COMPUTER VISION SYSTEM FOR AUTOMATED INSPECTION AND GRADING OF FRUITS

Project Title

COMPUTER VISION SYSTEM FOR AUTOMATED INSPECTION AND GRADING OF FRUITS

Project Area of Specialization Artificial IntelligenceProject Summary

Being a part of an agricultural and highly populated country, it is important to develop the techniques that could help increase the speed of processing of fruits and vegetables after collecting them from different farms around the country so that they can reach to local market as well as international market as soon as possible. This processing step mostly includes collection, washing, inspection, grading, packing and supplying. 

     In this processing step, inspection and grading are two most challenging steps. Inspection is the process of checking the quality, size, and shape while grading is the process of classifying the fruit or vegetable based on inspection. These steps are mostly done by human labor, which can be very tiring, slow and sometimes inaccurate. So, to relieve humans and to speed up the process, an automated system for sorting and grading is a requirement. As humans use their sight sense to sort and grade so a vision system is necessary that can use cameras to sort and grade like humans. In this thesis, we are proposing a design of a computer vision system that uses cameras to take pictures of incoming fruit on conveyor belt and inspect and grade it based on the size, color and defects on surface. This system will help us classify apples as red and yellow apples and as well as grade apples in two categories which are Grade ‘A’ and Grade ‘B’. We are using a KNN classifier to classify apples according to the dataset given to the KNN model as KNN is a supervised classifier. 

Project Objectives

The objective of this project is to present a model to make the inspection and grading process fast and inexpensive. To accomplish this, we sub-divided it into smaller objectives which includes:

Automated apples sorting hardware

The major aim is to develop reliable and automatic system as an alternate solution to human inspection and grading of apples. So, we are developing a computer  vision system for automated inspection and grading of apple. This objective directed us to use a conveyor belt for carrying separation of apples.

Development of fast and accurate algorithm

There was need for the computationally inexpensive algorithm which also provides the better accuracy. In this regard, different algorithm-based approaches were studied and tested. K-Nearest Neighbor (K-NN) algorithm was selected as it provides good accuracy at lower computation expense.

Project Implementation Method

The main challenges in Computer vision based smart inspection of Surgical Instruments are as:

• Low cost hardware development
• Development of fast and accurate algorithm

Low cost hardware development

The low-cost hardware development directed us to use a conveyor belt for carrying instruments. Once an apple is placed on it, conveyor belt will move it to specific position for real time image acquisition. IR sensors have been used to detect apple reached below cameras. Later, a separating plate is used to put the defected and good quality apple into the respective storage bins.

                      Development of fast and accurate algorithm

We are developing an algorithm using image processing and computer vision toolboxes in MATLAB. The overall algorithm for the operation of system is as followed,

After converting apple image into features, K-Nearest Neighbor (K-NN) algorithm is used to classify the apple as defected or grade A and grade B. As per decision, our project’s hardware uses a separating plate that can put the defected and good quality apple into the respective storage bins.

      Benefits of the Project

The major benefits which the project will demonstrate are at the sorting and grading stage are as mentioned below:

Automated grading of fruits is faster as compared to tradition methods (manual grading).

• It will reduce human labor.
• Reduce labor’ cost.
• It will provide more accuracy because it eliminates human errors.

Technical Details of Final Deliverable  

Project delivered would be a computer-vision systtem which can do the following:

1. Extract the area, color and texture features of apple for classification
2. Classifying the apples to A grade,  B grade and defective

Some of the major hardware components are:

Conveyor Belt

As per industrial requirement, we have prepared a slider bed belt conveyor for carrying apple. Following are specs of conveyor:

• Controlled by Vs code through Arduino
• Frame is made of aluminium and wood combination
• Rotation is controlled by 12V DC motor

HD webcam

The system includes 2 High definition camera providing high-resolution picture for apple @ 30fps.

Final Deliverable of the Project HW/SW integrated systemCore Industry AgricultureOther Industries IT Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Industry, Innovation and InfrastructureRequired Resources