Computer Vision Application on Controlling and Modeling the Food Drying Process: A Case Study for Radish (Raphanus Sativus)

The data (color, texture, shape, pattern) of food products obtained from photos taken with cameras have been used in all phases from production until they get through to end-user. In this view, the main object of this study is to analyze controlling the drying process with computer vision used in keeping and storing food and both manage the quality control phase and determine to finalize drying phase. Another object of the study is to show that the performance and quality of drying process and other parameters under different conditions (heat source, moisture, temperature etc.) can be observable by using image analyses. In accordance with these objects, food drying process with three different heat sources has been analyzed, food pictures have been taken with RGB cameras during drying process and the results have been modeled. According to the modeling results, the lowest error in training data and test data have been obtained with resistance drying (1.3493 MAE and 1.8446 RMSE) and infrared (6.3375 MAE and 7.0520 RMSE), successively. It has been seen that the resistance values have almost produced the best results in terms of R2 values. These error values staying between the acceptable limits while using the simple modeling can obtain more accurate results when they are used with improved modeling techniques by increasing the data number. This situation has shown that the drying process can be controlled with computer vision.