Integration CLAHE and Seeded Region Growing for Segmentation of Rubber Tree in HSI Color Space

Abstract

Rubber tree growth is excellent when in the tropics. Rubber trees that mature can be processed to extract the sap. The image segmentation process can be carried out first as the initial process of the maturity level classification. An accurate segmentation method and fast processing time are needed to support that process. We propose integrating CLAHE and Seeded Region Growing to segment rubber trees in HSI color space. This method uses a hue image as input, then enhancement of the sharpness image uses CLAHE. From this process, a seeded region growing segmentation method is used to separate the rubber tree object from the background. The result in this method shows that the average RAE is 31.02%, ME 21.61%, MHD 15.04%, and the processing time is 5.18 seconds. Based on these results, this can prove that the method is good enough to be applied on rubber tree images taken directly from a forest where the image has complexity texture, risk of multi-object, and complexity color.