机器学习在农作物品种识别中的应用研究进展  

The Application of Machine Learning in Crop Variety Recognition: A Review

作　　者：阳灵燕 张红燕[1] 陈玉峰[1] 刘亚文 Yang Lingyan;Zhang Hongyan;Chen Yufeng;Liu Yawen(College of Information and Intelligent Science and Technology,Hunan Agricultural University,Changsha 410128)

机构地区：[1]湖南农业大学信息与智能科学技术学院,长沙410128

出　　处：《中国农学通报》2020年第30期158-164,共7页Chinese Agricultural Science Bulletin

基　　金：国家重点研发计划“粮食丰产增效科技创新”重点专项(2017YFD0301506);长沙市工业科技特派员项目“农村农业信息化共性关键技术研究”(201845)。

摘　　要：机器学习在图像识别领域的成功应用,为农作物品种的自动识别提供了一种新的思路。为了全面了解机器学习在农作物品种识别中的应用现状,把握农作物品种识别的发展趋势和研究方向,本文归纳了农作物图像的常用获取方法,分析了光谱图像和RGB图像结合机器学习方法识别农作物品种的研究现状。基于RGB图像进行农作物品种识别研究起步较早,图像获取成本较低,识别率一般;基于高光谱图像进行农作物品种识别研究近年来发展迅速,识别精度较高,但图像获取成本较高,且易受环境因素影响。通过总结,指出了农作物品种识别研究中存在的问题,认为深度学习在农作物品种自动识别上具有广泛的应用前景.The successful application of machine learning in the field of image recognition provides a new idea for the automatic recognition of crop varieties. In order to fully understand the application status of machine learning in crop variety recognition and grasp the development trend and research direction of crop variety recognition, this paper summarized the common acquisition methods of crop image, and analyzed the research status of crop varieties recognition based on spectral image and RGB image with machine learning methods.The research on crop variety recognition based on RGB images started earlier, the cost of image acquisition is relatively low, and the recognition rate is general;the research on crop variety recognition based on hyperspectral images has developed rapidly in recent years, and the recognition accuracy is relatively high, but the cost of image acquisition is higher and easily affected by environmental factors. Through the summary, the existing problems in crop variety recognition research are pointed out, and it is considered that deep learning has a wide application prospect in the automatic recognition of crop varieties.

关 键 词：机器学习 光谱图像 RGB图像 农作物品种识别 图像分类 

分 类 号：S3-3[农业科学—农艺学]