A smart color sensor based on a multi-Fuzzy Inference strategy

Color detection task is of main importance in several applications. Available devices suffer for high cost or very specific application area. In this paper, a novel color detector is presented. The device is based on a RGB sensor and a smart signal processing algorithm implemented on a microcontroller architecture. In order to perform a suitable and reliable color identification a fuzzy paradigm has been used which dominates the task complexity. A crisp rule set has been also constructed to implement both the model selection and the fault recover tasks. The device performance has been investigated in terms of failure probability in recognizing color samples and results obtained demonstrated the effectiveness and the reliability of the proposed architecture.