基于支持向量机的作物预警设计及实现  

Design and Implementation of Crop EarlyWarning Based on Support VectorMachine

作　　者：黄怡宁 何金保[1] 任万川 韩玉静 

机构地区：[1]宁波工程学院,浙江宁波315211

出　　处：《大众科技》2020年第8期8-11,共4页Popular Science & Technology

基　　金：国家大学生创新训练项目(201911058002);宁波自然科学基金(2019A610096)。

摘　　要：在作物的成熟过程中往往会遇到农作物疾病、虫害。传统利用经验来人眼识别病虫害的效率低下并且容易将某些相似的疾病判别错误,文章采用基于“多分类”的支持向量机深度学习方法,对黄瓜叶片病害进行图像识别,对它们的纹理、颜色、形状进行提取与分析,先训练然后识别。实验表明,该方法识别的准确率高,在对于病虫害的预警的应用上有着很明显的效果。Crop diseases and insect pests are often encountered in the process of crop maturity.Traditional use of experience to identify diseases and insect pests is inefficient and it is easy to distinguish some similar diseases.In this paper,the deep learning method of support vector machine based on"multi classification"is used to identify cucumber leaf diseases.The texture,color and shape of cucumber leaf diseases are extracted and analyzed.First,training and then recognition.The experiment results show that the recognition accuracy of this method is high,and it has obvious effect on the application of early warning of diseases and insect pests.

关 键 词：病虫害诊断 图像识别 支持向量机 

分 类 号：TN1[电子电信—物理电子学]