Prediction of Potato Yield Using Unmanned Aerial Vehicle and Convolutional Neural Network

Abstract

We used UAV as a remote sensing device in potato cultivation and assessed yield prediction by CNN with aerial image and yield data. The experiment was performed in a potato field with six fertilization conditions, aerial images were acquired during the flowering period, and growth and yield were investigated. Yield prediction models were constructed by regression analysis and image regression method by CNN. In the aerial image, the difference in coverage was confirmed depending on the treatment section, and the treatment section difference occurred in the plant height as the growth progressed. Regarding yield prediction, the yield prediction model by image regression method by CNN was able to predict the potato yield with high accuracy than regression analysis.