Defected fruit detection

“Automatic Fruit Quality
Inspection System”

ABSTRACT
• An automatic fruit quality inspection system for sorting and grading of
tomato fruit and defected tomato detection discussed here.The main
aim of this system is to replace the manual inspection system.
• This helps in speed up the process improve accuracy and efficiency
and reduce time. This system collect image from camera which is
placed on conveyor belt.
• Then image processing is done to get required features of fruits such as
texture, color and size.
• Defected fruit is detected based on blob detection, color detection is
done based on thresholding.
• Size detection is based on binary image of tomato. Sorting is done
based on color and grading is done based on size.

CONTENTS
• Introduction
• Literature Survey
• Overview of Proposed System
• Discussion and Results
• Conclusion
• Future Scope
• References

INTRODUCTION
• India is an agriculture country. All the pre-harvest and post-harvest process are done
manually with help of labour.
• The post-harvest process includes sorting and grading of fruits. Different quality
factors like texture, shape, color, size and volume, and internal quality factors are
taste, sweetness, flavour, aroma, nutrients , carbohydrates present in that fruit are
considered for sorting and grading of fruits.
• Automation is playing important role in day today life. Their main source of
income is agriculture. Exporting of fresh fruit is increased day to day from India.
People are very conscious about their health; they prefer only fresh, good quality
fruit.
• There is confusion between tomato is a fruit or vegetable. As per Fruit definition
Fruits are developed from ovary. Fruit contains seeds of plant. So based on above
tomato is a fruit. Indian market export tomatoes to foreign countries.

LITERATURE SURVEY
• Texture, Color and Size are the important parameters for fruit quality
identification.
• The color recognition is very important process in ripeness detection.
The ripeness detection is external quality factor. But texture is also very
important.
• Because of texture defected fruit can be recognized. Texture analysis
detects the non-uniformity of fruit outer surface.
• The size is also important parameter. It clearly seen parameter all
customer select fruit based on size.

LITERATURE SURVEY
• Dah-Jye Lee et al.(2011) [1]designed a Direct color mapping technique for
obtaining the color of fruit. Advantage of this system is that it have
arrangement of adjusting color preferences or grading parameters as per
application. It is user friendly technique
• H. Dang et al (2010)[2] developed a system for fruit size detection. All
results are calculated based on image processing. Edge detection ,then fruit
size detection and based on size grading of fruit is done. OSTU (maximum
classes square error) is used to get binary image . For edge sequence
detection 8-connected boundary method is used. For diameter detection
symmetry is considered. Symmetry gives center coordinate. line diameter is
calculated based on center point and axis going from centre. For accurate
result two edge points are searched. If fruit is rotated then also diameter
shown is same then the diameter indicated true fruit size. Then based on
accurate size grading is done.

LITERATURE SURVEY
• H. Alimohamadi et al (2013)[5] designed a system for skin defect detection in
fruits. Gabor wavelet Filter is used for defect detection. Convert color image
into texture image and then on that image Bank of Gabor filter is applied.
Gabor Filter is linear filter and used as edge detector. Gabor filters with 4 scales
and 6 rotations used in this paper. Obtained response shows image pixel is as
defected or normal skin. Optimal filter is chosen from bank of Gabor filters
depending upon the response . Thresholding the response of the optimal filter.
Based on thresholding skin defect is detected.
• Y. Wang et al (2010) [6] designed a system in which Fruit quality inspection is
done based on fruits surface color. It is nondestructive method. Fruit image is
captured with camera, RGB image is converted into the HSI color model.
Image is segmented based on hue value ,separate fruit and its background.
Histogram of Hue and Saturation of fruits surface color is calculated. Input is
given as histogram ,output obtained earlier of Hue and Saturation of surface
color of fruit from back propagation network. Output as quality description of
given tested fruit. They performed experiment on banana and result obtained is
accurate.

LITERATURE SURVEY
• S. A. Khoje et al developed a system of fruit grading for automated skin defect
identification using Discrete Curvelet Transform. Discrete Curvelet transform is
used for texture analysis. Multi resolution approach is used in DCT, By using
low and high resolution capability local and global features of that fruit are
found . Energy, entropy, mean and standard deviation of each good and
defected fruit is calculated these are called as features.. Support vector machine
and Probabilistic neural network for classification of good fruits and defected
fruits. Based on obtained result SVM is more accurate than PNN.
• M. Satpute et al in this paper different technique of color, size, shape, texture
and volume detection of fruit are discussed.

OVERVIEW OF PROPOSED
SYSTEM
A. BLOCK DIAGRAM

SYSTEM
B. DESCRIPTION
• The Infrared sensor is placed on conveyor belt, when fruit is come in front of
infrared sensor message will display as fruit detected then conveyor belt moves with
small distance an stop when fruit come exactly in front of camera. Camera (High
Quality CMOS sensor, 25 MPs, 30fps) always in video mode. When fruit is detected
the image processing is done on that image captures and color is detected. Red,
Green, Yellow color are detected. The system is divided into hardware control and
image processing. The image processing results is based on camera image. The
results such color detected. Second part is hardware is controlled based on color
detection.
• The image processing is done by software OpenCv using a language python. The
software is divided into two parts first one is for image analysis and other is for
controlling hardware based on image processing results. As per Fig 1 the system is
operated in two different scenarios in first the image is captured with camera the all
the image processing is done in the control module. All the process are shown on
monitor and then based on decision taken by control module. The conveyor
assembly is operated.

SYSTEM
C. PROCESSING FLOW
Fig2. The fruit sorting and grading flow

Image Preprocessing
• In the process of fruit sorting and grading to work system efficiently proper
image acquisition is very important. The image is captured with camera that
image is with noise and its features are not clearly seen so image preprocessing
is done on that image.
• In this project the features required are color, texture and size. To get exact
feaure preprocessing is done on acquired image. The main aim of image
processing is an improvement of image so that unwanted distortions are
suppresed and enhance image features which are important for further
processing.
• The basic steps of preprocessing are first convert RGB image to gray scale
image. Then image histogram equilization is applied on gray image. This helps
in adjusting image intensities in order to enhance contrast. Remove noise with
filter, here we use median filter for removing noise.The laplacian is used for
edge detection as it highlights the region with rapid intensity change. So this
enhanced, noise free, filtered image is ready for further processing.

Defected Fruit Detection
This is the flow for defected fruit detection. The image taken is RGB image.
Firstly this image is converted to gray scale and the edge detection is performed
and the blob detection is performed and defected region is marked with red
circle. In [9] the Discrete Curvelet Transform is used for defected skin detection.
Another method used for defected fruit detection is that RGB image is converted
to YCR color space. Then lower and upper ranges are defined. Then ranges of
binary image are defined. Then convert single channel mask back into 3
channels.
Fig4. Flow for defected fruit detection

Color Detection
The figure.5 shows the flow of color detection
based sorting of fruit. Image captured that is RGB
image is given to preprocessing. Define the list of
boundaries for BGR color. Apply loop over the
boundaries. Find the colors within the specified
boundaries and apply the mask. In [14] the sorting
of mango is done based on color and grading is
done based on both color and size here fuzzy
logic is used for decision.

Size Detection
The image is captured. First convert RGB image to gray scale image.
Then OSTU thresholding the binary thresholding is performed on that
image. Then morphological operation such as dilation then erosion is
performed. For boundary detection opening is performed. Then major
axis and minor axis length is calculated. Then size is decided as small,
medium and large. Fruit size detection involves major axis calculation.
The palette used for sorting changes its direction when image
processing results are obtained.

DISCUSSION AND RESULTS
A. Defected Fruit Detection
Tomato is having different defects. Here blob
detection technology is used for defect
detection. This is the technique by which
specific region is detected which is differ in
properties compared to surrounding region
such as color or brightness. In the following
Fig.6(a). Color image, gray image and image
showing defect on tomato surface is shown.
The defects are highlighted with red circles.
The defect which are present on red, green
and yellow tomato are shown with green color
as shown in fig6(b).

Fig 7. Red color detected
Fig 9. Yellow color detected
Fig 8. Green color detected
B. Color Detection

C. Size Detection
Fruit size detection involves major axis calculation. The palette used for sorting
changes its direction when image processing results are obtained.

CONCLUSION
In this paper automatic vision based system is discussed for sorting and
grading of fruits based on its color and size respectively. The test
performed on tomato for defect detection detects defected fruit. This
test is performed for three color detection Red, Green and Yellow. And
for three different sizes large medium and small. The variation in speed
of conveyor and light, camera resolution affects the system. The
accuracy of green color detection is 94.28% which more than red and
yellow. The accuracy of defect fruit detection is up to 90%.This system
is much closer to manual expert judge.

FUTURE SCOPE
Further design can be modified by increasing size of
conveyor belt so that it is possible to perform quality
inspection of large fruit than tomato, and increase accuracy
of the system so that it can differentiate between artificial ,
hybrid color from original fruit color.

REFERENCES
[1] D.Lee, J.Archibald and G.Xiong,“ Rapid Color Grading for Fruit Quality Evaluation Using
Direct Color Mapping”, IEEE Transactions On Automation Science And Engineering, Vol 8,
No.2,pp.292-302, April 2011.
[2] H. Dang, J. Song and Q. Guo, “A Fruit Size Detecting and Grading System Based on Image
Processing” 2010 Second International Conference on Intelligent Human- Machine systems and
Cybernetics, pp.83-86.
[3] M. Alfatni , A. Shariff , M. Abdullah , M Hamiruce B. Marhaban , Osama M. Ben Saaed , “The
application of internal grading system technologies for agricultural products Review”, Journal of Food
Engineering 116(2013), pp.703-725.
[4] M.Dadwal and V. Banga, “Color Image Segmentation for Fruit Ripeness Detection: A Review”,
2nd International Conference on Electrical Electronics and Civil Engineering (ICEECE’ 2012)
Singapore, April 2829, 2012, pp.190-193.
[5] H. Alimohamadi, A Ahmadyfard, “Detecting Skin Defect of Fruits Using Optimal Gabor Wavelet
Filter”, International Conference on Digital Image Processing, pp.402-406.
[6] Y. Wang, Y. Cui, G. Huang, P. Zhang, S. Chen, “Study on Fruit Quality Inspection Based on Its
Surface Color in Produce Logistics”,2010 International Conference on Manufacturing
Automation,pp.107-111.
[7] B. Ojeda- Magana, R. Ruelas, J. Quintanilla-Domingvez and D. Andina, “ Colour Image
Segmentation by Partitional Clustering Algorithms”, pp.2828-2833.

REFERENCES
[8] C. Lin , C. Su, H. Haung and K. fan , “ Colour Image Segmentation Using Relative Values of
RGB in Various Illumination Circumstances”, International Journal of Computers, Issue 2, Vol-5,
2011,pp.252-261.
[9] S. Khoje, S. Bodhe, A. Adsul, “ Automated Skin Defect Identification System for Fruit Grading
Based on Discrete Curvelet Transform”, International Journal of Engineering and Technology (IJET).
Vol 5, No.4, pp.3251-3256, Aug-Sept 2013.
[10] R. Mahendran, GC Jayashree, K Alagusundaram, “ Application of Computer Vision Technique
on Sorting and Gradinng of Fruits and vegetables”,Abd El-Salam et al..,J Food Process Technol
2011,pp.1-7.
[11] H.C. Garcia, J. Villalobos, R. Pan and G. Runger, “A Novel Feature Selection Methodology for
Automated Inspection System”, IEEE Transactions on pattern analysis and machine intelligence,
Vol.31, NO.7,pp.1338-1344,July 2009.
[12] A. Mizushima , R. Lu, “An image segmentation method for apple sorting and grading using
support vector machine and Otsu’s method”, Computers and Electronics in Agriculture 94 (2013),pp.
29–37
[13] A. AL-Marakeby, A. Aly, F. Salem, “Fast Quality Inspection of Food Products using Computer
Vision”, International Journal of Advanced Research in Computer and Communication Engineering
Vol. 2, Issue 11, pp.4168-4171,November 2013
[14] C. Nandi, B. Tadu, C. Koley, “ Machine Vision Based Technique for Automatic Mango Fruit
Sorting and Grading Based on Maturity Level and Size”, Springer International Publishing
Switzerland 2014, pp.27-46.

REFERENCES
[15] S. Khoje, S. Bodhe, A. Adsul, “ Automated Skin Defect Identification System for Fruit Grading
Based on Discrete Curvelet Transform”, International Journal of Engineering and Technology (IJET).
Vol 5, No.4, pp.3251-3256,Aug-Sept 2013.
[16] M. Satpute, S. Jagdale, “Color, Size, Volume, Shape and Texture Feature Extraction Techniques
for Fruits: A Review” , International Research Journal of Engineering and Technology (IRJET),
Volume: 03 Issue: 02 , Feb-2016

Defected fruit detection

More Related Content

What's hot (20)

Similar to Defected fruit detection (20)

Recently uploaded (20)

Defected fruit detection