摘要
病虫害的及时防治是提高农作物产量与品质不可或缺的重要环节。为了及时有效地防治病虫害,首先应实现病虫害的自动识别。为此,提出一种基于机器视觉的欧氏距离病虫害自动识别检测方法,旨在为病虫害自动防治喷雾机器人的研制奠定基础。欧氏距离方法是以颜色特征为基础,利用几何阈值选取和RGB空间特征的变换,实现病虫害的自动识别。试验结果表明,该方法能够较准确地识别虫害区域和提取目标作物,准确率达88.33%。
Pest prevention and treatment are important parts for improving the yield and quality of crops. Therefore, this paper proposes a method which uses the plant pests identification and detection methods of Euclidean Distance to detect the plant pests. It aims to lay the foundation for the development of pest control spray robot. The euclidean distance method is based on the color feature, using the geometric threshold selection and spatial characteristic of RGB transform to identify of pest automatically. The experimental results show that these two methods can correctly identify the disease area; the accuracy rate reaches 88.33%
出处
《农机化研究》
北大核心
2015年第1期205-208,222,共5页
Journal of Agricultural Mechanization Research
基金
内蒙古自然科学基金项目(2011MS0717)
关键词
机器视觉
欧氏距离
菜青虫害
自动识别
machine vision
euclidean distance
rape pests
automatic recognition ]
