orcid.org/0000000333078579; http://lattes.cnpq.br/9110967421921927; https://orcid.org/0000000187394070; http://lattes.cnpq.br/5588447816273811; Callegaro, Rafael Marian; Almeida, André Quintão de; Moura, Cristiane Coelho de
In urban areas, trees play a crucial role in changing the landscape and local microclimate, in addition to promoting carbon sequestration and providing leisure and recreation spaces for the population. However, the establishment and maintenance of trees in cities pose a challenge for local administration, requiring environmental knowledge of the region, species, and deployment location. Currently, forest inventory enhanced with remote sensing data emerges as a facilitator of urban planning, expediting the tree inventory process and, consequently, decision-making. This study aimed to assess the accuracy of digital aerial photogrammetry (FAD) using images collected by a remotely piloted aircraft (RPA) in detecting trees and estimation of biometric variables in an urban forest inventory. The inventory was conducted on Governador Lindemberg Avenue, located in the municipality of Jerônimo Monteiro, Espírito Santo. High spatial resolution images were obtained by a multirotor RPA during the field inventory period. Subsequently, tree individuals were automatically identified, and their canopies were segmented using FAD-3D data. Finally, total height (H), diameter at 1,3m above ground (D), and canopy diameter (dc) values were estimated from regression models fitted with 3D point cloud height metrics. A total of 144 individuals were inventoried. For FAD validation, errors found were 0,32% for Digital Terrain Model (MDT) and 16,23% for total height. The windowed Variable detection algorithm (wV) using the point cloud as data source automatically identified 78% of individuals. For the comparison of canopy diameters, errors were 17,94%, 21,2% and 29,5% for manual measurements, FAD images, and field measurements with four rays, eight rays, and automatically obtained diameters through canopy identification and segmentation in FAD images, respectively. Regression models errors for H, D and dc were 8,97%, 36,76% and 15,68% respectively. The survey demonstrated the automatic identification of trees and extraction of traditional metrics for generating models to obtain variables of interest. The MDT obtained obtained provided satisfactory results for tree height estimation through FAD-RPA. Manual measurements with FAD images were considered satisfactory for canopy diameter, proving to be the best method for this variable. Additionally, regression models with tradicional metrics obtained were satisfactory for H and dc estimation, showing accurate RMSE and R² values. However, the trunk diameter model showed different results. In conclusion, conducting an aerial photogrammetric survey of urban areas using a remotely piloted aircraft is feasible and can provide valuable data for urban tree planning.

Navegar

Navegando Ciências Florestais por Autor "Almeida, André Quintão de"

Resultados por página

Opções de Ordenação