Application of Coconut Recognition Using Laplacian of Gaussian and HSV based on Android

ABSTRACT Madan coconut is a type of the coconut that can be used as a traditional medicine and means of upakara for Hindus in Bali, Indonesia. Madan coconut has various special characteristics or features in each species, but with many features of each type makes it difficult to remember and recognize the types of Madan coconut. Implementation of digital image processing technology can be a solution to facilitate in recognizing the types of Madan coconut. Shape and color are the main features that can be used to recognize the types of Madan coconut, while Laplacian of Gaussian and HSV methods are used to obtain both features. The Laplacian of Gaussian method was used to detect the coconut shape, meanwhile, HSV was used to detect the coconut color. Both features had to undergo the extraction stage through Invariant Moments and a matching process using Euclidean Distance calculations to find the smallest distance in the recognition process. This designed application was able to generate success in recognizing coconut type based on its form feature to 52.5% and 85% of its color feature, with a percentage success of matching was 70%.