Quality inspection of cereal grains using computer vision techniques
PROJECT SUMMARY: Great varieties of foods are attainable in grain structure and are essential for human nutrition caloric intake. Quality is an important factor in determining the price of rice grain at the time of procurement in the milling industry. Quality of rice&n
2025-06-28 16:34:39 - Adil Khan
Quality inspection of cereal grains using computer vision techniques
Project Area of Specialization Artificial IntelligenceProject SummaryPROJECT SUMMARY:
Great varieties of foods are attainable in grain structure and are essential for human nutrition caloric intake. Quality is an important factor in determining the price of rice grain at the time of procurement in the milling industry.
Quality of rice grain is determined by its morphological features. These morphological features include eccentricity, major axis length, minor axis length, perimeter, area and size of the grains. These grains are mostly received in mixed form. Previously, some automated systems have been designed for quality inspection. These systems are meant for partial quality inspection, but to best of our knowledge they can't distinguish rice into different variety (A class, B class, etc.). Thus, mills are left with only two methods to check contrasting variety which is DNA process and manual physical method.
Our aim is to make an automated system which will be capable of efficiently doing both tasks (complete quality inspection and classification of different varieties). In our project, we will be working on varieties of rice which are basmati, super basmati and saila rice. We will make an automated system which will work for these varieties of rice. The system will specify the quality and classification of these varieties of rice grain.
After gathering complete requirements, we will capture images. These images will be stored in the database. The captured images are then subjected to preprocessing. The system will determine the morphological features. Profiling will be done and its results will in turn help us in the quality inspection and classification using ML techniques.
The solution would be capable of automated quality inspection based on morphological parameters and also variety wise sorting. Since the system is completely automated, it will be fast and accurately available 24/7. This solution will also be cost effective.
Project ObjectivesIndustries are revolutionizing day by day. New technology is benefiting the industries and has provided a great benefit in terms of increase in quality and quantity.
The motivation here is to design a system which will efficiently analyze the quality and classify distinct types of rice grains. Grains are received mostly in mixed form thus it even gets difficult for the analyzer to differentiate between different varieties. Working on different ways for improving the quality will lead to more production.
PROJECT OBJECTIVES:
Following are the objectives of our project:
- This project is in collaboration with rice industry and these problems (quality of rice and its contrasting varieties ) were reported. The first step involves visiting the industry and understanding their requirements. After complete understanding, samples will be collected.
- The next step is to create a database and capture sample images of different varieties using some imaging modality. Algorithms training need to be performed. This automated system will perform quality inspection and will classify the variety of a particular grain.
- After completion of project, testing will be performed. Onsite testing will also be done.
This system will be implemented by using computer vision techniques. The suggested system can work well with minimum span of time and low cost.
Project Implementation MethodPROJECT IMPLEMENTATION:
Following will be the steps for implementation:
- Firstly, an image wll be acquired using an imaging module. The imaging module would particularly be a hyperspectral camera. The database is trained by feeding a number of images of each variety of grain.
- Images are then subjected to pre-processing. They will be segmented and will be prepared for further steps.
- The morphological features which include eccentricity, major and minor axis lengths, the perimeter and area of grains will be calculated which will help in calculating values.
- Then profiling will be done.
- After profiling based on results, we will use ML techniques.
It is very important to build an efficient solution that can help the industry in analysing various types of rice grain and its quality.
Currently, the need to emphasize upon this certain problem is to deliver something that would be helpful in resolving the problems that the industry is going through, that is the requirement of the time. If the previous solutions are inadequate and inefficient then there is a need in the current era to discover and ponder over a solution that emphasizes to solve the problems related to the different varieties of distinct grains. Thus, making it somehow time saving for large industries that would eventually lead to more production and more export.
- An expected outcome will be an efficient Image processing and Computer Vision technique which recognizes the sample images and will help in determining the quality of rice grains.
- Another outcome will be a Computer Vision technique to analyse different varieties of rice grain which would efficiently save time spent in manual detection.
TECHNICAL DETAILS OF FINAL DELIVERABLE:
Final deliverable would be a complete project along with whole report and paper related to it.
REPORT:
A complete report which will be comprised of following:
- Project specifications
- Approach used
- Algorithms and techniques applied
- Software and hardwares used
- Cost invested in the project
- Whole framework
PAPER:
The paper would contain whole literature review. What algorithms have previously been applied. What methodology have been used previously and what we will be using. This paper will also contain whole project flow and implementation procedure.
COMPLETE PROJECT:
Complete and tested framework.
Final Deliverable of the Project HW/SW integrated systemCore Industry FoodOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 80000 | |||
| Hyperspectral camera | Equipment | 1 | 35000 | 35000 |
| GPU | Equipment | 1 | 35000 | 35000 |
| Test bed | Miscellaneous | 1 | 5000 | 5000 |
| Connection cables | Miscellaneous | 1 | 2000 | 2000 |
| Stationery | Miscellaneous | 1 | 1000 | 1000 |
| Industry visit | Miscellaneous | 2 | 1000 | 2000 |