Eggs Detection Using Otsu Thresholding Method
Egg selection process can be performed by either natural or artificial way. The latter can be done by using image processing. It is segmenting the image of an egg to obtain objects contained in the image. This study is designed to identify fertile and infertile egg by using Otsu thresholding method. The importance of this study is designing model of egg detection system using image processing to obtain information of fertile and infertile egg. The results of egg testing show that algorithm and method that was used were able to distinguish the fertile and infertile eggs of domestic fowl.Keywords—segmentation, otsu method, fertile, infertile, egg detection machine
High demand of potentially hatched eggs offers wide business opportunity to the breeders. The rapid development of computer technology makes it possible to design a detection system of fertile and infertile eggs by using image processing. A digital image is one of the multimedia components that plays an important role as a form of virtual information. One part of the image processing which is able to identify the object of egg image is segmentation process [1]. Segmentation is a technique to identify an object in an image by using image thresholding [2-4]. However, its segmentation process is inefficient because the thresholding value is given manually. Therefore, in this study Otsu thresholding method plays a role in obtaining automatic thresholding values in the segmentation process. Some research on segmentation using the Otsu method [5-8]. In this study, we designed an egg detection box based on image processing where the lights and cameras were controlled by Arduino. The purpose of this study was to identify the fertile and infertile eggs of the domestic fowls using thresholding values obtained from Otsu method process.
This system consists of a camera used as an egg image capture, incandescent light is used as an egg shine when it will be captured by the camera where the lights and cameras were controlled by Arduino. The egg is placed between the lights and the camera. The eggs used in this study were eggs of domestic fowl. Fig. 1.Design Feature of an Egg Detection Box Each stage in Fig. 2, described in the following sections: First stage: Input stage, egg image of domestic fowls is inputted. The eggs were captured using a camera and stored in a format.jpg. Second stage: RGB image conversion to Grayscale Third stage: Segmentation process with Otsu method. The of Otsu thresholding is as follow: 1)Reading the original image as RGB color image (Red, Green, Blue). 2)Converting RGB image to Grayscale. 3)Reading grayscale image size. 4)Initializing histogram values = 0. 5)Calculating the histogram values. 6)Calculating p(i) value (the probability of intensity value). 7)Calculating mT value (the total average value). 8)Calculating t value (thresholding) namely a)Initialization of the threshold value (t = 0). b)Calculating the weighted values (w1 and w2) in both classes (object and background). c)Calculating the mean values of classes (m1 and m2). d)Calculating the BCV value (Between Class-Variance).