Image Processing Application on Automatic Fruit Detection for Agriculture Industry

The robot brings automation to every sector of human life, including agriculture. Automation in agriculture might be the solution to get a higher quality harvest and less dependency on human farming. The most suitable type of robot for harvesting is an arm robot manipulator. The harvesting robot needs "eye" to "see" the crop/fruit to be harvested. The detection is made possible by using image processing to get the fruit position. The fruit position is the input for a visual servoing robot. The image processing needs to be simple and effective to ensure less computational time to facilitate the limited memory of the available microcontroller. This paper proposes three image processing methods, i.e., image segmentation, edge detection, and blob analysis. The processes were conducted in SCILAB, and three fruit were used as the model, i.e., oranges, grapes, and tomato cherry. The results showed that all the fruit are detected and isolated by the vegetation background.