Bangladeshi Fresh and Rotten Vegetables Detection Using CNN

Abstract

Vegetables cultivation is part of our Bangladeshi Farming Industry. Vegetables also helps human to get lot of vitamins which keep body energetic and fit. Furthermore, in our nation almost every people get benefits from agriculture. Hence, it’s important to maintain vegetables freshness. Unhealthy vegetables can harm human body. Also, farmers and retailers or vendors can get losses if they buy rotten or spoiled vegetable. For this reason, I came up with an idea to distinguish all the rotten & fresh vegetables. I suggested a model that can identify the fresh-rotten veggies. It’s nearly impossible for humans to do this difficult task as there can be thousands of vegetables. To distinguish all of them will be challenging. Our model will classify the veggies as Fresh & Rotten. To do this task, I used CNN model. It will classify our data into fresh- rotten after giving input of vegetables image. Then I compared our model with SVM & KNN techniques too. Our proposed model performed better than these two algorithms. CNN model obtained accuracy of 93%. I’ll work with other types of vegetables & more images with other techniques to get better result in future. This research will be helpful for our country.