Fruit recognition based on multi-feature and multi-decision

In order to overcome the existing fruit recognition method only for single feature recognition which leads the problem of lower recognition rate, this paper proposes a recognition method based on multi-feature and multi-decision. Firstly, we preprocess the fruit image which is to be classified, separating foreground and background, and then we divide the target area. Secondly, in order to take full advantage of multi-independence and complementarity between features, we extract the color, shape and texture of fruits image and put these features into the BP neural network to classify. Finally, the results of different features are put into the decision-making mechanisms to obtain a final recognition result. The experiment results on object library and self-built library that is built by ourselves show that the method has good identification ability to fruit recognition and achieve the goal to identify different fruit.