How to Decide From the First View Where to Look Next

Thumbnail Image
Penn collection
Technical Reports (CIS)
General Robotics, Automation, Sensing and Perception Laboratory
Degree type
Grant number
Copyright date
Related resources
Maver, Jasna
Bajcsy, Ruzena

The task of constructing a volumetric description of a scene from a single image is an underdetermined problem, whether it is a range or an intensity image. To resolve the ambiguities that are caused by occlusions in images, we need to take sensor measurements from several different views. We have limited ourselves to the range images obtained by a laser scanning system. It is an active binocular system which can encounter two types of occlusions. An occlusion arises either when the reflected laser light does not reach the camera or when the directed laser light does not reach the scene surface. The task of 3-D data acquisition is divided into two subproblems: to see what is illuminated and to properly direct the illuminating plane to illuminate the entire scene. The first kind of occlusions are easily detected and can be used in designing an efficient algorithm. We develop a strategy to determine the sequence of different views using the information in a narrow zone around the occluded regions. Occluded regions are approximated by polygons. Based on the height information of the border of the occluded regions and geometry of the edges of the polygonal approximation, the next views in the same scanning plane are determined. From the acquired information in the first scanning plane the directions of the next scanning planes are computed.

Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
Volume number
Issue number
Publisher DOI
Journal Issue
University of Pennsylvania Department of Computer and Information Science Technical Report No. MS-CIS-90-39.
Recommended citation