Development of Vision Based Surface Measurement System for Stone Implements with Specular Reflection

本研究开发了一套基于视觉的镜面反射石器表面测量系统。在考古遗址出土的包括黑曜石在内的凿石工具，在剥落的表面上具有轮廓、环和裂缝等特征。为了实现对石质文物的分析和计算机自动配图，必须准确地描绘出石质文物的特征。我们首先提出了一种使用空间编码平面光源测量客观镜面法向矢量的方法。在此基础上，改进了三维距离像的获取方法，提出了一种空间编码模式观测误差的估计和校正方法。测量系统的原型已经开发，以显示我们的算法的有效性。

This research developed a Vision-based surface measurement system for stone implements with specular reflection. Chipped stone tools including obsidian, excavated at archeological sites, have features such as outline, rings, and fissures on the flaked surfaces. It is necessary to accurately delineate the features of stone artifacts in order to implement their analysis and computer-automated illustration and matching. We first proposed a method for measuring the normal vectors of objective specular surfaces, using a space coded flat light source. Then, we improved the method to obtain three-dimensional range images of surfaces, and also proposed a method for estimating and correcting observation errors of space coding patterns. A prototype of the measuring system has been developed in order to show the effectiveness of our algorithm.