Application of neuro-fuzzy inference system on wood identification

Abstract

Searching for systems with intelligent, flexible, and self-adjusting solutions on imaging, which could provide the contraction of the human operators' presence, a range of techniques is found. Each one of them can control the process through the assistance of autonomous systems, either software or hardware. Therefore, modeling by traditional computational techniques is quite difficult, considering the complexity and non-linearity of image systems. Compared to traditional models, the approach with Artificial Neural Networks (ANN) behaves well as noise elimination and non-linear data treatment. Consequently, the challenges in the wood industry justify the use of ANN as a tool for process improvement and, therefore, add value to the final product. Additionally, the Artificial Intelligence techniques, such as Neuro-Fuzzy Networks (NFN), have shown efficient, since they combine the ability to learn from examples and to generalize the learned information from the ANNs with the capacity of Fuzzy Logic, in order to transform linguistic variables in rules. Then, ANFIS plays active roles in an effort to reach a specific goal. © (2014) Trans Tech Publications, Switzerland.