A Robust SVM Color-Based Food Segmentation Algorithm for the Production Process of a Traditional Carasau Bread

In this paper, we address the problem of automatic image segmentation methods applied to partial automation production process a traditional Sardinian flatbread called pane Carasau for assuring quality control. The study focuses on one most critical activities obtaining an efficient degree automation: estimation size and shape bread sheets during phase, variations undergone by sheet depending some environmental variables. knowledge can thus be used create system capable predicting dough produced empower process. We implemented acquisition created machine learning algorithm, based support vector machines, measurements bread. Experiments demonstrated that method successfully achieve accurate images, ensuring dimensions extracted are representative coming from algorithm proved fast in estimating various scenarios occurred over year acquisitions. maximum error committed is equal 2.2% pixel worst scenario 1.2% elsewhere.