A Vision System for Surface Homogeneity Analysis of Dough Based on the Grey Level Co‐occurrence Matrix (GLCM) for Optimum Kneading Time Prediction

This paper presents an investigation of a methodology for predicting the optimum time to decide dough readiness by using two-dimensional imaging and statistical texture analysis. Based on the master baker experience and the torque curve generated by the dough mixer, the system will be correlated to stop the kneading process with the minimum error on the optimum time or point of maximum dough development (peak consistency). This development takes advantage of the gray level co-occurrence matrix (GLCM) texture analysis for the development of an algorithm to emulate the visual inspection performed by the master bakers, where the homogeneity feature of dough is analyzed during the kneading time. An automatic graphical user interface software for offline and inline analysis was developed for the homogeneity analysis. The programmed algorithm was implemented in many experiments (online and offline) for testing, producing an average error of 33.9 s with respect to average optimal mixing time. Practical Applications The development of a dough analyzing method, on the basis of an optical surface evaluation, represents an entirely new way to look at the dough properties. The possibility of applying an online dough judgment method without any losses will be given. Simple upgrading of existing production lines can help to prevent erroneous processing and to increase the quality of the end product. Instantaneous state of the art for determining the optimum dough development or dough readiness is the visual and tactile evaluation of the technical staff and laboratory investigations. The expansion of the indirect method by a digital interface-based analytic approach provides a significant improvement in industrial bakery production and a potential direct method for dough analysis.