Search

Accurately mapping farmlands is important for precision agriculture practices. Unmanned aerial vehicles (UAV) embedded with multispectral cameras are commonly used to map plants in agricultural landscapes. However, separating plantation fields from the remaining objects in a multispectral scene is a difficult task for traditional algorithms. In this connection, deep learning methods that perform semantic segmentation could help improve the overall outcome. In this study, state-of-the-art deep learning methods to semantic segment citrus-trees in multispectral images were evaluated. For this purpose, a multispectral camera that operates at the green (530–570 nm), red (640–680 nm), red-edge (730–740 nm) and also near-infrared (770–810 nm) spectral regions was used. The performance of the following five state-of-the-art pixelwise methods were evaluated: fully convolutional network (FCN), U-Net, SegNet, dynamic dilated convolution network (DDCN) and DeepLabV3 + . The results indicated that the evaluated methods performed similarly in the proposed task, returning F1-Scores between 94.00% (FCN and U-Net) and 94.42% (DDCN). It was also determined the inference time needed per area and, although the DDCN method was slower, based on a qualitative analysis, it performed better in highly shadow-affected areas. This study demonstrated that the semantic segmentation of citrus orchards is highly achievable with deep neural networks. The state-of-the-art deep learning methods investigated here proved to be equally suitable to solve this task, providing fast solutions with inference time varying from 0.98 to 4.36 min per hectare. This approach could be incorporated into similar research, and contribute to decision-making and accurate mapping of plantation fields.

In soybean, there is a lack of research aiming to compare the performance of machine learning (ML) and deep learning (DL) methods to predict more than one agronomic variable, such as days to maturity (DM), plant height (PH), and grain yield (GY). As these variables are important to developing an overall precision farming model, we propose a machine learning approach to predict DM, PH, and GY for soybean cultivars based on multispectral bands. The field experiment considered 524 genotypes of soybeans in the 2017/2018 and 2018/2019 growing seasons and a multitemporal–multispectral dataset collected by embedded sensor in an unmanned aerial vehicle (UAV). We proposed a multilayer deep learning regression network, trained during 2000 epochs using an adaptive subgradient method, a random Gaussian initialization, and a 50% dropout in the first hidden layer for regularization. Three different scenarios, including only spectral bands, only vegetation indices, and spectral bands plus vegetation indices, were adopted to infer each variable (PH, DM, and GY). The DL model performance was compared against shallow learning methods such as random forest (RF), support vector machine (SVM), and linear regression (LR). The results indicate that our approach has the potential to predict soybean-related variables using multispectral bands only. Both DL and RF models presented a strong (r surpassing 0.77) prediction capacity for the PH variable, regardless of the adopted input variables group. Our results demonstrated that the DL model (r = 0.66) was superior to predict DM when the input variable was the spectral bands. For GY, all machine learning models evaluated presented similar performance (r ranging from 0.42 to 0.44) for each tested scenario. In conclusion, this study demonstrated an efficient approach to a computational solution capable of predicting multiple important soybean crop variables based on remote sensing data. Future research could benefit from the information presented here and be implemented in subsequent processes related to soybean cultivars or other types of agronomic crops.

In cotton cultivars, an insect that causes irreversible damage is the Spodoptera frugiperda, known as the fall armyworm. Since the visual detection of plants is a burdensome task for human inspection, the spectral information related to plant damage, registered on a spectral scale, can be useful. These measurements, associated with machine learning techniques, produce useful information for a rapid and non-invasive inspection method development. To contribute to this gap fulfillment, this paper proposes a machine learning framework to model the spectral response of cotton plants under the attack of the fall armyworm. Additionally, a theoretical model is presented, built from the results of the machine learning analysis, to infer this damage with up-to-date orbital sensors. The data was composed of the reflectance measurements collected at a cotton field with control plants and plants submitted to Spodoptera frugiperda damage. Their spectral response was recorded with a hand-held spectroradiometer ranging from 350 to 2,500 nm, for eight consecutive days. Different machine learning models were evaluated and the overall best model was defined by accuracies comparisons on a testing-set. A ranking approach was adopted based on the model accuracy, returning the most contributive wavelengths for the classification. Sequentially, an unsupervised neural network (Self-Organizing Map - SOM) was implemented to reduce data-dimensionality and assist in the definition of important spectral regions. The regions were associated with the spectral bands of the two sensors (OLI and MSI) and a theoretical model using a band simulation process with the overall best machine learning model was proposed. The results indicated that the Random Forest (RF) algorithm is the most suitable to predict cotton-plants damaged by insects and that the last day of analysis (8th day) was better to separate it, with F-measure equals 0.912. The ranking approach combined with the SOM method indicated the spectral regions at the red to near-infrared (650 to 1,350 nm) and shortwave infrared (1,570 to 1,640 nm) as the most important regions to the analysis. The proposed theoretical model simulated with the OLI and MSI sensor-bands returned an F-Measure of 0.865 and 0.886, respectively. In conclusion, this framework can be used to map cotton-plants under insect-attack. The theoretical model presents high accuracy to infer the insect-damaged on cotton plants based on multispectral bands from other sensors, being a useful tool for future research that intends to evaluate it in other areas and at different field scales.

Riparian zones consist of important environmental regions, specifically to maintain the quality of water resources. Accurately mapping forest vegetation in riparian zones is an important issue, since it may provide information about numerous surface processes that occur in these areas. Recently, machine learning algorithms have gained attention as an innovative approach to extract information from remote sensing imagery, including to support the mapping task of vegetation areas. Nonetheless, studies related to machine learning application for forest vegetation mapping in the riparian zones exclusively is still limited. Therefore, this paper presents a framework for forest vegetation mapping in riparian zones based on machine learning models using orbital multispectral images. A total of 14 Sentinel-2 images registered throughout the year, covering a large riparian zone of a portion of a wide river in the Pontal do Paranapanema region, São Paulo state, Brazil, was adopted as the dataset. This area is mainly composed of the Atlantic Biome vegetation, and it is near to the last primary fragment of its biome, being an important region from the environmental planning point of view. We compared the performance of multiple machine learning algorithms like decision tree (DT), random forest (RF), support vector machine (SVM), and normal Bayes (NB). We evaluated different dates and locations with all models. Our results demonstrated that the DT learner has, overall, the highest accuracy in this task. The DT algorithm also showed high accuracy when applied on different dates and in the riparian zone of another river. We conclude that the proposed approach is appropriated to accurately map forest vegetation in riparian zones, including temporal context.

In recent years, the impacts associated with changes in land use and land cover (LULC) over natural processes have received attention. Examining these changes can assist in urban/rural planning activities, temperature change analysis, and environmental monitoring. Land surface temperature (LST) is one of the main physical parameters used to measure the impacts proportioned by LULC changes. Several studies have investigated the implications of LULC changes over LST variations using remote sensing data. However, this dynamic remains unknown in multiple environmentally important areas, like the western region of Sao Paulo state, Brazil, where one last remnant of the Atlantic Forest biome is located. This paper demonstrates that significant LST differences exist in distinct LULC classes. Results indicated that bare soil and pasture areas contribute to increasing LST values, while water bodies and arboreous vegetation attenuate them. As the LULC pattern changes, it reflects on both LST and air temperature. Over a 30-year analysis in the studied area, the LST increased 3.50 °C (0.117 °C per year). This multi-temporal analysis was based on four winter/autumn periods: 1987, 1997, 2007, and 2017, using images from Landsat 5 Thematic Mapper sensor for 1987 to 2007 and Landsat 8 Operational Land Imager and Thermal Infrared Sensor sensors for 2017. A supervised classification was applied to create LULC maps for each period. These findings bring a multi-temporal environmental diagnosis to the region, which is important to optimize future environmental planning actions like the restoration of degraded areas. It is inferred that continuous monitoring of LULC dynamics is required to devise sustainable land use policies in favor of environmental protection and regional economic development.

Defining the area for final solid waste disposal is a complex task. In Brazil, there are mandatory (restrictive) and guidance regulations for landfilling. However, the analysis of landfills in operation under these regulations is still little explored, but it can bring important information to territorial management. This paper investigates whether landfills in western São Paulo state meet the restrictive criteria and their level of environmental and socioeconomic adequacy. 32 landfills were analyzed, according to four restrictive criteria and a suitability map produced in a previous work. Approximately 41% of operating landfills fail to comply with legislation. Of those in legal areas, only 25% are in areas of high level of environmental and socioeconomic adequacy. This demonstrates the importance of investigating the social and environmental impacts related to landfill according to both mandatory and guiding regulations, aiming at effective territorial management. A definição de área para disposição final de resíduos sólidos é uma tarefa complexa. No Brasil, existem regulamentos de caráter obrigatório (restritivo) e de orientação para a implantação de aterros. Todavia, a análise de aterros em operação segundo esses regulamentos é ainda algo pouco explorada, mas que pode trazer importantes informações à gestão territorial. Este trabalho investiga se os aterros no oeste do estado de São Paulo atendem aos critérios restritivos e qual o seu nível de adequabilidade ambiental e socioeconômica. Analisou-se 32 aterros, segundo quatro critérios restritivos, e um mapa de adequabilidade produzido em um trabalho anterior. Aproximadamente 41% dos aterros em operação descumprem a legislação. Daqueles em áreas legais, somente 25% encontram-se em áreas de alto nível de adequabilidade ambiental e socioeconômica. Isso demonstra a importância de se investigar os impactos socioambientais relacionados aos aterros segundo ambos os regulamentos, de caráter obrigatório e de orientação, visando uma eficaz gestão territorial.

Made available in DSpace on 2020-12-12T02:00:26Z (GMT). No. of bitstreams: 0 Previous issue date: 2020-03-01 This paper presents a framework based on machine learning algorithms to predict nutrient content in leaf hyperspectral measurements. This is the first approach to evaluate macro-and micronutrient content with both machine learning and reflectance/first-derivative data. For this, citrus-leaves collected at a Valencia-orange orchard were used. Their spectral data was measured with a Fieldspec ASD FieldSpec® HandHeld 2 spectroradiometer and the surface reflectance and first-derivative spectra from the spectral range of 380 to 1020 nm (640 spectral bands) was evaluated. A total of 320 spectral signatures were collected, and the leaf-nutrient content (N, P, K, Mg, S, Cu, Fe, Mn, and Zn) was associated with them. For this, 204,800 (320 x 640) combinations were used. The following machine learning algorithms were used in this framework: k-Nearest Neighbor (kNN), Lasso Regression, Ridge Regression, Support Vector Machine (SVM), Artificial Neural Network (ANN), Decision Tree (DT), and Random Forest (RF). The training methods were assessed based on Cross-Validation and Leave-One-Out. The Relief-F metric of the algorithms' prediction was used to determine the most contributive wavelength or spectral region associated with each nutrient. This approach was able to return, with high predictions (R2), nutrients like N (0.912), Mg (0.832), Cu (0.861), Mn (0.898), and Zn (0.855), and, to a lesser extent, P (0.771), K (0.763), and S (0.727). These accuracies were obtained with different algorithms, but RF was the most suitable to model most of them. The results indicate that, for the Valencia-orange leaves, surface reflectance data is more suitable to predict macronutrients, while first-derivative spectra is better linked to micronutrients. A final contribution of this study is the identification of the wavelengths responsible for contributing to these predictions Federal University of Mato Grosso do Sul (UFMS) Environmental and Regional Development University of Western São Paulo (UNOESTE) Department of Cartographic Science São Paulo State University (UNESP) Department of Agronomy University of Western São Paulo (UNOESTE) Forest Engineering Department Santa Catarina State University (UDESC) National Research Center of Development of Agricultural Instrumentation Brazilian Agricultural Research Agency (EMBRAPA) Department of Geography and Environmental Management and Department of Systems Design Engineering University of Waterloo (UW) Department of Cartographic Science São Paulo State University (UNESP)

Thiswork aimsto identifythe restrictive and ponderable criteria for the identification of potential areas for implantation of a sanitary landfill. The method wasdivided into (1) consultation of laws, technical standards and resolutions; (2) consult the recommendations Recebido em: 27/08/2018Aprovado em: 05/09/2018 24Colloquium Exactarum, v. 10, n.3,Jul-Set. 2018, p.23–37. DOI: 10.5747/ce.2018.v10.n3.e243related to the object of study (landfill); and (3) description of restrictive and ponderable criteria. The first result was a general overview of the main laws, resolutions, norms and ordinances related to the implementation of a sanitary landfill. As a second result, we obtained the criteria of restrictive and ponderable order. There are several ponderable criteria, however, due to the difficulty of finding some data and the generality of some information, one can choose to consider only the ponderable criteria pertinent to the study area. We conclude that this work contributes to the ease of understanding and access to the laws, resolutions, norms and ordinances in force, related to the implantation of a sanitary landfill. However, there is still a need for the development of environmental permits, as required by the environmental agency, in which they require more detailed studies and specific analyzes to define the area for landfill.

O processo de selecao de area para aterros sanitarios pode ser realizado com uma Analise Multicriterio de Decisao (MCDA) em um ambiente SIG. No entanto, as variaveis ambientais precisam ser modeladas adequadamente antes que qualquer MCDA seja realizada. Uma maneira de fazer isso e integrar o MCDA com a Inferencia Fuzzy. Ainda assim, ate onde sabemos, essa estrategia nao foi avaliada na selecao de areas para aterros do consorcio. Aqui, propomos uma nova estrutura para selecionar areas de aterro sanitario de consorcio, integrando MCDA e logica Fuzzy com caracteristicas ambientais e socioeconomicas em ambiente GIS. Para isso, consideramos multiplos niveis de adequacao socioambiental para atender um grande e pequeno numero de cidades em um periodo de 20 anos. Uma regiao do interior de Sao Paulo, Brasil, com 32 municipios foi selecionada como area de estudo e o metodo da Combinacao Linear Ponderada (WLC) foi aplicado para combinar variaveis ambientais e socioeconomicas. Os pesos foram determinados utilizando o Analytical Hierarchy Process (AHP) e a logica Fuzzy foi aplicada para modelar as variaveis. Nossa abordagem foi capaz de identificar areas com niveis satisfatorios de adequacao a aterros sanitarios, atendendo a um grande e pequeno numero de municipios, minimizando os impactos socioambientais. Essa abordagem e importante porque pode ajudar a direcionar as entidades governamentais para uma selecao apropriada de uma area dentro de suas fronteiras regionais. Tambem recomendamos o metodo proposto para outras regioes e locais. A new strategy to define consortium sanitary landfill siting location based on multicriteria analysis A B S T R A C T The area selection process for sanitary landfills can be performed with a Multicriteria Decision Analysis (MCDA) in a GIS-environment. However, environmental variables need to be properly model before any MCDA is performed. One manner to perform this is to integrate MCDA with Fuzzy Inference. Still, to the best of our knowledge, this strategy was not evaluated in the selection of areas for consortium landfills. Here, we propose a new framework to select consortium landfill areas by integrating MCDA and Fuzzy logic with environmental and socioeconomic features in GIS environment. For this, we considered multiple levels of social-environmental suitability to attend a large and a small number of cities in a 20-years period. A region in the interior of Sao Paulo, Brazil, with 32 municipalities was selected as study area and the Weighted Linear Combination (WLC) method was applied to combine environmental and socioeconomic variables. Weights were determined using the Analytical Hierarchy Process (AHP) and Fuzzy logic was applied to model the variables. Our approach was able to identify areas with satisfactory suitability levels for landfills, both serving a large and a small number of municipalities while minimizing social-environmental impacts. This approach is important as it may help directs governmental entities to a appropriate selection of an area within their regional boundaries. We also recommend the proposed method for other regions and locations. Keywords: Analytical Hierarchy Process, Environmental Impact, Fuzzy Logic, Solid Waste.

Rare-earth elements have gained significant attention as they are currently widely used in high tech, chemical, and pharmaceutical industries. Here we used the fiddler crabs Leptuca leptodactyla as bioindicators to verify the presence of rare-earth elements in two mangrove areas of the Ubatuba, northern littoral of São Paulo state, Brazil. The specimens were collected in the mangrove areas of the Tavares River and Quiririm-Puruba river system, separated by season (dry and rainy). A total of 243 individuals were collected and analyzed. For determination and quantification of the elements we used the instrumental neutron activation analysis (INAA) technique. In both the dry and rainy season, the elements La, Ce, Sm, Eu, Tb, Yb and Sc were detected in samples of both mangroves, with La and Ce presenting higher concentrations. Samples from Tavares River mangrove had higher concentration levels of rare-earth elements than those of the Quiririm-Puruba river system. That is probably due to the central geographic position of the Tavares River in Ubatuba, which crosses a large portion of the city and receives a great amount of sewage. On the other hand, the Quiririm-Puruba river system has less anthropogenic inputs, thus, it receives much fewer rare-earth elements when compared to the Tavares River.

O processo erosivo e um fenomeno que acontece devido as condicoes climaticas ou uso inadequado da terra. O mapeamento dos niveis de vulnerabilidade a erosao de uma area pode ocorrer usando diferentes modelos de inferencia geografica. No entanto, definir o metodo apropriado e ainda uma questao a ser respondida. Este trabalho apresenta uma abordagem de validacao de mapa de vulnerabilidade a erosao elaborado por diferentes metodos de inferencia. Como estudo de caso, adotou-se uma bacia hidrografica e considerou-se os seguintes criterios: geomorfologia, pedologia, declividade, densidade de drenagem e cobertura da terra. Dentre os metodos testados tem-se: Combinacao Linear Ponderada (CLP) e tres operadores Fuzzy: soma algebrica, produto algebrico e gamma, variando o expoente “γ” entre os valores 0,4; 0,6 e 0,8. Os pesos dos criterios foram definidos com base no Processo Analitico Hierarquico. A validacao dos mapas ocorreu usando 1902 pontos, sendo 951 pontos de erosao na area, definidos com base em imagens do Google Earth Pro, e 951 pontos sem erosao, gerados aleatoriamente no QGIS 3.8. O modelo de regressao logistica foi usado parar comparar o desempenho de cada mapa ao apontar as areas com maior e menor grau de vulnerabilidade. A melhor modelagem foi alcancada com o operador Fuzzy gamma quando parametrizado com γ = 0,6. Embora o CLP seja a abordagem recorrente em estudos ambientais envolvendo inferencia geografica, nossos resultados demostram que outros operadores podem produzir resultados mais proximos aos encontrados com a realidade observada em campo. Machine learning erosion and vulnerability map validation A B S T R A C T Erosion is a natural phenomenon that happens in all ecosystems, whether due to weather conditions or inappropriate land use. Mapping the erosion vulnerability levels of an area can occur using different methods of geographic inference. However, defining the appropriate method is still a question to be answered. This paper presents an erosion vulnerability map validation approach elaborated by different inference methods. As a case study, a watershed was adopted and the following criteria were considered: geomorphology, pedology, slope, drainage density and land cover. Among the tested methods are: Weighted Linear Combination (WLC) and three Fuzzy operators: algebraic sum, algebraic product and gamma, varying the exponent “γ” between the values 0.4; 0.6 and 0.8. The weights of the criteria were defined based on the Hierarchical Analytical Process. The validation of the maps took place using 1902 points, with 951 erosion points in the area defined based on Google Earth Pro images and 951 points without erosion randomly generated in QGIS 3.8. The logistic regression model was used to compare the performance of each map by pointing out the areas with the highest and lowest degree of vulnerability. The best modeling was achieved with the Fuzzy gamma operator when parameterized with γ = 0.6. Although WLC is the recurring approach in environmental studies involving geographic inference, our results show that other operators can produce results closer to those encountered with the reality observed in the field. Keywords: Geographical inference; multicriteria analysis; data validation; environmental impact.

Doenças transmitidas por mosquitos estão cada vez mais frequentes e diversificadas em todo mundo. Isto é provável que esteja apenas no estágio inicial de ondas epidêmicas. O objetivo deste trabalho é mapear os casos de dengue na região do Pontal do Paranapanema no período de 2007 a 2012. Devido à falta de informações sobre a dengue no Pontal, a principal contribuição deste trabalho é auxiliar na visualização de quais municípios correspondem às áreas mais críticas dessa doença e se há correlação da mesma com as questões de urbanização e de saneamento do município. A distribuição espacial dos casos de dengue foram representadas em mapas temáticos do tipo coroplético. Conclui-se que a urbanização possui uma relação direta com o aumento dos casos de dengue, visto que municípios com maiores índices de urbanização são também com maiores taxas de dengue. Outros resultados serão discutidos.