A sensor has been designed and tested in a pilot mill to characterize granular products in the food industry. It consists of: a mechanical system which takes a representative part of the product a CCD camera to capture images, a software package for image analysis and data processing. The method consists of comparing a sample with a predetermined ＇quality class＇. The decision system is built on example learning: real cases fed into the system allow its configuration. Three qualify classes have been defined, they correspond to the rolls gap (0.30, 0.40, 0.50 mm) of the first break rolls of a semolina pilot mill. In these conditions, the classification accuracy rate achieved by the system is higher than 80%.