Gelatin is a protein substance derived from collagen, a natural protein present in skin, bones, and animal tissue. Gelatin's ability to form strong, transparent gels and flexible films has made it a valuable commodity in food processing, pharmaceuticals, photography, and paper production. Gelatin recovered from bone is used primarily for pharmaceutical purposes. In its manufacturing process a number of steps are taken such as chopping the bones, degreasing them to reduce the fat content and dehydrating them in industrial dryers. Next, an acid/alkaline treatment follows by soaking the bones for approximately five days. This process removes most of the minerals and bacteria and facilitates the release of collagen. Finally, the pieces of bone are boiled in distilled water and the liquid that now contains gelatin can be drawn off.

To be able to predict the quality of the resulting gelatin a number of characteristics need to be known from the raw sample bones such as percentages of fat, protein, water and mineral content. Generally, a chemical analysis is performed on the degreased and dehydrates bones to extract this information. Ideally we would like to know the characteristics of the sample batch in order to be able to discard it before proceeding to the acid/alkaline treatment. However, current chemical analysis cannot be performed in a timely manner and the acid treatment is started before the bone characteristics and quality level are known. Moreover, due to efficiency constraints the analysis is only performed on a small subset of the batch. For a more efficient use of resources a faster and robust evaluation technique of these bone properties is desired. This would allow to screen wider amounts of bone samples while rapidly extracting the required information before the bones undergo any further time consuming process.

For this purpose we want to evaluate if hyper-spectral imaging is suitable to perform the required fast and accurate prediction of these parameters (water, fat, protein and mineral content) based on the spectral response of dehydrated bone samples.