Inline discrete tomography system: Application to agricultural product inspection

•It is proposed an inline X-ray Computed Tomography imaging setup.•Discrete Tomography is well suited for the proposed scanning setup.•The use of a conveyor belt with object rotation allows a full angular sampling.•The use of prior knowledge about the external object shape leads to better results. X...

Full description

Saved in:
Bibliographic Details
Published in:Computers and electronics in agriculture 2017-06, Vol.138, p.117-126
Main Authors: Alves Pereira, Luis F., Janssens, Eline, Cavalcanti, George D.C., Tsang, Ing Ren, Van Dael, Mattias, Verboven, Pieter, Nicolai, Bart, Sijbers, Jan
Format: Article
Language:English
Subjects:
Agricultural commodities
Agricultural engineering
Belt conveyors
Computed tomography
Computer simulation
Discrete tomography
Image acquisition
Image reconstruction
Inline scanning geometry
Inspection
Inspections
Studies
System effectiveness
Tomography
X ray sources
X-ray astronomy
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:•It is proposed an inline X-ray Computed Tomography imaging setup.•Discrete Tomography is well suited for the proposed scanning setup.•The use of a conveyor belt with object rotation allows a full angular sampling.•The use of prior knowledge about the external object shape leads to better results. X-ray Computed Tomography (CT) has been applied in agriculture engineering for quality and defect control in food products. However, conventional CT systems are neither cost effective nor flexible, making the deployment of such technology unfeasible for many industrial environments. In this work, we propose a simple and cost effective X-ray imaging setup that comprises a linear translation of the object in a conveyor belt with a fixed X-ray source and detector, with which a small number of X-ray projections can be acquired within a limited angular range. Due to the limitations of such geometry, conventional reconstruction techniques lead to misshapen images. Therefore, we apply a Discrete Tomography reconstruction technique that incorporates prior knowledge of the density of the object’s materials. Moreover, we further improve the reconstruction results with the following strategies: (i) an image acquisition involving object rotation during a linear translation in the conveyor-belt; and (ii) an image reconstruction incorporating prior knowledge of the object support (e.g., obtained from optic sensors). Experiments based on simulation as well as real data demonstrate substantial improvement of the reconstruction quality compared to conventional reconstruction methods.
ISSN:0168-1699
1872-7107
DOI:10.1016/j.compag.2017.04.010