Robust Approach for Fruit and Vegetable Classification

Efficient detection of ‘species and variety’ of fruits and vegetables is one of the major challenges in the fruit recognition. This paper presents a texture feature based on the sum and difference of the intensity values of the neighboring pixels of the color images. The proposed improved sum and difference histogram (ISADH) texture feature is an extension of the Unser's descriptor. The result shows that proposed texture feature performs better than the other state of the art color and texture features in the fruit and vegetables classification. The classification accuracy for the proposed ISADH texture feature is achieved up to 99%.