الفهرس | Only 14 pages are availabe for public view |
Abstract Laser scanners are already accepted as a tool for (3D) three dimensional modeling and analysis in the engineering, architecture, and construction fields. Technological advances have led to laser scanners capable of acquiring range measurements at rates of tens to hundreds of thousands of points per second, at distances of up to few hundred meters, and with uncertainties on the scale of millimeters to few centimeters. One of the important issues in laser scanning is registering the multiple scan positions. Despite all modern techniques in registering terrestrial laser scanner point clouds, registration of multiple scan positions is still a crucial issue, especially in large projects, Placing targets in fields Processes and detection of corresponding points in scans and consume effort, time and money. Three points are the minimum number of common points required for registration. In this work, a new Two Point Registration (TPR) algorithm is developed to register terrestrial laser scanner point clouds using only two common points, which reduces one-third of the registration work in both field and laboratory, by using Two Point Registration (TPR) algorithm no need to tilt the laser scanner, this reduce the cost of laser scanner manufacturing. The Two Point Registration algorithm (TPR), consists of three main stages which are: (i) Detecting possible corresponding points in the two scans. (ii) Finding the correct corresponding points at the two scans. (ii) Computing the registration parameters. An accuracy assessment study is then performed to compare the registration accuracy of the developed algorithm with the traditional algorithm using both the time-of-flight and the phase shift laser scanner.5 The results show that the accuracy of the two algorithms is approximately equal and within the considered point clouds accuracy in both cases. Also this thesis studied the different techniques of registration and comparison between then as follows: 1. Basic target free registration is not frequently used in practice because it requires accurate instrument installation over the control points. Also, any resulted error made by mistake would result in unusable collected data. 2. In Feature-based registration a large overlapping area is usually required which in turn increases the number of scans. This technique still fails to register scans at many times. 3. The target-based registration depends on the type of the used targets (natural/artificial). By using natural targets like corners or edges appeared in point clouds, much manual laboratory work is consumed to find the corresponding natural targets in the scans in addition to the lower accuracy obtained. Artificial targets on contrarily are automatically detected from scans which guarantee full automatic registration. However, planning for targets placing in the field and the setting-up time are the main obstacles of this technique. 4. the Iterative Closest Point ICP method can be applied to achieve the fine registration. |