Hk. Aghajan et T. Kailath, SLIDE - SUBSPACE-BASED LINE DETECTION, IEEE transactions on pattern analysis and machine intelligence, 16(11), 1994, pp. 1057-1073
An analogy is made between each straight line in an image and a planar
propagating wavefront impinging on an array of sensors so as to obtai
n a mathematical model exploited in recent high resolution methods for
direction-of-arrival estimation in sensor array processing. The new s
o-called SLIDE (Subspace-Based Line Detection) algorithm then exploits
the spatial coherence between the contributions of each line in diffe
rent rows of the image to enhance and distinguish a signal subspace th
at is defined by the desired line parameters. SLIDE yields closed-form
and high resolution estimates for line parameters, and its computatio
nal complexity and storage requirements are far less than those of the
standard method of the Hough transform. If unknown a priori, the numb
er of lines is also estimated in the proposed technique. The signal re
presentation employed in this formulation is also generalized to handl
e grey-scale images as well. The technique has also been generalized t
o fitting planes in 3-D images. Potential application areas of the pro
posed technique include road tracking in robotic vision, mask-wafer al
ignment and linewidth measurement in semiconductor manufacturing, aeri
al image analysis, text alignment in document analysis, particle track
ing in hubble chambers, and similar applications.