Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning

In order to obtain the depth information of the target in the scene and realize three-dimensional (3D) reconstruction, in this paper, a target reconstruction method combining monocular focus stack image and deep neural network is proposed. This method makes full use of the advantages of light field...

Full description

Saved in:
Bibliographic Details
Main Authors: Yanzhu Hu, Yingjian Wang, Song Wang
Format: Article
Language:English
Published: IEEE 2020-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/9187884/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849430699664211968
author Yanzhu Hu
Yingjian Wang
Song Wang
author_facet Yanzhu Hu
Yingjian Wang
Song Wang
author_sort Yanzhu Hu
collection DOAJ
description In order to obtain the depth information of the target in the scene and realize three-dimensional (3D) reconstruction, in this paper, a target reconstruction method combining monocular focus stack image and deep neural network is proposed. This method makes full use of the advantages of light field imaging technology and can generate the all focus image. The method first collects multiple frames of continuous images at different focal lengths of the scene, using a divide and conquer algorithm strategy, uplink uses YOLO neural network to identify the target in 3D space and track the position information; the downlink reconstructs the four-dimensional (4D) light field data based on the focus stack image frequency domain back projection, and then uses light field imaging technology to invert the scene parallax; subsequently, achieve scene depth estimation and reconstruction of all focus image; finally, the uplink and downlink are merged to realize the reconstruction of the 3D point cloud of the space target. Experimental results on real scenes show the effectiveness of the proposed algorithm.
format Article
id doaj-art-e741d202f1144f33a56dde549206ae7e
institution Kabale University
issn 2169-3536
language English
publishDate 2020-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-e741d202f1144f33a56dde549206ae7e2025-08-20T03:27:52ZengIEEEIEEE Access2169-35362020-01-01816809916811010.1109/ACCESS.2020.30226309187884Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep LearningYanzhu Hu0https://orcid.org/0000-0002-0115-3143Yingjian Wang1https://orcid.org/0000-0002-3539-3110Song Wang2https://orcid.org/0000-0001-9589-4956College of Modern Post, Beijing University of Posts and Telecommunications, Beijing, ChinaCollege of Modern Post, Beijing University of Posts and Telecommunications, Beijing, ChinaCollege of Modern Post, Beijing University of Posts and Telecommunications, Beijing, ChinaIn order to obtain the depth information of the target in the scene and realize three-dimensional (3D) reconstruction, in this paper, a target reconstruction method combining monocular focus stack image and deep neural network is proposed. This method makes full use of the advantages of light field imaging technology and can generate the all focus image. The method first collects multiple frames of continuous images at different focal lengths of the scene, using a divide and conquer algorithm strategy, uplink uses YOLO neural network to identify the target in 3D space and track the position information; the downlink reconstructs the four-dimensional (4D) light field data based on the focus stack image frequency domain back projection, and then uses light field imaging technology to invert the scene parallax; subsequently, achieve scene depth estimation and reconstruction of all focus image; finally, the uplink and downlink are merged to realize the reconstruction of the 3D point cloud of the space target. Experimental results on real scenes show the effectiveness of the proposed algorithm.https://ieeexplore.ieee.org/document/9187884/Focus stack imagedeep learninglight field reconstructionall focus image3D reconstruction
spellingShingle Yanzhu Hu
Yingjian Wang
Song Wang
Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning
IEEE Access
Focus stack image
deep learning
light field reconstruction
all focus image
3D reconstruction
title Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning
title_full Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning
title_fullStr Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning
title_full_unstemmed Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning
title_short Scene Target 3D Point Cloud Reconstruction Technology Combining Monocular Focus Stack and Deep Learning
title_sort scene target 3d point cloud reconstruction technology combining monocular focus stack and deep learning
topic Focus stack image
deep learning
light field reconstruction
all focus image
3D reconstruction
url https://ieeexplore.ieee.org/document/9187884/
work_keys_str_mv AT yanzhuhu scenetarget3dpointcloudreconstructiontechnologycombiningmonocularfocusstackanddeeplearning
AT yingjianwang scenetarget3dpointcloudreconstructiontechnologycombiningmonocularfocusstackanddeeplearning
AT songwang scenetarget3dpointcloudreconstructiontechnologycombiningmonocularfocusstackanddeeplearning