The shape measurement and geometric modeling are strongly connected. And with the development of reverse engineering, reconstruction CAD models from real parts has become an important geometric modeling method. Human and animal can judge the surrounding 3-D geometry relationship accurately according to the 2-D image on the retina, which implies that creating 3-D shape from its 2-D images under common light conditions is feasible. And this method differs from the traditional methods in that it need to prove up the correlation between the 3-D shape of object and the visual information such as contour, luminance and color etc. contained in the intensity image, and to realize the conversion algorithm for visual information and geometry dimension.In this project, research has been made mainly on two aspects: in the first aspect, the shape measurement based on intensity images under various conditions and geometric modeling methods have been researched; and in the second aspect, combined with engineering field, the systems to realize these methods have been researched.The contributions of this project in the first aspect are followed:1. Introduce a new measurement method combinng shape from shading (SFS) and point correspondences.2. Introduce a new way to solve SFS problem using B-spline variation algorithm.3. Introduce a new visual method for 3-D reconstruction of the textureless free-form urface.Intensive foundational studies have also been made on the mathematics models of aboving methods.And in the application aspect:1. Develop a 3D scanner system using a hand-held laser source, which combines the shape data collected from laser scan and camera and simplifies the measurement process. The precision of 3-D reconstruction surface is higher.2. For cars and ceramic products etc., develop a 3-D reconstruction system from 2-D image obtained by a digital video.This project is about to deepen in the theory research and application generalization on the basis of above results, to introduce numerical manifold method to improve the algorithm's efficiency, and combining with medical filed to cooperate with relevant hospitals to develop applicable medical system used for dental treatment etc.
复杂形体的几何造型与形式设计是困扰CAD界的一大难题。本项目研究一般光照条件下物体耐枷裥蛄卸晕锾寮负翁卣骷靶问缴杓铺卣鞯南拗疲捎弥鸫畏指睢⒂纱旨熬牟呗裕岢霾⑹迪忠恢中碌哪嫦蛟煨头椒ǎ凰唤瞿芑指慈堤宓募负涡畔ⅲ鼓芑指次锾灞砻娴纳省⑼及傅刃问缴杓菩畔ⅲ┐驝AD用互提供一种全面、快速的复杂形体设计手段。.
{{i.achievement_title}}
数据更新时间:2023-05-31
硬件木马:关键问题研究进展及新动向
内点最大化与冗余点控制的小型无人机遥感图像配准
基于细粒度词表示的命名实体识别研究
滚动直线导轨副静刚度试验装置设计
基于混合优化方法的大口径主镜设计
微光照条件下CMOS图像传感器设计研究
微分和形态学算子混合的图像特征检测理论与方法
藏文实体语义关系抽取理论与方法研究
由图像序列测量三维物体的方法