Help us make Indico better by taking this survey! Aidez-nous à améliorer Indico en répondant à ce sondage !

May 24 – 25, 2021
Faculty of Science, Prince of Songkla University
Asia/Bangkok timezone

Towards Accurate Non-Contact Moisture Inspection using THz Imaging and Thickness Information

Not scheduled
15m
Faculty of Science, Prince of Songkla University

Faculty of Science, Prince of Songkla University

Oral Instrumentation, Metrology and Standards

Speaker

Mr Napat Cota (National Electronics and Computer Technology Center)

Description

This study proposes an approach for non-contact moisture inspection in dried food products, which is crucial to maintain optimal quality and shelf-life, using terahertz signal intensity and thickness of a test sample. To achieve this, a sample-specific calibration curve needs to be determined first. The HAITAI crackers were chosen in this work for demonstration purposes. Four different stacks of crackers were moisturized by spraying water on a tissue paper and then covering it on the sample. Actual moisture percentages were also determined immediately after data collection by a gravimetric method of the moisture analyzer, whose results served as ground-truth measurements. The moisture of the samples varied between 3 and 40%. The terahertz source emitted wave continuously at 0.1 THz while the detector was a 1D-array camera having 256 pixels and receiving waves in 0.05 – 0.7 THz frequency range. Transmitted terahertz signal of each sample, placed on a conveyor belt system between a terahertz source and detector, was measured 5 times, while the sample thickness was measured by a vernier. The 3D RGB camera was also used for measuring thickness of sample as well as obtaining RGB image. All dataset of known-thickness, and actual moisture measurements were used to calculate the coefficients necessary to define the signal absorption equation according to Beer-Lambert’s Law. Once a calibration curve was obtained, it was used to estimate the moisture percentages in samples with different thicknesses. The mean absolute error (MAE) of moisture is found to be less than 11%, when the sample thickness was taken into account for constituting the calibration curve. Therefore, the utilizing of 3D RGB camera for color-terahertz image registration and automatic thickness determination is promising for an accurate non-contact moisture inspection. This approach can be also integrated into production line to improve quality control in the dried food industry without interrupting existing processes.

Keywords: terahertz moisture imaging, calibration curve, correction

Primary authors

Mr Napat Cota (National Electronics and Computer Technology Center) Dr Rungroj Jintamethasawat (National Electronics and Computer Technology Center) Dr Kiattiwut Prasertsuk (National Electronics and Computer Technology Center) Mr Patharakorn Rattanawan (National Electronics and Computer Technology Center) Mr Natcha Cota (National Electronics and Computer Technology Center) Mr Piyaphat Phukphan (National Electronics and Computer Technology Center) Ms JIA YI CHIA (National Electronics and Computer Technology Center) Mr Woraprach Kusolthossakul (National Electronics and Computer Technology Center) Ms Panichard Poomvised (King Mongkut's University of Technology North Bangkok) Dr Thawatchart Chulapakorn (National Electronics and Computer Technology Center)

Presentation materials

Peer reviewing

Paper