A fuzzy sensor for color matching vision system

This article presents the development of a vision system according to a fuzzy sensor concept with the aim to improve the wooden boards color matching. The main difficulty consists in the recognition of gradual color in an industrial context. This problem implies the use of a flexible and robust vision system. It must be focused on the techniques that are able to work with the imprecision of physical measurements and the subjective human definitions of colors. Indeed, the color classes are not often well separated and the different users can have different perceptions of the colors. So, the vision system must be easy to tune. To achieve such a system, the recognition step is done with a fuzzy reasoning classifier that is well-adapted for this application case. The proposed method is quite efficient with small learning data sets because of the generalisation capacity of both feature selection and recognition steps. This technique allows to keep the interpretability of the model. The obtained results are compared with more classical methods and show the efficiency of the proposed fuzzy sensor, especially for merging outputs of several sensors used together in multi-face application case.