Two-View Structure-from-Motion with Multiple Feature Detector Operators

Abstract

This paper presents a novel two-view Structure-from-Motion (SfM) algorithm with the application of multiple Feature Detector Operators (FDO). The key of this study is the implementation of multiple FDOs into a two-view SfM algorithm. The two-view SfM algorithm workflow can be divided into three general steps: feature detection and matching, pose estimation and point cloud (PCL) generation. The experimental results, the quantitative analyses and a comparison with existing algorithms demonstrate that the implementation of multiple FDOs can effectively improve the performance of a two-view SfM algorithm. Firstly, in the Oxford test dataset, the RMSE reaches on average 0.11 m (UBC), 0.36 m (bikes), 0.52 m (trees) and 0.37 m (Leuven). This proves that illumination changes, blurring and JPEG compression can be handled satisfactorily. Secondly, in the EPFL dataset, the number of features lost in the processes is 21% with a total PCL of 27,673 pt, and this is only minimally higher than ORB (20.91%) with a PCL of 10,266 pt. Finally, the verification process with a real-world unmanned aerial vehicle (UAV) shows that the point cloud is denser around the edges, the corners and the target, and the process speed is much faster than existing algorithms. Overall, the framework proposed in this study has been proven a viable alternative to a classical procedure, in terms of performance, efficiency and simplicity.

Enhancing Building Point Cloud Reconstruction from RGB UAV Data with Machine-Learning-Based Image Translation

“…However, this process decreases the number of features significantly [7]. Therefore, multiple feature detector operators are applied to increase the number of features [3]. Vegetation features often decrease the feature matching performance, especially when dealing with remotely sensed images, because of their dynamic appearances and characteristics in different images, even with only insubstantial imaging parameters.…”

“…Among the available GAN algorithms, the conditional GAN called Pix2Pix [2] can generate a corresponding image output based on input image information. In addition, the modified two-view Structure from Motion (SfM) algorithm can generate a 3D point cloud from two 2D images [3,4]. The process of reconstructing a 3D model with SfM involves taking images from different viewpoints (camera motion).…”