Crop Classification on Deep Learning

In this paper, we combine image processing and deep learning to achieve crop classification recognition. Crop detection is based on deep learning of convolutional neural networks to capture image features and determine various types of crops. Through the steps of image pre-processing, feature extraction, data augmentation, identification and classification, the crop images are identified and classified. The proposed crop species recognition system is divided into three major parts. We use image pre-processing then cut down the crop feature images, and use Efficientnet-B0 as the backbone for crop classification in this paper. The accuracy of 99.3% is obtained in crop classification recognition.