Depth Optimization for Accurate 3D Reconstruction from Light Field Images

被引:0
作者
Wang, Xuechun [1 ]
Chao, Wentao [1 ]
Duan, Fuqing [1 ]
机构
[1] Beijing Normal Univ, Sch Artificial Intelligence, Beijing 100875, Peoples R China
来源
PATTERN RECOGNITION AND COMPUTER VISION, PRCV 2023, PT II | 2024年 / 14426卷
关键词
Light field; Depth map; Optimization; 3D reconstruction;
D O I
10.1007/978-981-99-8432-9_7
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Because the light field camera can capture both the position and direction of light simultaneously, it enables us to estimate the depth map from a single light field image and subsequently obtain the 3D point cloud structure. However, the reconstruction results based on light field depth estimation often contain holes and noisy points, which hampers the clarity of the reconstructed 3D object structure. In this paper, we propose a depth optimization algorithm to achieve a more accurate depth map. We introduce a depth confidence metric based on the photo consistency of the refocused angular sampling image. By utilizing this confidence metric, we detect the outlier points in the depth map and generate an outlier mask map. Finally, we optimize the depth map using the proposed energy function. Experimental results demonstrate the superiority of our method compared to other algorithms, particularly in addressing issues related to holes, boundaries, and noise.
引用
收藏
页码:79 / 90
页数:12
相关论文
共 50 条
[41]   3D box method for 3D reconstruction of an object from multi-images [J].
Alam, J ;
Hama, H .
VISION GEOMETRY IX, 2000, 4117 :81-90
[42]   Progress and Prospect of 3D Reconstruction Based on Light Field Cameras [J].
Liu Y.-X. ;
Zhang L. ;
Ai H.-B. ;
Xu B. ;
Sun Y.-S. ;
Fan Z.-L. .
Tien Tzu Hsueh Pao/Acta Electronica Sinica, 2022, 50 (07) :1774-1792
[43]   Dense 3D Object Reconstruction from a Single Depth View [J].
Yang, Bo ;
Rosa, Stefano ;
Markham, Andrew ;
Trigoni, Niki ;
Wen, Hongkai .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2019, 41 (12) :2820-2834
[44]   Fast 3D Reconstruction of UAV Images Based on Neural Radiance Field [J].
Jiang, Cancheng ;
Shao, Hua .
APPLIED SCIENCES-BASEL, 2023, 13 (18)
[45]   3D depth sensing of active structured light field based on EPI [J].
Lin YingSi ;
Wang Meng ;
Wang Ziwei ;
Tang Qijian ;
Liu Xiaoli .
OPTICAL SENSORS 2021, 2021, 11772
[46]   3D Kidney Reconstruction from 2D Ultrasound Images [J].
Teresa Alvarez-Gutierrez, Mariana ;
Rodrigo Mejia-Rodriguez, Aldo ;
Alejandro Cruz-Guerrero, Ines ;
Roman Arce-Santana, Edgar .
VIII LATIN AMERICAN CONFERENCE ON BIOMEDICAL ENGINEERING AND XLII NATIONAL CONFERENCE ON BIOMEDICAL ENGINEERING, 2020, 75 :393-400
[47]   Fast and Accurate 3D Reconstruction of Plants Using MVSNet and Multi-View Images [J].
Chen, Zhen ;
Lv, Hui ;
Lou, Lu ;
Doonan, John H. .
ADVANCES IN COMPUTATIONAL INTELLIGENCE SYSTEMS, 2022, 1409 :390-399
[48]   Light field camera and integral 3D display 3D image reconstruction based on lightfield data [J].
Iwane, Toru .
2014 13TH WORKSHOP ON INFORMATION OPTICS (WIO), 2014,
[49]   Light-field angular super-resolution reconstruction via fusing 3D epipolar plane images [J].
Chen C. ;
Fan X. ;
Hu X. ;
Yu H. .
Guangxue Jingmi Gongcheng/Optics and Precision Engineering, 2023, 31 (21) :3167-3177
[50]   Accurate and robust 3D reconstruction of wind turbine blade leading edges from high-resolution images [J].
Sterckx, Jonathan ;
Vlaminck, Michiel ;
De Bauw, Koenraad ;
Luong, Hiep .
AUTOMATION IN CONSTRUCTION, 2025, 175