Volume Estimation Of Curved Surfaces Using Structured Lighting And Their Similarity Measure
Loading...
Date
2006-07
Authors
Abd Elmagid Elnaggar, Ahmed Mohamed
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis presents the investigation of a non-contact method called structured
lighting to determine volumes of smooth surfaces. This method uses a camera that
captures a deformed pattern of light projected upon a surface. The system was verified
using objects having a smooth surface. Firstly, a simulation study was carried out to
determine a suitable method to determine the volume. A suitable method was found by
integration of a fitted curve across coordinate points and multiplied by its depth along xdirection.
The effect of different fringe pitches on volume calculation was studied.
Secondly, an experiment was carried out to determine the volume of six different
spherical caps. A projector was developed to project a collimated fringe pattern. A
digital camera was calibrated to convert the pixel values into millimetres. The error in
volume of 26 measurements was found to be less than 8.52%. A segmentation method
was developed to extract the surface from the background. The segmentation method
consists of three steps: (i) a triangular mesh is generated for the known surface points,
(ii) the area of each triangle is calculated and (iii) the standard deviation of triangle
areas that surround each surface point is calculated. All surface points that have similar
standard deviation value are segmented into one patch. Finally, the shape similarity
among the measured 3-0 shape is calculated. Each 3-0 model was brought into its
canonical position to ensure its rotation and translation invariance. At this position, a
set of features is extracted from the 3-0 model and the similarity between all models
are measured. This set of features is the core length of the bounding box. The new
features show better accuracy (43.75%) in sir,;!arity measure when applied to shapes
with smaller height and diameter as compared to the slightly larger shapes. The work
of this thesis can be used to develop a non-invasive system to measure skin tumours.
Description
Keywords
Mechanical Engineering , Similarity Measure