Machine learning base fruit grading system using Feature Fusion

Project Title

Machine learning base fruit grading system using Feature Fusion

Project Area of Specialization Artificial IntelligenceProject Summary

The aim of this project is to designed a fruit grading system using the Machine learning and feature
fusion techniques. A publically available fruit dataset is used to train the proposed CNN architecture.
We created a secondary dataset by making it the binary class dataset of apples only. A complete fruit
grading system is designed and developed, composed of software and hardware modules. Initially, an
apple is placed on a conveyer belt and image is acquired which is further pre-processed by applying
cropping and background removal. Afterwards, handcrafted and deep features are extracted and these
features are fused together and forwarded to classifier for binary classification of apple categories. On
the basis of classification result, the apple is moved to its relevant class/category. The grading
algorithm is designed using MatLab, whereas, the hardware modules include: a conveyer system, DC
and Servo Motors, Camera with uniform illumination and a processor to execute the algorithm. This
system can be used in the agriculture sectors to provide best quality fruits.

Project Objectives

1. To designed machine vision system that is capable of automating the visual inspection of fruits in real-time.
2. To attain more prominent Features information by using Features Fusion method.
3. To improve the accuracy of system.
4. To reduce time and maintain the quality of fruits.

Project Implementation Method

1^st Phase:

Algorithm designing (CNN)

2nd Phase:

Training and testing

3rd Phase:

Hardware implementation

4th Phase:

Final testing of software and hardware.

Benefits of the Project

1. Quality control applications
2. Agriculture industries
3. Liquor industries

This project  will be use for fuits grading qualitatively wherever, it is needed .

Technical Details of Final Deliverable

The final deliverable of the project is complete fruit grading system The grading algorithm is designed using MatLab, whereas, the hardware modules include: a conveyer system, DC and Servo Motors, Camera with uniform illumination and a processor to execute the algorithm. A publically available fruit dataset is used to train the proposed CNN architecture.
We created a secondary dataset by making it the binary class dataset of apples only. A complete fruit
grading system is designed and developed, composed of software and hardware modules. Initially, an
apple is placed on a conveyer belt and image is acquired which is further pre-processed by applying
cropping and background removal. Afterwards, handcrafted and deep features are extracted and these
features are fused together and forwarded to classifier for binary classification of apple categories. On
the basis of classification result, the apple is moved to its relevant class/category. 

Final Deliverable of the Project HW/SW integrated systemCore Industry AgricultureOther Industries Food , Manufacturing , Health Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Good Health and Well-Being for People, Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	69270
DC Gear motor	Equipment	1	1300	1300
Servo motors	Equipment	2	700	1400
IR sensors	Equipment	3	100	300
Motor Driver IBT2	Equipment	1	1800	1800
Arduino UNO	Equipment	1	850	850
Camera	Equipment	1	3600	3600
Conveyor belt system	Equipment	1	15000	15000
Power Supply	Equipment	1	2000	2000
PC or KIT	Equipment	1	40000	40000
Jumpers Wires	Equipment	80	5	400
HDMI cable	Equipment	1	150	150
Ply wood	Miscellaneous	1	600	600
Glue	Miscellaneous	1	120	120
Nuts Bolts	Equipment	15	10	150
Screws	Equipment	20	5	100
Acrylic sheet	Miscellaneous	1	500	500
Glue Gun /sticks	Miscellaneous	1	600	600
Led light	Equipment	1	100	100
Connection wires	Equipment	3	100	300