REGISTRATION OF TERRESTRIAL LASER SCANS VIA IMAGE BASED FEATURES

The complexity of natural scenes and the amount of information acquired by terrestrial laser scanners turns the registration among scans into a complex problem. This problem becomes even more complex when considering the relatively low angular resolution of terrestrial scanner compared to images, the monotonicity of manmade surfaces that makes the detection of corresponding objects difficult, and the lack of structure of vegetated objects that makes the detection of meaningful features difficult. Since most modern scanners are accompanied with consumer cameras of relatively high quality, it stands to reason making use of the image content for the registration process. Such alternative will benefit from the large body of image based registration work that has been carried out for several decades and therefore has the potential of providing an alternative and simple approach for the registration of pairs and multiple scans simultaneously. In this paper, we study the registration of terrestrial scans via image-based information. For this purpose, we propose an efficient autonomous model that supports the registration of multiple scans. Following the presentation of the model, we analyze its application to outdoor, complex scenes, ones that are common to find in actual laser scanning projects. 1