An approach for estimating adulteration of virgin olive oil with soybean oil using image analysis

This paper describes the development of an image processing algorithm, which can estimate the amount of adulteration of olive oil with soybean oil from a captured photo. This algorithm is intended to be implemented into an application for modern smartphones, where the user can measure the quality of a sample of olive oil, only by taking photos from the sample. The determination of the adulteration percentage is done by separating the captured image into two regions: one that contains only the oil sample and another one, which contains the rest of the image. The colour difference between these two regions, for known adulteration percentages is used to determine the appropriate model that combines these quantities. Then, any other mixture of these oils, can be identified using the derived model and the methodology that is described in this paper.