Abstract
In this paper, we put forward a new method for surface reconstruction from image-based point clouds. In particular, we introduce a new visibility model for each line of sight to preserve scene details without decreasing the noise filtering ability. To make the proposed method suitable for point clouds with heavy noise, we introduce a new likelihood energy term to the total energy of the binary labeling problem of Delaunay tetrahedra, and we give its s-t graph implementation. Besides, we further improve the performance of the proposed method with the dense visibility technique, which helps to keep the object edge sharp. The experimental result shows that the proposed method rivalled the state-of-the-art methods in terms of accuracy and completeness, and performed better with reference to detail preservation.
Highlights
Image-based scene reconstruction is a fundamental problem in Computer Vision
When it comes to large scale scenes with multi-scale objects, current reconstruction methods have some problems with the completeness and accuracy, especially when concerning scene details [3]
We present a new surface reconstruction method
Summary
Image-based scene reconstruction is a fundamental problem in Computer Vision. It has many practical applications in fields such as entertainment industry, robotics, cultural heritage digitalization and geographic systems. As far as small objects under controlled conditions are concerned, the performance of current scene reconstruction methods could achieve results comparable to those generated by laser scans or structured-light based methods [1,2]. When it comes to large scale scenes with multi-scale objects, current reconstruction methods have some problems with the completeness and accuracy, especially when concerning scene details [3]. To preserve scene details without decreasing the noise filtering ability, we propose a new visibility model with error tolerance and adaptive end weights. Experimental results show that the proposed method rivals the state-of-the-art methods in terms of accuracy and completeness, and performs better with reference to detail preservation
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have