www.tjprc.org editor@tjprc.org
AUTOMIZATION OF AGRICULTURE PRODUCTS DEFECT DETECTION
AND GRADING USING IMAGE PROCESSING SYSTEM
GAURI DEEPAK PATNE
1
& P. A. GHONGE
2
1
Research Scholar, Department of Computer Engineering, Yadavrao Tasgaonkar College of
Engineering & Management, Karjat, Maharashtra, India
2
Associate Professor, Yadavrao Tasgaonkar College of Engineering & Management,
Karjat, Maharashtra, India
ABSTRACT
Grading of fruit is important phase after harvesting and before marketing. The automatic fruit grading system
extracts its defective region and grading, according to its level of defection. The classification of Fruit is done into four
categories by considering smaller changes in defective parts, so it increases output efficiency and user acceptance level of
different fruit categories.
In this proposed system of image processing algorithm, rgb2gray method and median filter are used to pre-
processed the input image and convert it into gray scaled image; after that input image is segmented using modern
iterative tri-class threshold based on Otsu’s method for classifications in terms of defect, extract statistical features
texture and shape, with the normalized Symmetric GLCM method. For testing purpose, Apple images data collected from
the database provided by Mechanics and Construction Department of Gem-bloux Agricultural University of Belgium [8]
have been used. Fruits are classified into four categories by using kNN method with 95% accuracy (Category 1, Category
2, Category 3 and Category 4).
KEYWORDS: Otsu’s, Tri-class Thresholding, GLCM, Multi Spectral, Four Categories & KNN
Received: May 31, 2018; Accepted: Jun 21, 2018; Published: Jul 28, 2018; Paper Id.: IJCSETIRAUG20184
INTRODUCTION
Image processing applications are extended by giving computer vision in various fields. In image
processing, different information is extracted from image, also provide necessary theory and algorithm in the field
of multimedia, medical and agriculture. Here, combination of hardware and software is used for the purpose of
processing images. Now a day, computer vision technology is adopted in agriculture field to detect defects in
products. In India, agriculture is most important field because; most of the people depend on farming. So, there is
needed to detect defect and improve product quality, it will automatically increase market price. In food industries,
before sending product to market, there is need of grading. Grading is normally performed on the basis of external
defects on skin of fruits or vegetables. Traditionally, food products are inspected by manual inspection, which is
more time consuming and less efficient for large industries. The computer vision techniques are more helpful, as it
gives consistent result.
In the area of Image processing using advanced technology, it is possible to perform automatic grading of
fruits and vegetables. This reduces the cost; improve quality, increase market price and user acceptance ratio. Some
automatic grading systems are available, that uses classification techniques on the basis of color, texture, size and
Original Article
International Journal of Computer Science Engineering
and Information Technology Research (IJCSEITR)
ISSN (P): 2249-6831; ISSN (E): 2249-7943
Vol. 8, Issue 3, Aug 2018, 25-32
© TJPRC Pvt. Ltd.