基于深度学习的玉米抽雄期判识

Maize tasseling period recognition based on deep learning

农作物自动化观测是农业现代化和自动化进展的重要标志和不可缺少的一部分。目前作物生长自动化观测主要通过获取作物生长图像,进而对图像进行处理分析来获取作物生长特征等信息,利用传统的图像分割、特征点检测等方法检测作物的生长特征误差较大。近年来,深度学习迅速发展,且在多个领域得到广泛应用。在ImageNet数据集和大量的玉米作物图像基础上,对深度学习中常用的CNN进行训练和微调,充分利用迁移学习的优势,得到识别模型,对玉米雄穗进行识别,进而对玉米抽雄期进行判识。实验证明深度学习方法在作物生长特征识别上有良好的效果,比传统方法有显著提高。

Automatic observation of crop is one of the core components of modern agriculture and automation progress. At present, automatic crop growth observation is mainly by obtaining crop growth images and then processing and analyzing the images to obtain growth characteristics information of crop. Traditional methods such as image segmentation and feature point detection have obvious error when used to detect crop growth characteristic. In recent years, deep learning has developed rapidly and has been widely used in many fields. In this paper, based on ImageNet data set and a large number of images of corn crops, the CNN which frequently used in deep learning is trained and fine-tuned. The advantage of transfer learning is fully utilized to obtain the recognition model, to identify the male ear of maize and then to judge the tasseling period of maize. The experimental results show that the deep learning method has a good effect on the recognition of crop growth characteristics, and is significantly improved than the traditional method.