Fuzzy Image Segmentation for Urban Land-Cover Classification

Abstract

ABSTRACT:In this paper a general fuzzy approach for segmentation-based classification is proposed. Traditional segmentation techniques focus on partitioning imagery into image-objects with well-defined boundaries. Instead, the proposed methodology aims to produce and analyze fuzzy image-regions expressing degrees of membership to different target classes. This approach, called Fuzzy Image-Regions Method (FIRME), is suitable to deal with the spectrally and spatially complexity of urban landscapes. The main stages of the FIRME approach are described, including techniques to produce such regions, alternatives to measure region attributes, and a number of methods for defuzzification. The FIRME method is tested for an urban classification experiment using multi-spectral imagery from Bogota, Colombia. Results suggest that, in complex environments, the FIRME method may be a suitable alternative to hard segmentation as it performs well in discriminating between spectrally mixed geographic objects.

Batch Mode Active Learning for Multimedia Pattern Recognition

“…We propose an optimization-based strategy for batch selection from an unlabeled set and exploit the properties of sub-modular functions to derive an efficient solution with provable performance guarantees. Although validated only on facial expression data in this work, the proposed framework is generic and can be used in any application involving fuzzy labels (document classification [244] or image segmentation [245], for instance). Our BMAL framework for fuzzy label problems is based on the notion of fuzzy sets and membership functions.…”

Supporting Global Environmental Change Research: A Review of Trends and Knowledge Gaps in Urban Remote Sensing

“…Raw data sources typically include aerial photography and multispectral and hyperspectral optical sources, such as those described thus far. Methods for classifying urban land covers include manual classifiers [19], fuzzy and hard classifiers [20], expert systems [14,21], object-based methods [22], machine learning [23], subpixel [24], and urban spectrometry [25].…”

Quantitative land cover change analysis using fuzzy segmentation

“…3 shows the image analysis workflow for estimating impervious surface percentages using the FIRME approach. Similarly to the procedure proposed for discrete classification (Lizarazo and Barros, 2010), it involves three main stages: (i) fuzzy segmentation (i.e. fuzzification); (ii) feature analysis; and (iii) regression (i.e.…”

“…This trend has been highlighted in recent review of GEOBIA literature (Hay and Blaschke, 2010). A large percentage of most recent articles focus on evaluating advantages of GEO-BIA classifications over traditional pixel-based classifications, see, for example, Johansen et al (2010), Kim et al (2010), Lizarazo and Barros (2010). GEOBIA classifications are esentially discrete or crisp classifications of land cover, which allocate categorical labels to pixels.…”

“…This paper explores the potential of the Fuzzy Image Regions Method (FIRME) for continuous classification of land cover. FIRME was recently proposed as a new GEOBIA method for discrete classification of urban land cover (Lizarazo and Barros, 2010). Fuzzy image regions are image objects with indeterminate boundaries and uncertain thematic content which are produced by a fuzzy Table 1 Landsat TM-5 spectral bands.…”