摘要
作物生长期自动识别是精准农业支持技术的核心部分之一。为了实时准确地识别玉米不同的生长期,获取玉米生长信息,提出一种基于深度卷积神经网络特征提取的玉米生长期识别方法。首先,对拍摄的玉米图像进行预处理以滤除噪声;然后通过深度卷积神经网络提取玉米不同生长期的特征,再结合粒子群优化算法优化SVM参数,构造多级支持向量机模型对玉米生长期分类识别。实验结果表明,结合深度学习和SVM分类建模的方法能自动提取玉米的关键特征并能有效识别不同生长期。
Automatic recognition of crop growth techndogy accurately identi{y the di period fferent is one of the core components of precision agriculture support growth periods of maize and obtain maize growth information in real time, a new automatic growth identification method based on Firstly, the maize images were preprocessed to filter out the noise. maize were extracted by deep convolutional neural network deep convolution neural network was proposed. Then, the features of different growth periods of with particle swarm optimization algorithm to optimize SVM parameters, and a multi-level support vector machine model was constructed for maize growing category. A multi-level support vector machine model was constructed to identify the growth period of maize. The experimental results show that the combination of deep learning and SVM classification method can automatically extract the key characteristics of maize and can effectively identify different growth period.
作者
张芸德
刘蓉
刘明
龚永丽
Zhang Yunde;Liu Rong;Liu Ming;Gong Yongli(College of Physical Science and Technology,Central China Normal University,Wuhan 430079,China;School of Computer,Central China Normal University,Wuhan 430079,China)
出处
《电子测量技术》
2018年第16期79-84,共6页
Electronic Measurement Technology
基金
湖北省技术创新专项重点项目(2017AFB188)资助
关键词
卷积神经网络
深度学习
生长期识别
特征提取
支持向量机
convoiutional neural network（CNN）
deep learning
growth period recognition
feature extraction
support vector machine（SVM）
