High Speed Separation of Potato and Clod Using an Acoustic Based Intelligent System

Separating clods from potato tubers is one of the most challenging jobs in a potato harvester. In this study، an acoustic-based intelligent system was developed for high speed discriminating between potato tubers and clods. About 500kg mixture of potato tubers and clods were put on the belt conveyer and impacted on a steel plate in four different velocities. The resulting acoustic signals were recorded، processed and potential features were extracted from the analysis of sound signals in both time and frequency domains. A multilayer perceptron (MLP) neural network with a back propagation algorithm was used for pattern recognition. Altogether، 17 potential discriminating features were selected and fed as input vector to artificial neural network (ANN) models. Optimal network was selected based on mean square error، correct detection rate and correlation coefficient. At the velocity of 1 ms-1 of the belt، detection accuracy of the presented system was about 97. 3% and 97. 6% for potato and clod، respectively. Detection accuracy decreased by increasing belt velocity. A potato harvester by using this system can perform at capacity of 20 ton hr-1 by accuracy of about 97%.