Ripe Fuji Apple Detection Model Analysis in Natural Tree Canopy

In this work we develop a novel approach for the automatic recognition of red Fuji apples within a tree canopy using three distinguishable color models in order to achieve automated harvesting. How to select the recognition model is important for the certain intelligent harvester employed to perform in real orchards. The L*a*b color model, HSI (Hue, Saturation and Intensity) color model and LCD color difference model, which are insensitive to light conditions, are analyzed and applied to detect the fruit under the different lighting conditions because the fruit has the highest red color among the objects in the image. The fuzzy 2-partition entropy, which could discriminate the object and the background in grayscale images and is obtained from the histogram, is applied to the segment the Fuji apples under complex backgrounds. A series of mathematical morphological operations are used to eliminate segmental fragments after segmentation. Finally, the proposed approach is validated on apple images taken in natural tree canopies. A contribution reported in this work, is the voting scheme added to the natural tree canopy which recognizes apples under different light influences.