REFERENCING OF IMAGES TO LASER SCANNER DATA USING LINEAR FEATURES EXTRACTED FROM DIGITAL IMAGES AND RANGE IMAGES

摘要

This paper deals with the referencing of single images taken independently from the laser scanner (camera not mounted on the scanner) to a laser scanner point cloud of the same scene. The single image orientation is a standard task in photogrammatric applications and can be performed by a classical photo resection. Thereby, the determination of the exterior orientation is realised indirectly by measuring imaged object points in the photo. In architectural applications, linear features can often be extracted easier than points. Hence, the image orientation in this paper is based on linear features instead of point features. To decrease the effort of user interaction for feature measurement, this paper will present a method for line extraction from image data, which are eligible for image orientation. The first step of line extraction is the edge detection using a Canny edge detector. Afterwards, the resulting 1 pixel wide edges are vectorised in parts of straight lines by a customised Douglas-Peucker algorithm. The presented method is applied to extract 2D lines from digital image data and 3D lines from laser scanner data. At this, the basis for the 3D line extraction is a range image and an intensity image generated from the point cloud. The line extraction is applied to an image of a building facade, which subsequently is oriented with different levels of user interaction involving the extracted lines. To validate the results, the image orientation additionally was obtained with existing control points.