Segmentation for embryonated Egg Images Detection using the K-Means Algorithm in Image Processing

Abstract

Image segmentation is often used in the process of detecting separated objects. In this study, the application of image segmentation in the detection of egg fertility. The fertility of eggs in hatching is checked between the seventh day to separate eggs that have embryos (fertile). Application of technology, one of which is image processing, requires a preprocessing process to detect the presence of embryos in eggs. In this research, the preprocessing process can help divide the color image of chicken eggs using K-means Algorithm. K-means used are based on a matrix of color images (three color components, red, green, and blue) with a value of k = 50. The result is a segmented color image. The K-means segmentation image is converted to a grayscale image and processed with image enhancement. The final process is the result of image enhancement morphological processes (dilated with string size six) and converted to black and white images to clarify the segmentation process occurs. Based on experiments, the process can run well, with the value of MSSIM = 0.9995 (Mean of the SSIM), which means that the image information is under the original image. Besides, the processed object gives a clear picture of the embryo in the egg, which shows that k-means segmentation can help the process of detecting the presence or absence of embryos in the egg.