Fruit Detection using Deep Convolutional Neur...

Fruit Detection using Deep Convolutional Neural Network (DCNN)

Author Name : Amol Wankhade, Shrihari Sirsikar, Rushabh Pundkar, Prof. Vishal Meshram

ABSTRACT

Fruit classification has recently become a focus of research. Fruit detection was achieved using a convolution neural network (CNN) technique. The goal is to develop a fruit identification system that is accurate, rapid, and reliable, as well as a fundamental component of an autonomous agricultural robotic platform for fruit yield estimation and automated harvesting.Thisresearch looks at different fruit detecting methods. We used a high-quality, fruit-containing image dataset to train a neural network to recognize fruits in this study. A dataset with 300 images of India's top fruits divided into six classes. This method, in addition to being more accurate, is also significantly faster to implement for new fruits since it uses bounding box annotation rather than pixel-level annotation. The model is retrained to detect six different fruits, and the entire procedure takes three hours per fruit to annotate and train the new model.

Keywords: CNN, Fruit recognition, Dataset, Annotation