Abstract
This paper describes an oil identifi cation system that was developed to verify food safety and measures.
We studied the absorption property of oils by spectrometer. The principle component analysis of the
absorption spectrum of each oil identifi ed it if we used fi ve adjacent wavelengths at 1750, 1727, 1713,
1685 and 1657 nm. Our experimental prototype system consisted of fi ve laser diodes (LD) as a light
source, one photo diode as a receiver, and a 1 mm-length absorption cell. Each of the fi ve LDs operated
one of the fi ve wavelengths in turn. Oil identifi cation, which distinguished between inedible and edible
oil, was clearly shown on a two-dimension map. Our results confi rmed the feasibility of a LD-based
sensor-type identifi cation system that has many advantages such as compactness, low power and easy to
use.