I. INTRODUCTION
For its wide field of applications, surfaces reconstruction is one of the most important topics in computer vision. It is usually related to structure from motion, stereovision, or pose determination, which results in applications including object modeling, mobile robot localization, environments building. The environment may be controlled, like in most of industrial visual inspection cells, or it may be complex due to illumination changes or motions like it is with outdoor environments or inside the human body for medical applications. Actually, the use of imagery in this field remains limited because of the nature, generally 2-D, of the information provided by the most of acquisition devices. Thus, a device providing, in real time, an accurate 3-D information of a very close environment would be helpful to achieve robotized surgical acts.