Strawberry Fruit Identification and Ripeness Discrimination under Multi-Color Model

Abstract

Aiming at the problem of inaccurate fruit recognition and ripeness judgment of strawberries in natural environment due to lighting and shading, a method of strawberry fruit recognition and ripeness discrimination under multi-color model is proposed. The method firstly fuses segmented images of YCrCb, Lab and HLS color models, secondly separates strawberries one by one using the watershed algorithm, then uses the Hough transform to locate strawberries, and finally judges whether strawberries are ripe by comparing all fruit and ripe fruit data. The experiments show that the method achieves 93.6% recognition success rate for elevated strawberries and 93.25% accuracy for green-ripe and red-ripe strawberries.