Morphological operators for segmentation of high contrast textured regions in remotely sensed imagery

Abstract

We develop a transformation based on morphological filters that measures the contrast of image texture. This transformation is proportional to texture contrast, but insensitive to its specific type. Though the transformation provides a high response in textured areas, it suppresses individual high contrast features that stand apart from textured areas. It can serve as an effective texture descriptor for unsupervised or supervised segmentation of textured regions, provides high accuracy of localization and does not involve heavy computations. The method is robust to variations of illumination and works on different types of images without needing to be tuned. The only parameter is a scale related parameter. We illustrate the use of the proposed method on satellite and aerial images.

A morphological approach for distinguishing texture and individual features in images

“…Contrast operator and the ASF difference Below, we define the morphological texture contrast (MTC) transformation ψ MTC ( f ) that we recently introduced in Zingman et al (2012) for distinguishing texture regions (such as forests, urban areas and rocky mountains) in satellite images from smooth areas, which may also contain individual structures that should not to be assigned to texture 2 . Qualitatively, MTC's response is summarized in the first row of Table 1. The MTC is based on alternating morphological filters, γ r ϕ r and ϕ r γ r , which are closing ϕ followed by opening γ and opening followed by closing, respectively.…”

“…2, measures the difference between upper and lower texture envelopes estimated by means of alternating morphological filters (Serra and Vincent, 1992;Soille, 2003). Its qualitative performance was illustrated in Zingman et al (2012), where only few remotely sensed images were used and no quantitative comparison was provided. In Sec.…”

A Comparison of Two Morphological Techniques in the Classification of Urban Land Cover

“…Commonly, edge-based and region-based segmentation are used to partition discrete surface spectral characteristics. Edge-based segmentation, on the other hand, seeks boundaries by distinguishing areas within the image and segments of the complete enclosure by edge pixels [21]. Consequently, pixels that characterize objects can either form part of the in-segment or constitute a segment as its boundary [22].…”

“…Consequently, pixels that characterize objects can either form part of the in-segment or constitute a segment as its boundary [22]. However, the major limitations related to edge-based techniques are insensitivity to noise and are edge-based, hence they are highly dependent on the analysis window which blurs the borders of textured regions [18,21].Morphological image analysis techniques, generally referred to as mathematical morphology (MM) [23], exploit the spatial domain in images using various techniques based on set theory to estimate and measure many useful geometrical features such as shape, size, and connectivity [21,23]. These techniques are developed based on concatenation of mathematical operations grounded in a set of operations such as union, intersection, complementation, and translations [23].…”

Detection of incomplete enclosures of rectangular shape in remotely sensed images

“…We discarded all candidate points having a distance smaller than 10 or greater than 90 pixels, which limits the distances between opposite walls of the structures. High contrast texture regions were filtered out using the Morphological Texture Contrast descriptor [40,39,37] thresholded with the Otsu method [27]. This filters out urban areas, forests, rocky mountains, and other high contrast texture regions, but preserves individual structures.…”