Image analysis with the computer vision system and the consumer test in evaluating the appearance of Lucanian dry sausage

The object of the investigation was the Lucanian dry sausage appearance, meant as color and visible fat ratio. The study was carried out on dry sausages produced in 10 different salami factories and seasoned for 18 days on av- erage. We studied the effect of the raw material origin (5 producers used meat bought from the market and other 5 producers used meat from pigs bred in their farms) and of the salami factories or brands on meat color, fat color and visible fat ratio in dry sausages. The sausages slices were photographed and the images were analysed with the computer vision system to measure the changes in the colorimetric characteristics L*, a*, b*, hue and chroma and in the visible fat area ratio. The last parameter was assessed on the slice surface using image binarization. A consumer test was conducted to determine the relationship between the perception of visible fat on the sausage slice surface and acceptability and preference of this product. The consumers were asked to look carefully at the 6 sausages slices in a photo, minding the presence of fat, and to identify (a) the slices they considered unacceptable for consumption and (b) the slice they preferred. The results show that the color of the sausage lean part varies in relation to the raw material employed and to the producer or brand (P b 0.001). Besides, the sausage meat color is not uniform in some salami factories (P b 0.05–0.001). In all salami factories the sausages show a high unifor- mity in fat color. The visible fat ratio of the sausages slices is higher (P b 0.001) in the product from salami facto- ries without pig-breeding farm. The fat percentage is highly variable (P b 0.001) among the sausages of each salami factory. On the whole, the product the consumers consider acceptable and is inclined to eat has a low fat percentage (P b 0.001). Our consumers (about 70%) prefer slices which are leaner (P b 0.001). Women, in particular, show a higher preference for the leanest (P b 0.001).