Automatic Classification of Chestnuts Based on Decision Fusion

Image acquisition system of chestnuts has been designed in this study. After the preprocessing on the images, including image de-noising, image enhancement, image segmentation, edge extraction and edge linking, the parameters of the shape features of chestnuts such as the areas, the perimeter, the circularity, the eccentricity and the degree of circularity have been extracted. This study proposes a decision fusion strategy using L-M neural network for automatic classification of chestnuts. The simulation results present that the classification accuracy has been improved using the method based on decision fusion compared with the traditional methods. Keywords-decision fusion; image processing; L-M neural network; automatic classification; chestnut