Automated Rice Classification and Grading using Deep Learning
Pakistan is one of the top producers of rice in the world and is well recognized for producing and exporting high quality rice. However, we are still using manual practices for classification and grading. Manual grading and classification leads to several problems including mixing of different varie
2025-06-28 16:30:23 - Adil Khan
Automated Rice Classification and Grading using Deep Learning
Project Area of Specialization Computer ScienceProject SummaryPakistan is one of the top producers of rice in the world and is well recognized for producing and exporting high quality rice. However, we are still using manual practices for classification and grading. Manual grading and classification leads to several problems including mixing of different varieties of rice and different qualities of same rice variety. This causes problems in producing quality export product. To overcome this problem, we are proposing an automatic rice classification and grading system using machine learning algorithms. Classification of Pakistani basmati rice varieties based on rice grain features including size, shape, and color. The classified varieties are graded into quality grades (A, B, and C) by using SVM (Support Vector Machine) to differentiate between good, average, and less than the average rice grain on the basis of parameters including head rice, broken, and half rice grains.
Project ObjectivesThis study, aims to disrupt the traditional and manual system of grading and classification of rice grains by automatic system. The proposed system is envisioned to satisfy the exports requirement that will increase the international demand of rice.
OBJECTIVES:
The proposed study will achieve following objectives:
- To automatically classify the rice grains variety.
- To classify the rice grains into quality grades (A, B, and C) on the basis of broken head, half, and less than the half rice grains.
To collect image data, a digital camera will be mounted on stand at a fixed location with the distance between the lens and sample to be around 14cm. All images will be captured with black background and uniform light intensity to improve the data collection quality. We select fifteen different varieties of rice grains for experimental evaluation. All images will be stored in JPG format in separate folders named after that variety. The proposed methodology comprises of four main stages as given below
| Image Acquisition |
| Pre-processing |
| Classification of variety |
| Grading |
Algorithm 1
Input: Colored rice grains images
Output: predicted rice grains variety and grading.
Start
Step1: Data collection.
Step2: Preprocessing
2.1) Scaling
2.2) Image Enhancement.
2.3) Perform image segmentation.
2.4) Feature Extraction
Step3: Classification module.
Step4: Grading module.
Stop
Image Acquisition
Pre-processing
Classification of variety
Grading
Benefits of the ProjectRice is an important food crop and it is cultivated in several areas across Pakistan including in Punjab it is sown in Gujranwala, Sheikhupura, Wazirabad, Sialkot, Faisalabad, Sargodha, Kasure, and district Gujrat,. In Sindh, Thatta, Shikarpur, and Jacobabad, Dadu, Larkana, Badin districts are important in farming of rice crop. In Pakistan, different varieties of rice grains are mixed together causing rice adulteration that effects the national as well as an international trade and exports. There is strong need to overcome this problem by developing an automatic system for automatic grading and classification of rice grains in Pakistan. The benefits of this system include:
- Detection of rice adulteration
- Detection of quality of rice grain
- Grading of rice on the basis of quality (full rice, splitted, half)
To collect image data, a digital camera will be mounted on stand at a fixed location with the distance between the lens and sample to be around 14cm. All images will be captured with black background and uniform light intensity to improve the data collection quality. We select fifteen different varieties of rice grains for experimental evaluation. All images will be stored in JPG format in separate folders named after that variety. The proposed methodology comprises of four main stages as given below
| Image Acquisition |
| Pre-processing |
| Classification of variety |
| Grading |
Algorithm 1
Input: Colored rice grains images
Output: predicted rice grains variety and grading.
Start
Step1: Data collection.
Step2: Preprocessing
2.1) Scaling
2.2) Image Enhancement.
2.3) Perform image segmentation.
2.4) Feature Extraction
Step3: Classification module.
Step4: Grading module.
Stop
Image Acquisition
Pre-processing
Classification of variety
Grading
Final Deliverable of the Project Software SystemCore Industry AgricultureOther Industries IT , Food Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and Infrastructure, Life on LandRequired Resources| Pre-processing |