The food industry requires a high level of quality for huge product volumes. But individual product quality inspections are not always feasible, due to volume, processing speed and technology limitations. Many of the current quality control methods are destructive, indirect, sample based and time consuming. The food processing industry would benefit by a non-destructive inspection and sorting method. This led to cosine researching whether various properties of tomatoes could be discerned using their spectral features. Another major challenge was to use COTS (Commercial Off The Shelf) components instead of lab components to guarantee a robust and cost effective solution for companies.
We measured three different kinds of tomatoes next to a white reference: Beef tomato, Small cherry and Big cherry, using a hyperspectral camera in the 380-800 nm spectral range. We used the Principal Component Analysis (PCA) method to identify differences between the spectra of the different kinds of tomatoes. PCA is a mathematical procedure that uses an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables called principal components. This transformation is defined in such a way that the first principal component has the largest possible variance and each following component has the highest variance possible under the constraint that it is orthogonal to the preceding components. We used a python code to perform PCA on the spectral data.
We were able to distinguish tomatoes by their cultivar by looking at the spectrum. This was achieved by using the specular reflection filter and the analysis technique PCA. We also drafted a first concept on how to measure tomato ripeness.
condi food was founded by cosine to develop a commercial product based on these research insights. Since then, these insights have been further improved and inline equipment for the food processing industry is now available.