Browsing by Author "HO, SY"

Now showing 1 - 5 of 5

AN EFFECTIVE SEARCH APPROACH TO CAMERA PARAMETER-ESTIMATION USING AN ARBITRARY PLANAR CALIBRATION OBJECT
(1993-05-01)
CHEN, Z
;
WANG, CM
;
HO, SY
;
交大名義發表
;
資訊科學與工程研究所
In this paper we shall present a new method for camera parameter estimation using an arbitrary planar calibration object. We begin with an approximate model for the perspective view and give a parameter-free normalized-line-segment-ratio relationship between the corresponding reference and sensed views. This representation makes parameter decoupling possible and reduces the library dimension from six to one. Then the estimation problem can be viewed as a library search problem. Various mechanisms are provided to reduce the library search time, including (a) library partition by clustering, (b) sensed cluster identification by binary search, and (c) sequential testing for view matching using sorted features. In addition, feature perturbation is also considered to achieve better robustness against feature noise. Both computer-generated and real data are included in experiments, illustrating how our method works.
COMPUTER VISION FOR ROBUST 3D AIRCRAFT RECOGNITION WITH FAST LIBRARY SEARCH
(1991)
CHEN, Z
;
HO, SY
;
交大名義發表
;
資訊科學與工程研究所
;
National Chiao Tung University
A fast but accurate classification method is presented which takes the fundamental problems of aircraft recognition in 3D space into consideration. Elliptic Fourier descriptors are used for aircraft classification and pose determination from a two-dimensional image recorded at an arbitrary viewing angle. The computation model of the nearest neighbor classification rule (or, simply, NNR) is analysed to come up with proper necessary conditions used by our method for reducing the set of near neighbors of a test projection for a fast partial library search. On the other hand, our method is robust to noise by incorporating a perturbation analysis of feature variation into the classifier design. Both graphics-generated data and real data of five aircraft have been used to demonstrate the applicability of the new method. The results indicate that the method is almost as accurate as the NNR method, but is much faster. The method is also shown to outperform other existing classification methods including the decision-tree classification method.
INCREMENTAL MODEL-BUILDING OF POLYHEDRAL OBJECTS USING STRUCTURED LIGHT
(1993-01-01)
CHEN, Z
;
HO, SY
;
交大名義發表
;
資訊科學與工程研究所
;
National Chiao Tung University
A vision method using active sensing to construct the complete geometric model of a polyhedral object is presented. The model is constructed incrementally from multiple views. Each view produces a partial object model. Three different methods are described for integrating all the partial models into a final complete model. Theorems regarding the sufficient number of views required and how these views can be set up during the reconstruction process are also given. From theoretical analysis and experimental results, our method is shown to produce an accurate object geometric model, not just an approximate model. Furthermore, it does not rely on a precise object-to-image space transformation. Moreover, the method is fast and robust, since the basic geometric entity it derives is a face instead of a point.
MEASURING 3-D LOCATION AND SHAPE-PARAMETERS OF CYLINDERS BY A SPATIAL ENCODING TECHNIQUES
(1994-10-01)
CHEN, Z
;
CHIA, TL
;
HO, SY
;
交大名義發表
;
資訊科學與工程研究所
We are concerned with the problem of estimation of true cylinder radius, height, location, and orientation. We present mathematical models for measuring these location and shape parameters using a spatial encoding technique. A crucial step in the proposed method is to convert the estimation problem with complex curved stripe patterns to an equivalent, but simpler, estimation problem with line stripe patterns. The notions of silhouette edges and virtual plane are introduced. Various projective geometry techniques are applied to derive the cylinder location and shape parameters. The actual experiment apparatus is set up to employ the developed mathematical models to measure the cylinder geometric parameters. Description of the image-processing tasks for extracting perceived curved stripes and stripe endpoints is given. Sensitivity analysis and proper measures are taken to consider the effects of the uncertainties of stripe endpoints and silhouette edge locations on the measurement. Experiments have confirmed that our measurement method yields quite good results for different cylinders under various measurement environment conditions.
POLYHEDRAL FACE RECONSTRUCTION AND MODELING FROM A SINGLE IMAGE WITH STRUCTURED LIGHT
(1993-05-01)
CHEN, Z
;
HO, SY
;
TSENG, DC
;
資訊工程學系
;
Department of Computer Science
The determination of the 3-D geometric model of visible polyhedral faces from a single view is addressed. It applies a grid coding technique to derive the normal vector of each visible polyhedral face and the depth parameter of the race equation based on the given dimensions of the grid on the code plate. It is shown that the method is correspondenceless. Furthermore, it also determines the 3-D face model for all visible polyhedral faces including the occlusion relations between faces. Two edge reconstruction methods are given and their goodness are compared. For the determination of the final object model, three different integration methods are given. Some experiments are reported to illustrate the method.