National Institute for Space Research
Biblioteca Digital da Memória Científica do INPENot a member yet
6175 research outputs found
Sort by
Influência da interpolação na geração de MDTs a partir de pontos classificados LiDAR
No full textThe generation of digital terrain models (DTMs) from airborne LiDAR (Light Detection and Ranging) data is an important task for both ecological studies and forest management purposes. However, besides a proper ground filtering, a correct choice of the interpolation method is also important step. We evaluated three interpolation methods for the generation of DTMs (i.e. kriging, inverse distance weighted and nearest neighbor). Our study area is located in western Washington State (USA). The area contains different land use classes ranging from bare soil to shrubland, log forest and then dense forest. We compare the different interpolated DTM results each other and with a topographic map. The results showed that significant differences between interpolation methods were associated with abrupt changes in relief and also related to specific land use classes. However, the higher differences (i.e. up to 5m) were found using the nearest neighbor approach. The best methods for the selected area were kriging and inverse distance weighted. Further investigation should take into account the influence of different land use types on the interpolation results and evaluate the impact of topography on the extraction of individual trees.Pages: 6105-611
Classificação de espectros de dados hiperespectrais pelo método de sequência típica e modelo oculto de Markov
No full textThis work presents a new methodology for hyperspectral image classification based on two concepts. Former is the Typical Sequence determination, derived from the Information Theory, and the last is the Hidden Markov Chains (HMM). The HMM gives the probability of a given mixture belongs to the HMM model of an endmember (EM) and the Typical Sequence determination does the association of a mixture with a given EM. Five EM from an AVIRIS Scene and six mixtures, generated by a linear model, were used to perform a classification test. It was also done a classification with the SAM (Spectral Angle Mapper), the ED (Euclidian Distance) and with the SID (Self Information Divergence). In the test the proposed method produced the best results showing that it can be used as an alternative method for hyperspectral image classification.Pages: 9004-901
Estudo da alteração do uso e cobertura do solo no centro oeste maranhense entre os anos de 2000 e 2011
No full textThis study aims to analyze the temporal use and land cover between the years 2000 and 2011 in the Municipality of Grajaú-MA, located in the center west of the state, having cerrado vegetation and secondary forest. Using geoprocessing tools that aid through Software 5.0 SPRING worked up LANDSAT satellite images to quantify the entire area of use. What was found was an acceleration in land use. In a margin space of 11 years we can observe a change in the vegetation, which among the different years this vegetation has been replaced by another that would meet the economic needs of man. It was also observed a shift of the peasant to the periphery of the city, which in turn had to cede its space to house vegetation that meets the need of extensive livestock and monoculture. The analysis in question allowed quantifying the data caused by human action in a spatial field and confront the census data collected from other sources, which runs on an accelerated land use of 39.127,31 ha in 2000 to 157.360,79 ha in 2011, a representativeness of 75%, which corresponds to 118.233,48 ha.Pages: 7186-719
Monitoramento de mudanças na cobertura da terra na região da Floresta Nacional do Tapajós, Amazônia Brasileira, utilizando segmentação de imagens multitemporais
No full textThe main purpose of this paper is to present a semi-automated procedure for monitoring land cover changes in Brazilian Amazonia. For this, the region around the Tapajós National Forest in Pará State, Brazilian Amazonia was selected. The study area comprises approximately 10,650 km2 covered by the Landsat Thematic Mapper (TM) images acquired on 22 August 1989; 02 August 1999; and 28 July 2009. According to Radambrasil vegetation map, the study area is primarily covered by dense tropical rain forest (Floresta Ombrófila Densa) with a high number of emergent tree species. Firstly, the image segmentation technique was applied to multitemporal Landsat Thematic Mapper acquired over Tapajos National Forest region to generate the georeferenced polygons. Then the vegetation, soil and shade fraction images derived from TM images using Linear Spectral Mixing Model were used to classify these polygons in the land cover classes for each image date. Finally the Geographic Information System tools are used to map the changes occurred on the land cover and land use in the study site between the dates analysed. Then the results achieved by the proposed methodology were evaluated using the MODIS time series available in the LAF (Remote Sensing Laboratory for Agriculture and Forest applications) website (https://www.dsr.inpe.br/laf/series/en/index.html) at INPE and field data. The proposed methodology presents a potential tool for mapping and monitoring land cover and land use changes in the Brazilian Amazonia.Pages: 7786-779
Aplicação de métodos de interpolação espacial em dados batimétricos: estudo de caso do Alto curso do rio Paraná
No full textBathymetric map is a very important product for understand the dynamics morphological and hydraulics variables of rivers. The precision of bathymetric map is determined by spatial interpolation employed for to spatialize discrete points or cross sections. The aim of work was analysis methods of spatial interpolation of bathymetric data in generation bathymetric surface in one section of Upper Parana River. For this, were applied ordinary kriging and IDW spatial interpolations. The results obtained were bathymetric maps showed the continue variations of depth of channel. Overview the maps demonstrated similarity with exception of IDW spatial interpolation that created shallow and deep regions with concentric forms. By statistical evaluation the less valor of RMSE was found in anisotropic ordinary kriging with spatial trend extracted increased the performance in 16.6% at relation the same technique with spatial trend. However with extraction of spatial trend of bathymetric data for IDW increased about 50% of RMSE smoothing the depth of channel. Cross section in Upper Parana River validated that anisotropic ordinary kriging without spatial trend obtained more similarity with actual bathymetric variation, beyond showed less RMSE. The others spatial interpolation techniques used smoothing the morphology of channel not representing satisfactorily the variation of depth of Parana River.Pages: 4250-425
Análise de sensibilidade do sensor MERIS a gradientes espaciais de constituintes opticamente ativos
No full textIn order to assess the sensitivity of the MERIS sensor to changes in the concentration of optically active constituents (chlorophyll, suspended solids and dissolved organic matter) in the Amazon floodplain, regular grids were created by interpolating a set of in situ samples gathered during a seven day field work. Four interpolation algorithms (simple average; distance weighted average; nearest neighbour; ordinary kriging) were applied to the dataset in order to assess which one could be the best spatial descriptor to our data. The interpolation by Ordinary kriging algorithm was able to better represent the variability of all the three constituents. Horizontal profiles were plotted over the grids in order to verify the change gradient of each optically active constituent in a given direction. These same profiles were then applied on a MERIS image acquired concurrently with a day of sampling, and analysis of the sensor's ability to detect spatial changes of the constituents concentration were made, using the most appropriate spectral bands and band ratio to each constituent. Finally, the results of these two processes were compared. The MERIS sensor showed different sensibility to each type of optically active constituent, according to the spectral band considered, however, only the substantial variations were noted.Pages: 1517-152
Classificação de cobertura vegetal baseado na integração de dados multiespectrais TM-Landsat, mapa geomorfológico e dados altimétricos numa abordagem baseada em regras
No full textThe present work aims at presenting an approach adopted to produce a map of vegetation cover of the RPPN Foz do Rio Aguapeí. This map will help to design a management plan of this Conservation Unit known as RPPN Foz do Rio Aguapeí. This map was made based on TM-Landsat multispectral images, height data and geomorphological map. The classification system was established based on field observations as well as expert technical support. The vegetation class system was improved by applying a generalization process which was also carried out with experts and by associating the vegetation classes with geomorphological classes. Considering that it is a region that includes a plain of the Aguapei river and several raised shore lines, it was decided to adopt a classification approach based on an integration of multispectral data and local landscape information. It is important to notice that such landscape has great correlation with the relief shapes since the degree of humidity of each position depends on the proximity to the river and its elevation. Supervised classification techniques were applied with a region approach on scene of the dry season. Band 5 TM-Landsat image was sliced as well. In addition, a region classification was done following the geomorphology and the spectral response of each region. Boolean rules were specified to combine all these results together and height classes of the relief. Those rules were coded in Legal, from the Spring system. The findings are in accordance to the information needs of the vegetation cover in order to design a management plan for this RPPN.Pages: 3298-330
Monitoramento Ambiental para Identificação de Áreas Degradadas no Bioma Caatinga do Estado do Ceará - Brasil
No full textThe Caatinga Biome is threaten by firewood fetching for ceramics, lime and coal industries, besides by aperture of new agriculture areas. The remote sensing monitoring system widely used by IBAMA (DETER) was built to monitor the Amazon Biome and does not have efficacy on the Caatinga Biome. Aiming somehow to outmatch this lack and especially to provide data to assit IBAMA's Ceará Control Division, Geoprocessing Center of IBAMA Ceará developed a remote sensing procedure to monitor deforestation and biomass burning at the Caatinga Biome. The monitoring data produce subsidized the deflagration of 05 environmental control operations in Ceará, that resulted in R$ 1.139.067,00 penalties and the arrest of 825,49 hectares.Pages: 6985-699
Avaliação de Modelos Digitais de Elevação para aplicação na caracterização do ambiente cafeeiro na Região da Serra da Mantiqueira, MG
No full textThe digital elevation models (DEMs) are being widely used. Its applications can be related to prevention, planning and monitoring of activities in space, whether urban or rural. One of its main applications is linked to agriculture being a tool for strategic decision making to improve agricultural practices or reducing impacts on the environment. Measuring the accuracy of a DEM is a job that requires going to the field and check sampled points for the best models. This study aims to evaluate and compare the DEM with MDE generated from the contours of the IBGE. We performed trip to the field to collect sample points randomly in the study area. We applied the square root of the mean square error (RMSE) of the elevation of the extent of agreement Kappa. The altimetric variation between MDE made a small difference, and SRTM DEM presented the highest values in both the maximum altitude as the minimum. Regarding spurious depressions observed that the MDE SRTM showed a total of 10 times the value identified in the MDE curves IBGE. The drainage network extracted from DEMs showed a slight difference in their extensions, the MDE curves showed greater extent than the SRTM DEM. The RMSE of the two MDEs were low but still standard in accordance with a scale of 1:50,000.Pages: 4716-472
Abundância e diversidade da avifauna em quatro parques urbanos no município de Jundiaí, SP analisadas com o uso de sistemas de informação geográfica
No full textThe urban landscape promotes a complex relationship between the natural environment and people, and through the use of geoprocessing tools can spatially analyze these relations. In this study we are considering the hypothesis that the number of bird species is related to the areas of urban parks and their woodland surroundings areas. We selected four parks with different characteristics in the municipality of Jundiaí, SP and we use the methodology by point count and transects to inventory of birds. Through software (Webshot and GoogleMaps) an aerial image of the study area was obtained and georeferenced (ArcGIS 10). Through visual interpretation a shapefile was created, corresponding to existing woodland areas. This shapefile was confronted with circumferences centered in parks and radius of 1.000m and 2.000m to find coverage the woodland area around each park. Analyzing the number of species recorded in the park with the area of each park, and the woodland areas in buffer regions, no significant relationship was found until now. Relations was found between the number of contacts with the park area (F = 94.6917, p = 0.0077). Although our preliminary database shows that there is no relationship between number of species and the woodland area, such study is being continued, in order to confirm or reject the hypothesis above mentioned for a longer time.Pages: 1077-108