|
|
Bimonthly Since 1986 |
ISSN 1004-9037
|
|
|
|
|
Publication Details |
Edited by: Editorial Board of Journal of Data Acquisition and Processing
P.O. Box 2704, Beijing 100190, P.R. China
Sponsored by: Institute of Computing Technology, CAS & China Computer Federation
Undertaken by: Institute of Computing Technology, CAS
Published by: SCIENCE PRESS, BEIJING, CHINA
Distributed by:
China: All Local Post Offices
|
|
|
|
|
|
|
|
|
|
Abstract
In recent years, people have been using their eyes more and more to grade fruit. Image segmentation is an important part of developing a system for grading fruit because it lets the fruit be automatically analysed. Segmenting an image is a way to get the foreground of a digital photo while getting rid of the background that doesn't matter. Image segmentation can be done on both analogue and digital images. Object recognition has already been done with a number of different segmentation techniques. Methods like thresholding and cluster analysis are two examples of these kinds of approaches. Because natural light doesn't light up an object's surface in the same way, it's hard for standard algorithms to separate different parts of a picture of fruit. Having multiple lights on the objects of interest changes how they look, which can lead to wrong object analysis. The results of the study led to the development of a new way to divide up images of fruit. This method makes segmentation more accurate. The new method (DSDL-FCM) is made up of three algorithms: Fuzzy C-Means, adaptive K-means and modified thresholding. Combining the two methods is the only way to get more accurate results when separating images of fruits with different colours on the surface. The results showed that the new way of doing things worked well and accurately extracted individual fruits from photos.
Keyword
Shape Fruit Image , Ripened Fruit , FCM.
PDF Download (click here)
|
|
|
|
|