Vegetable Detection Using Tensorflow Object Detection API

Abstract

At present, object detection is one of the popular platforms in whole the world. Because of uses automation technology in most of the sectors, object detection is most important to know the accurate object of the machine. Realizing this concept, we made this project. This project will be detected seven different types of vegetables that are available in our county. It is implemented by using TensorFlow object detection API that is making use of OpenCV. This API makes it easy to detect our selected objects by using a pre-trained object detection model. A pretrained model simply means that it has been trained on another dataset. The model we have used in our project that is ssd_mobilenet_v2_coco. There are 2100 images in seven separate kind of vegetable used in our project. Using these datasets in our proposed model we are getting up to 99% accuracy based on types of vegetable.