Interational Conference on Communication and Signal Processing, April 3-5, 2014, India A Comparative Analysis of Edge and Color Based Segmentation for Orange Fruit Recognition R. Thendral, A. Suhasini, and N. Senthil Abstract-In this paper, we presented two segmentation methods. Edge based and color based detection methods were used to segment images of orange fruits obtained under natural lighting conditions. Twenty digitized images of orange fruits were randomly selected from the Internet in order to fnd an orange in each image and to determine its location. We compared the results of both segmentation results and the color based segmentation outperforms the edge based segmentation in all aspects. The MATLAB image processing toolbox is used for the computation and comparison results are shown in the segmented image results. Index Terms- color based segmentation, edge based segmentation, machine vision, Orange harvesting. I. INTRODUCTION Fresh fuit harvesting is a sensitive operation. According to the [1] cost of harvesting by labors is very expensive and time consuming. In addition, picking of fuits by hand is very tedious. To solve these problems, human works can be replaced by automatic robots. Automatic harvesting operations reduce the harvesting costs. Therefore, automation and use of image processing methods in agriculture have become a major issue in recent years [2] - [4]. Jimenez et al. [5] presented a review of fuit recognition systems. Schertz and Brown [6] considered both individual fuit harvest and mass harvest. The fuit surface was identifed by photometric comparison. They reported that ten times more light refected fom a fuit than fom a leaf Parrish and Goksel [7] reported an experimental automated apple harvesting system in a laboratory. They used different green and red optical flters and a black-and-white TV camera to obtain images for apple orchard and developed an automated experimental system for harvesting apple. Problems accompanying with mechanical harvesting resulted in the development of robotic harvesting methods, thereby prototype machine vision based harvesters has been increasingly being developed. Several studies have been carried out to design a harvesting robot to pick up fuits fom the trees or plants [8 - 12]. Bulanon and Kataoka [13] developed an algorithm for R. Thendral is with the Department of Computer Science and Engineering, Research scholar, Annamalai University, Annamalai Nagar, Chidambaram (e· mail: thendralamutha@gmail. com). A. Suhasini is with the Department of Computer Science and Engineering, Associate Professor, Annamalai University, Annamalai Nagar, Chidambaam (e-mail: suha_babu@yahoo.com). N. Senthil is with the Department of Computer Science and Engineering, PG Student,Annamalai University,Annamalai Nagar, Chidambaram (e-mail: engzenthil@gmail. com). 978-1-4799-3358-7114/$31.00 ©2014 IEEE the automatic recognition of ripe Fuji apples fom the tree; they enhanced the difference between fuit fom other objects within the image, based on the difference between luminance and red color (R-Y). Hanan et al. [14] developed a vision system to pick orange using a harvesting robot. The RI(R+G+B) feature was used for recognition of orange fuits on the tree. The automated harvesting system [15] should perform the following operations: (1) recognize and locate the fuit; (2) reach for the fuit; (3) detach the fuit without causing damage both to the fuit and the tree; and (4) move easily in the orchard. There are an increasing number of robotics applications aimed at detecting fuits fom images or videos [16 - 19]. The frst major task of a harvesting robot is to recognize and localize the fuit on the tree. This paper focuses on recognition of orange fuits by using edge and color based segmentation methods and we compare the results of both segmentation results. In the next section the details of our proposed edge and color based segmentation methods are presented. The results and discussion is given in section III. Finally, in section IV, conclusions of the proposed approach were presented. II. MATERIALS AND METHODS Twenty digitized images of orange fuits were randomly selected fom the Interet and all the images are of different pixel size. These images were captured in different lighting conditions with different background and different camera distances. A. Image Processing Algorithm The input sectional tree images were having different lighting conditions. The fuit regions in many images were under the shadow of the leaves and branches. Recognition is the process of separating an object of interest fom the background. This is an image processing procedure called segmentation [20]. In order to segment the acquired images, two algorithms were developed: edge based and color based segmentation. B. Edge Detection Based Algorithm The Canny edge detector [21] is a popular method for fmding edges that begins by smoothing an image by convolving it with a Gaussian of a given sigma value. Based on the smoothed image, derivatives in both the x and y direction are computed; these in turn are used to compute the gradient magnitude of the image. Once the gradient magnitude +IEEE Advancng Technology for Humanit 463