On the Extensive Reconstruction of Hough-type Transform for Line Pattern Detection
スポンサーリンク
概要
- 論文の詳細を見る
The basic function of the Hough transform, defined byρ=x cosθ+y sinθ, is not a unique functional expression for line pattern detection. This paper extends the functional expression so that line patterns can be detected by means of an extended version of Hough transform functions, called EHT-functions. Three conditions required by a family of EHT-functions are introduced, and a few basic examples such as the "Hough-Parabolic function" are presented. Furthermore, two kinds of fast transform functions, Piecewise Linear Hough Transform (PLHT) and Fast Incremental Hough Transform (FIHT), are proposed in order to show the practical usefulness of the EHT-functions. As PLHT uses a piece-wise linear transform function and FIHT can be executed without multiplication operations, the computational cost of these fast Hough transforms can be reduced to about 1/4 or 1/6 of that of the traditional Hough transform. Definitions, basic properties, and the effects of the computational cost reduction of the two functions are briefly presented. It is also shown that a line detector based on the extended Hough-type function is more sensitive in detecting the angles of line pattern segments.
- 一般社団法人情報処理学会の論文
- 1992-10-31
著者
-
Koshimizu H
Chukyo Univ. Toyota‐shi Jpn
-
Numada Munetoshi
Lossev Technology Coap
-
Numada M
Nippei Toyama Co. Ltd. Toyama‐shi Jpn
-
KOSHIMIZU HIROYASU
Department of Computer Science, School of Computer and Cognitive Sciences (SCCS), Chukyo University
-
Koshimizu Hiroyasu
Department Of Computer Science School Of Computer And Cognitive Sciences (sccs) Chukyo University
関連論文
- A RELATIONSHIP BETWEEN OK-QUANTIZATION THEORY AND SAMPLING THEOREM AND SOME EXPERIMENTAL CONSIDERATIONS(International Workshop on Advanced Image Technology 2006)
- On the Extensive Reconstruction of Hough-type Transform for Line Pattern Detection
- An Efficient Algorithm for Generating Slanted Ellipse Using Simultaneous Recurrences