Robotic Apple Harvesting Using Computer Vision Based on Shape & Colour Analysis and Object Positioning

In this paper, the new automatic apple harvesting algorithm is designed for harvesting the apple to overcome the labour shortage problem. The apple detection and positioning is the major issue in the automatic robotic systems for apple harvesting. The major hurdle in the fruit detection (apple detection in our case) is the overlapping fruits. The robots are not capable of detecting and separating the layered objects or objects placed over objects. This problem leads to the misdetection or no-detection of the objects. The new algorithm has been developed to address all of the above discussed issues. This algorithm has been developed using color and shape analysis, and has been designed for the object positioning in 2-D and 3-D space. The major objective of this algorithm is to eliminate the overlapping objects, calculating 2-D distances, 3-D distances, etc. The accuracy of the robotic arm movement is increased by using this algorithm. [1] The results have shown that the proposed algorithm is very effective and almost 94-95% accurate in detecting the ripen apples on the apple trees.