A survey of cloud lighting and rendering techniques

Abstract

Scanning 3D objects has become a valuable asset to many applications. For larger objects such as buildings or bridges, a scanner is positioned at several locations and the scans are merged to one representation. Nowadays, such scanners provide, beside geometry, also color information. The different lighting conditions present when taking the various scans lead to severe luminance artifacts, where scans come together. We present an approach to remove such luminance inconsistencies during rendering. Our approach is based on image-space operations for both luminance correction and point-cloud rendering. It produces smooth-looking surface renderings at interactive rates without any preprocessing steps. The quality of our results is similar to the results obtained with an object-space luminance correction. In contrast to such an object-space technique the presented image-space approach allows for instantaneous rendering of scans, e.g. for immediate on-site checks of scanning quality.