Multi Thresholding Berbasis Algoritma Sinus Cosinus Untuk Segmentasi Citra Pantai

Abstract

This paper examines the coastal image segmentation approach to clustering the intensity of land pixels and ocean pixels. This beach image segmentation is generated by the Otsu algorithm which is based on the principle of bilevel threshoding that will be optimized using a metaheuristic algorithm, namely the Sine Cosine Algorithm or known as the SCA method. This sine cosine algorithm is based on the principle of multilevel thresholding and uses a mathematical model based on trigonometric functions to solve optimization problems and to obtain more optimal settings with shorter computation time. Image histograms are also added to obtain a graph that can describe the distribution of pixel intensity values from a beach image. The measurement of this coastal image segmentation optimization uses FSIM (Feature Similarity Index Measure), SSIM (Structural Similarity Index Measure) and the computational time of processing the coastal image. The result of this research is the image of the beach as the result of segmentation optimization of the sine cosine algorithm with the value of FSIM, SSIM and computational processing time which increases along with the increase in the threshold value used.