- Version
- Download 31
- File Size 828.65 KB
- File Count 1
- Create Date 19/04/2023
- Last Updated 19/04/2023
Tomato Grading with Respect to Maturity Levels Using CNN
Tomato Grading with Respect to Maturity Levels Using CNN
| M Jnanesh A Shetty | M Vignesh |
| Department of Electronics and Instrumentation Engineering | Department of Electronics and Instrumentation Engineering |
| St Joseph’s College of Engineering | St Joseph’s College of Engineering |
| Chennai-600119 | Chennai-600119 |
| mjnaneshshetty@gmail.com |
|
ABSTRACT
Food is a basic requirement for every living being to survive. It provides us with energy to perform our daily routine, being one of the basic requirements a majority of the population doesn’t have the access to quality food. Due to improper food management over1.6 billion tons of food waste was generated in the year 2022.
In order to tackle this problem we need to reach the root of the problem. The root of the problem lies during the harvesting of fruits vegetable, grains, pulses etc... Were due to improper segregation and management of waste the quality of food being produced as the end product. In the wastage being produced this paper focuses on the agricultural wastes being generated. In India over 350 million tonnes of agricultural waste every year. This waste generated can be regulated by performing a precise harvesting. This project was carried out with a future sight of use age of robots while harvesting and so to expand the domain of deep learning in the field of agriculture i.e, towards harvesting. The solution proposed for the above problem statement was developed by taking tomato as an example and pre-grading to the vegetable unlike the traditional post grading methodology that is by grading the tomato into three grades with respect to the six ripening stages of a tomato and an additional classification the unfit category is added to promote waste management during harvest.
As working proof for the above idea a CNN model was developed by using tomatoes an example and the grading scale and the waste management system is developed on the basis of the six ripening stages of a tomato. Out of which the three distinguishable stages namely the mature green, turning stage, red ripe were used as the classification part and in order to perform waste management and additional classification called the unit called unfit was used (contains data on stem, leaves, flowers, dry twigs, sepals). A dataset of around 1838 images where each classification contains around 450 to 470 images. A CNN model with an accuracy of around 91.4% was developed.
Keywords:
Fruit/vegetable grading, Deep learning, Image processing, Food waste management, AI in harvesting, CNN.
