Differences
This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision | ||
thales:leila09062010 [2010/06/08 16:39] tkorting |
thales:leila09062010 [2010/06/10 15:29] (current) tkorting |
||
---|---|---|---|
Line 1: | Line 1: | ||
====== Leila ====== | ====== Leila ====== | ||
+ | |||
+ | Reunião do dia 09/06/2010: | ||
+ | |||
+ | * [[thales:leila09062010#conteudo|Conteúdo]] | ||
+ | * [[thales:read_papers|Papers lidos]] | ||
+ | * [[thales:ongoing_papers|Papers em andamento]] | ||
===== Conteúdo ===== | ===== Conteúdo ===== | ||
Line 21: | Line 27: | ||
{{:thales:sits_scatterplot.png?300}} | {{:thales:sits_scatterplot.png?300}} | ||
- | //Uma imagem SITS NDVI, ao ser mensionada, não se pode esquecer qeu vários outros atributos são tidos como estáticos, para poder dizer que uma mudança no scatterplot indique uma mudança ambiental. (Gilberto) // | + | //Ao referenciar uma imagem SITS NDVI não se pode esquecer qeu vários outros atributos são tidos como estáticos, para poder dizer que uma mudança no scatterplot indique uma mudança ambiental. (Gilberto) // |
==== Segmentação ==== | ==== Segmentação ==== | ||
Line 36: | Line 42: | ||
{{:thales:fuzzy_decision_trees.png}} | {{:thales:fuzzy_decision_trees.png}} | ||
- | |||
- | ===== Papers lidos ===== | ||
- | |||
- | ===== Papers em andamento ===== | ||
- | |||
- | * **Interpreting Images with GeoDMA.** [[http://www.dpi.inpe.br/~tkorting/docs/korting2010interpreting.pdf|Paper]]. Thales Sehn Korting, Leila Fonseca, Gilberto Camara. **GEOBIA 2010.** //Object oriented analysis offers effective tools to represent the knowledge in a scene. Knowledge-based interpretation arises as an effective way to interpret remote sensing imagery. In this approach, human’s expertise is organized in a knowledge base to be used as input of automated interpretation processes, thus enhancing performance and accuracy, and reducing at the same time the subjectivity in the interpretation process. Some systems such as Definiens, ENVI-FX, and more recently, InterIMAGE, have incorporated useful tools to aid the object-oriented classification. In this context, this work presents a system for object image analysis, called Geographical Data Mining Analyst (GeoDMA), which implements several mechanisms for automatic image interpretation. It performs image segmentation, objects feature extraction, supervised and unsupervised classification with raster, shape and cellular data sets. GeoDMA has been used for intra-urban land cover classification of high spatial resolution images and also to detect deforestation patterns in the Brazilian Amazon. Such different applications warrant the generalist behavior of the system. In this article, we discuss each module implemented in the system, the integration of objects obtained by segmentation to extract texture features and landscape metrics. We also show the main tools for data preprocessing, sample selection and feature visualization. Finally, one application of automatic urban interpretation is shown.// | ||
- | |||
- | * **GeoDMA - A software for geographical pattern recognition.** Thales Sehn Korting, Leila Fonseca, Gilberto Camara. **EMS 2010?** //This work presents a novel software for pattern recognition applied to geographical and remote sensing data. GeoDMA stands for Geographical Data Mining Analyst, and is a free software designed to include data mining techniques to aid domain-experts in the detection of land cover and land use patterns in geographical databases. For this purpose, GeoDMA includes spatial and spectral features extraction, visual data analysis and supervised and unsupervised classification algorithms. GeoDMA has already been applied to detect intra-urban classes of land cover and to detect patterns of deforestation in the Brazilian Amazon.// | ||
- | |||
- | * **A re-segmentation approach to detect rectangular objects in high resolution imagery.** Thales Sehn Korting, Luciano Dutra, Leila Fonseca. **GRSL, 2010?** //Image segmentation covers techniques for splitting one image into its components. This paper presents a re-segmentation approach applied to urban images. The interest elements, such as houses roofs, are considered to have a rectangular shape. Our technique finds and produces rectangular objects, setting to background the remaining elements to background. With an over-segmented image we connect bordering elements in a graph, known as {\it Region Adjacency Graph} -- RAG. By going into the graph, we search for the best cuts that may result in more rectangular objects, using a relaxation-like approach. Results show that the method was able to match rectangles, according to user-defined parameters, such as maximum levels of graph depth search and minimum degree of rectangularity for interest objects. // | ||
- | |||
- | * **GeoSOM applied to image segmentation.** Thales Sehn Korting, Leila Fonseca. **SIBGRAPI 2010?** //Image segmentation is one of the most challenging steps in image processing. Its results are used by many other tasks regarding information extraction from images. In remote sensing, segmentation generates regions according to found targets in a satellite image, as roofs, streets, trees, vegetation, agricultural crops, or deforested areas. Such regions are used to differentiate land uses by classification algorithms. In this paper we investigate a way to make segmentation using a strategy to classify and merge spectrally and spatially similar pixels, using an extension of the Self-Organizing Maps algorithm, named GeoSOM. Neurons will have similar properties to the main elements found in the image. By associating a different label to each image pixel, according the most similar neuron, neighboring pixels with the same label are merged into segments.// | ||
- | |||
- | * **Use of an Improved Expectation-Maximization approach for Remote Sensing Data Classification.** Thales Sehn Korting, Luciano Dutra, Leila Fonseca, Guaraci Erthal. **CIARP 2010?** //In statistical pattern recognition, mixture models allow a | ||
- | formal approach to unsupervised learning. This work aims to present a modification of the Expectation-Maximization clustering method applied to remote sensing images. The stability of its convergence has been increased by supplying the results of the well-known k-Means algorithm, as seed points. Hence, the accuracy has been improved by applying cluster validity measures to each configuration, varying the initial number of clusters. High-resolution urban scenes has been tested, and we show a comparison to standard supervised classification results. Performance tests were also realized, showing the improvements of our proposal, in comparison to the original one.// |