Estimation Crop types and Growth Stages With Hierarchical Classification Models

Abstract

In smart agriculture, a technology that automatically classifies crop types and growth stages for each crop is required in case of hybridization. In this paper, we propose a model that classifies crop types and growth stages together with crop images. Based on VGG-16 pre-trained with ImageNet, a hierarchical class structure is applied to design a model to classify crop types and growth stages. Image data of 15 types of crops and 13 types of growth stages are used. The test result showed an accuracy about 90.1%.