Nelson D. A. Mascarenhas

Use of synthetic bands derived from mixing models in the multispectral classification of remote sensing images

An analysis of the use of features derived from class proportions in a pixel for the multispectral classification of reforested areas in Landsat images is performed. Through a linear mixing model, synthetic bands derived from those proportions are obtained either using the constrained or the weighted least squares procedures. The method indicates that the synthetic bands offer an alternative to well-known dimensionality reduction techniques such as principal components or canonical analysis. Furthermore, those bands provide a useful tool for visual interpretation, since they contain information that is related to physical concepts (proportions) more easily assimilated than class spectral signatures.

Multispectral image data fusion under a Bayesian approach

Abstract Multispectral image data fusion is understood here as a process that generates synthetic images from a combination of primary images, by attempting to preserve the best characteristics of each primary image. The obtained product is important in helping the users of remote sensing in visual analysis. This paper describes a new method for multispectral image data fusion using a Bayesian framework. As an example, the method is applied to the synthesis of new bands for the SPOT satellite. The method comprises two steps: (a) a new interpolator for the multispectral bands, obtained through the orthogonality principle and leading to the estimator and its covariance.matrix, which will be used as a priori information for the second step; and (b) a new statistical synthesis formulation, also obtained through the orthogonality principle, using as observations the panchromatic and the multispectral bands. Experimental results with SPOT images over Guarulhos Airport, Sao Paulo, Brazil, are presented, includin...

Multispectral image data fusion using POCS and super-resolution

The problem of image data fusion coming from different sensors imaging the same object is to try to obtain a result that integrates the best characteristics of each one of those sensors. In this work, we want to combine the characteristics of multispectral (better spectral definition) and panchromatic (better space definition) images, using the bands from the satellites Landsat-7 (panchromatic) and CBERS-1--China-Brazil Earth Resources Satellite (four multispectral bands). The process proposes solutions using projection onto convex sets (POCS) techniques divided in two steps: (a) interpolated image processing, regularizing the block artifacts and using super-resolution techniques based on POCS and (b) synthesis, obtained by sequential and parallel projections or by the least squares method.

Digital image restoration under a regression model

In this paper, image-restoration techniques based upon a regression model are analyzed and verified by computer simulation. A regression model is formulated to describe image blurring, additive noise, physical image sampling, and quadrature representation. Classical estimation methods utilized for image restoration are described and related to one another. Restorations obtained by these classical techniques are shown to be poor because of noise disturbances and the ill conditioning of the image-degradation regression model. Constrained restoration methods which avoid ill conditioning problems are introduced. Computer simulations demonstrate that a boundedness constraint on the brightness of a reconstructed image provides significantly improved restorations as compared to unconstrained methods.

Some experiments with spatial feature extraction methods in multispectral classification

Abstract Feature extraction is an important factor in determining the accuracy that can be attained in the classification of multispectral images. The traditional per point classification methods do not use all the available information, since they disregard the spatial relationships that exist among pixels belonging to the same class. In this paper, methods are developed to extract additional image spatial features by means of linear and non-linear local filtering. Feature selection methods are also developed, since it is usually not possible to use all the generated features. The classification stage is performed in a supervised mode using the maximum likelihood criterion. A quantitative analysis of the performance of the spatial features show that an overall increase in precision of classification is achieved, although at the expense of increased rejection levels, particularly on the borders between different fields.

Edge preserving wavelet speckle filtering

A new edge preserving wavelet filtering is proposed and it is applied to real SAR images that are generally affected by a multiplicative noise, called speckle, which degrades the quality of these images. The new approach attempts to look for the neighborhood area in the detail images of a wavelet decomposition, that identifies homogeneous areas and edge information by using masks in order to reduce speckle while edges are preserved The improved filtering method uses the Nagao and Matsuyama and Tomita and Tsuji masks to detect edge locations in wavelet subspaces. The information provided by the masks is used to distinguish which of the detail coefficients, are to be shrunk.

A discrete approach for supervised pattern recognition

We present an approach for supervised pattern recognition based on combinatorial analysis of optimum paths from key samples (prototypes), which creates a discrete optimal partition of the feature space such that any unknown sample can be classified according to this partition. A training set is interpreted as a complete graph with at least one prototype in each class. They compete among themselves and each prototype defines an optimum-path tree, whose nodes are the samples more strongly connected to it than to any other. The result is an optimumpath forest in the training set. A test sample is assigned to the class of the prototype which offers it the optimum path in the forest. The classifier is designed to achieve zero classification errors in the training set, without over-fitting, and to learn from its errors. A comparison with several datasets shows the advantages of the method in accuracy and efficiency with respect to support vector machines.

Cell nuclei segmentation in noisy images using morphological watersheds

A major problem in image processing and analysis is the segmentation of its components. Many computer vision tasks process image regions after segmentation, and the minimization of errors is then crucial for a good automatic inspection system. This paper presents an applied work on automatic segmentation of cell nuclei in digital noisy images. One of the major problems when using morphological watersheds is oversegmentation. By using an efficient homotopy image modification module, we prevent oversegmentation. This module utilizes diverse operations, such as sequential filters, distance transforms, opening by reconstruction, top hat, etc., some in parallel, some in cascade form, leading to a new set of internal and external cell nuclei markers. Very good results have been obtained and the proposed technique should facilitate better analysis of visual perception of cell nuclei for human and computer vision. All steps are presented, as well as the associated images. Implementations wee done in the Khoros system using the MMach toolbox.

Transmission tomography under Poisson noise using the Anscombe transformation and Wiener filtering of the projections

Abstract A minitomograph scanner for soil science was developed by the National Center for Research and Development of Agricultural Instrumentation (EMBRAPA/CNPDIA). The purpose of this paper is twofold. First, a statistical characterization of the noise affecting the projection measurements of this scanner is presented. Second, having determined the Poisson nature of this noise, a new method of filtering the projection data prior to the reconstruction is proposed. It is based on transforming the Poisson noise into Gaussian additive noise, filtering the projections in blocks through the Wiener filter and performing the inverse tranformation. Results with real data indicate that this method gives superior results, as compared to conventional backprojection with the ramp filter, by taking into consideration both resolution and noise, through a mean square error criterion.

Pseudolikelihood Equations for Potts MRF Model Parameter Estimation on Higher Order Neighborhood Systems

This letter presents pseudolikelihood equations for the estimation of the Potts Markov random field model parameter on higher order neighborhood systems. The derived equation for second-order systems is a significantly reduced version of a recent result found in the literature (from 67 to 22 terms). Also, with the proposed method, a completely original equation for Potts model parameter estimation in third-order systems was obtained. These equations allow the modeling of less restrictive contextual systems for a large number of applications in a computationally feasible way. Experiments with both simulated and real remote sensing images provided good results.