Special Issue on Registration and Fusion of Range Images

✍ Scribed by Marcos Rodrigues; Robert Fisher; Yonghuai Liu

• Publisher: Elsevier Science
• Year: 2002
• Tongue: English
• Weight: 57 KB
• Volume: 87
• Category: Article
• ISSN: 1077-3142
• DOI: 10.1006/cviu.2002.0978

✦ Synopsis

This special issue contains nine high-quality papers representative of the state-of-the-art technologies used to acquire and process range image data. Following the call for papers, 23 manuscripts were received which were reviewed by a pool of 37 expert referees. We are thankful to these anonymous referees for their invaluable work which, as ever, is essential to maintain the high standards of CVIU. All selected papers carry an element of novelty, and we hope that this issue will be useful to theoreticians and practitioners alike.

Directly and indirectly, the papers deal with a central problem in working with 3D data sets, namely the fusion of single 2.5D range images into full 3D data sets describing all surfaces of an object or environment. In Section 2 we highlight what we perceive as the main issues and relevant approaches to this central task.

The issue opens with a paper by Vanden Wyngaerd and Van Gool on automatic prealignment of surfaces, a task that is usually performed manually by current approaches. The method assumes Monge patch data from which subsequent good registration can be achieved. The paper by Krsek, Pajdla, and Hlavac also deals with automatic registration based on information from features found in the data which are determined from differential geometry. Robertson and Fisher also tackle automatic registration by using an evolutionary approach to estimate transformation parameters making it a pose search method as 1