An algorithm for segmenting images based on local texture properties and which requires only linear filtering operations has been described in a previous paper. We now report on the implementation of this algorithm in a hybrid electro-optical system which performs the linear filtering operations in the spatial frequency domain. Liquid-crystal video displays (LCDs) are used for input of the image and the frequency-domain filter. The system has the potential for rapid segmentation of simple images by texture properties. . INTRODUCTIONNatural scenes often are composed of irregularly shaped macroscopic regions which are distinguished only by the small-scale ("microscopic") local order of pixel gray levels. This local patterning commonly is called 'texture", and may be described as a pseudoperiodic nonrandom arrangement of gray levels'. Identification and segmentation of these textural regions is a common image processing problem, particularly in remote sensing.The information about the texture of a region appears in the Fourier spectrum as concentrations of amplitude about the spatial frequency(s) characteristic of the microscopic texture pattern. The shape of the macroscopic region which contains that texture determines the local distribution of amplitude in the Fourier spectrum about the characteristic spatial frequency(s). The spectrum of a strictly periodic texture pattern will exhibit sharp concentrations of amplitude at the characteristic spatial frequencies in the spectrum, while that of a pseudoperiodic texture will be less well delineated.Stromberg and Farr2 described an algorithm for segmenting regions based on texture by applying a sequence of annular filters to the Fourier spectrum to identify and isolate regions in the image that contain specific texture patterns. Use of annular filters ensures that the result is insensitive to orientation of the texture pattern, but the raw filtered images include the carrier frequencies of the texture components. The carrier was removed by a subsequent lowpass filter applied to the magnitude of the bandpass filtered images. The resulting set of filtered images can be used to define a signature for each texture class. The method was used to discriminate geological textures in synthetic aperture radar (SAR) images.Previously, we reported a modified algorithm which gives results similar to those of Stromberg and Farr but may be implemented more easily in a coherent electro-optical system3. The method generates a large number of images from single-sideband bandpass filters. By computing the magnitude (or squared magnitude) of the filtered images, the lowpass filter required by the Stromberg method is eliminated. The bandpass-filtered magnitudes within each passband may be digitized and summed to create a single image for each annular band analogous to the images obtained by the earlier method. O-8194-1538-3/94/$6.OO SPIE Vol. 2234 / 265 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/24/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.