An intelligent approach based on adaptive neuro-fuzzy inference systems (ANFIS) for walnut sorting

Abstract

In the present paper, an efficient walnut recognition system was developed by combining acoustic emissions analysis, Principle Component Analysis (PCA) and Adaptive Neuro-Fuzzy Inference System (ANFIS) classifier. The system was tested later and classified walnuts into two classes. In order to produce sound signals, a 60 degrees inclined polished steel plate was used. This intelligent system had three phases (stages). In pre-processing phase, the data acquisition and pre-processing for impact signals performed and 281 sample data were used to evaluate the performance of the system. In feature extraction stage, some statistical parameters of impact signals in the time domain were selected as a feature source for sorting, and then the feature reduction was carried out using PCA. In classification phase, selected statistical features were used as the input of the ANFIS classifier. The classification accuracy of proposed PCA-ANFIS intelligent system was 100%.