Image Selection Method from Image Sequence to Improve Computational Efficiency of 3D Reconstruction: Analysis of Inter-image Displacement Based on Optical Flow for Evaluating 3D Reconstruction Performance
Toshihide Hanari, Kuniaki Kawabata, Keita Nakamura · 2022 IEEE/SICE International Symposium on System Integration (SII) · 2022
This paper describes the image selection method for an efficient three-dimensional (3D) reconstruction computation from an image sequence. Adequate images must be selected from the image sequence to improve the computational efficiency of the 3D reconstruction. Thus, we investigated a threshold based on the displacement among images obtained from a camera mounted on a remotely operated robot. The results confirmed that the proposed method can select adequate images for efficient 3D reconstruction by the threshold based on the optical flow from the image sequence. Therefore, the computational cost could be reduced by eliminating the duplicate and high-similarity images to perform the efficient 3D reconstruction.