3D reconstruction of apple fruits using consumer-grade RGB-depth sensor

Three-dimensional reconstruction has great potential to improve not only the post-harvest quality control but also the breeding efficiency in horticulture. The depth information of the consumer-grade RGB-depth sensor was unreliable compared to that obtained from industrial sensors. To cope with this...

Full description

Saved in:
Bibliographic Details
Published in:Engineering in Agriculture, Environment and Food Environment and Food, 2018, Vol.11(4), pp.159-168
Main Authors: Yamamoto, Satoshi, Karkee, Manoj, Kobayashi, Yuichi, Nakayama, Natsuki, Tsubota, Shogo, Thi Thanh, Loan Nguyen, Konya, Tomoko
Format: Article
Language:English
Subjects:
3D model
Apple
Kinect
Point cloud
Size estimation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Three-dimensional reconstruction has great potential to improve not only the post-harvest quality control but also the breeding efficiency in horticulture. The depth information of the consumer-grade RGB-depth sensor was unreliable compared to that obtained from industrial sensors. To cope with this disadvantage, the generated point cloud was corrected within a region of interest of the target fruit, which was extracted from the color image of the sensor. Evaluating more than a hundred apple fruits, the root-mean-square error of the volume and the largest diameter were less than 6 cm3 and 1 mm, respectively. Reconstruction of various kinds of fruits and vegetables were demonstrated. The proposed method can be applied to accelerate the quantification of three-dimensional features of agricultural products.
ISSN:1881-8366
1881-8366
DOI:10.1016/j.eaef.2018.02.005