选自《3D modeling by means of Videogrammetry and Laser Scanners for reverse engineering》即《三维建模的优势和激光扫描仪用于逆向工程》
1. Introduction
The automatic generation of virtual 3D models is of major interest in many scientific areas and for many application domains. Several tools and techniques have been developed due the need to know accurately the geometry of an element. Fundamentally it could be differentiated between those based on images and those based on distances.
Currently, laser scanner is the most used tool in this type of works. It offers exceptional performance as to the amount of data provided in a short time in engineering works [1] or reverse engineering [2]. However it suffers from certain problems in data gatering, in the subsequent data processing and in terms of accuracy. For data gatering it is necessary to make multiple data captures around the element to capture all geometry and to be able to identify the object smoothly [3], which implies a high capture time. In regard to processing, the first problem is the filtration of clouds (most of the registered information does not correspond to the element of interest) and subsequently the clouds have to be joined [4]. Both processes require high processing time. And in terms of accuracy, geometric and dimensional tolerance control is necessary [5] as well as digitizing errors [6]. Finally we must add problems such as weight and size of the equipment, and mainly its price (which it makes hard accessible for many applications).
For image processing, stereoscopic photogrammetry with calibrated cameras [7] has been conventionally the more used technique [8]. Also, Photomodelling techniques based on automatic stereoscopic processes suffers from similar problems than laser scanner. Data gatering requires multiple setups and image processing requires high skills and modeling time [9, 10], although it is much cheaper than the laser scanner [11,12]. To exploit the potential of laser scanner and solve the problems arising from the lack of image is increasingly common to find works that join both techniques [13,14]. However, we still have problems of cost and processing ones are increased.
Videogrammetry has been used in several fields for years [15-19]. Its potential results great since it has shown the advantages of the two previous systems and none of its drawbacks. Therefore, videogrammetry is increasingly applied to precise measurements [20-24]. It is a faster technique in data gatering compared to the laser scanner or conventional photogrammetry, it offers a huge amount of points and the processing time is much lower.
This work shows the generation of 3D models by using different laser scanners and videogrammetry. The data gatering and the data processing results much faster with videogrammetry than with laser scanners; it results completely automatic and it does not depend on the type of material. The method for generating 3D models by videogrammetry improves usual techniques used in photomodelling. The process developed makes no approximations sample points, more or less dense. The method developed uses all object points on every edge of every frame generating an accurate 3D model. Results shows that the videogrammetry is a real alternative with high accuracy. It provides precision of 1-2 mm that is better than the architecture laser scanner (whose accuracy is of 3-6mm) and almost as precise than portable 3D scanner (whose accuracy is of 1mm). Finally 3D visualization through anaglyphs is possible in all desired positions, while laser systems can not provide this kind of information.
Both scanner lasers are a real option for 3D modeling for reverse engineering but each one of them can not be applied in all cases. Thus, architectural scanner laser will be adecuate in cases of big objects (for example in car or naval industry where portable 3D scanner can not be applied due to the big amount of scans requiered) while portable 3D scanner will be adecuate for small ones (in which architectural scanner laser has a lower accuracy). On the other hand, videogrammetry can be applied in all cases with many advantages.
2. Automatic generation of 3D models by videogrammetry
2.1. System design and calibration
以上是毕业论文文献综述，课题毕业论文、任务书、外文翻译、程序设计、图纸设计等资料可联系客服协助查找。
