Improving Passive 3D Model Reconstruction using Image Enhancement

Recently, 3D model reconstruction becomes a reality. That is because of using many algorithms that are implemented as automated tools and the ability to process dozens of images which might be calibrated or uncalibrated. When dealing with uncalibrated images, more processing time is usually required to perform "self-calibration". When using uncalibrated images, the images are usually not in perfect quality due to the uncalibrated nature of the capturing process. For this reason, this paper aims to study the impact of using image enhancement techniques on the quality of the building process of 3D models. This work mainly focus on the contrast of images. After applying the preprocessing procedures, the image sets where used to build the 3D models using three software kits (COLMAP, VSFM and Agisoft Photoscan). The results show that applying histogram equalization enhances the quality of the generated 3D models.

[1]  J. G. Lahoz,et al.  LASER SCANNING OR IMAGE-BASED MODELING? A COMPARATIVE THROUGH THE MODELIZATION OF SAN NICOLAS CHURCH , 2005 .

[2]  Alexandru Tupan,et al.  Triangulation , 1997, Comput. Vis. Image Underst..

[3]  Katrin Eismann,et al.  Adobe Photoshop Restoration & Retouching (3rd Edition) (Voices That Matter) , 2005 .

[4]  Changchang Wu,et al.  Towards Linear-Time Incremental Structure from Motion , 2013, 2013 International Conference on 3D Vision.

[5]  Jan-Michael Frahm,et al.  Structure-from-Motion Revisited , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Zhiguo Jiang,et al.  3D Reconstruction of Space Objects from Multi-Views by a Visible Sensor , 2017, Sensors.

[7]  Maarten Vergauwen,et al.  3D Reconstruction from Multiple Images: Part 1 - Principles , 2009, Found. Trends Comput. Graph. Vis..

[8]  Jan-Michael Frahm,et al.  A Vote-and-Verify Strategy for Fast Spatial Verification in Image Retrieval , 2016, ACCV.

[9]  Aziz Makandar,et al.  Computation Pre-processing Techniques for Image Restoration , 2015 .

[10]  N. A. Zakaria,et al.  3D Model Reconstruction from Multi-views of 2D Images using Radon Transform , 2015 .

[11]  Jose Luis Lisani,et al.  Simplest Color Balance , 2011, Image Process. Line.

[12]  Jan-Michael Frahm,et al.  Pixelwise View Selection for Unstructured Multi-View Stereo , 2016, ECCV.

[13]  Steven M. Seitz,et al.  Photo tourism: exploring photo collections in 3D , 2006, ACM Trans. Graph..

[14]  Y. Hung,et al.  Three-dimensional curve reconstruction from multiple images , 2012 .

[15]  Fabrizio Ivan Apollonio,et al.  ADVANCES IN IMAGE PRE-PROCESSING TO IMPROVE AUTOMATED 3D RECONSTRUCTION , 2015 .

[16]  Andrew W. Fitzgibbon,et al.  Bundle Adjustment - A Modern Synthesis , 1999, Workshop on Vision Algorithms.

[17]  Yizhou Yu,et al.  3D Human Model Reconstruction from Sparse Uncalibrated Views , 2016, IEEE Computer Graphics and Applications.