A Comparative Analysis of Speeded Up Robust Features (SURF) and Harris Algorithms in Point Cloud Generation

Abstract

The precise documentation and preservation of cultural heritage elements are of fundamental importance since they represent a social, historical and cultural improvement of societies throughout history. There have been different documentation techniques from past to present such as classical architectural drawing methods, classical surveying methods, terrestrial photogrammetry and aerial photogrammetry, etc. Photogrammetry emerges as a popular method with the developing technology. On the other hand, different evaluation techniques have been evolved to evaluate the collected photogrammetric data especially with the improvements in computer vision technologies. In this aspect, Structure from Motion (SfM) became a widely used approach for producing a 3D view from photogrammetric images. This paper presents the comparison of point clouds produced by Speeded Up Robust Features (SURF) and Harris Algorithms. The point cloud production was carried out in MATLAB environment. 20 images at total captured by Nikon Coolpix A10 digital camera were used for the point cloud generation. Therefore, the resulted points clouds were investigated comparatively.