基于图像处理的玉米叶片姿态测量方法研究  

Measurement Method of Maize Leaf Posture Based on Image Processing

作　　者：邓立苗[1] 于仁师[1] 马文杰[2] 

机构地区：[1]青岛农业大学理学与信息科学学院,山东青岛266109 [2]青岛农业大学动漫与传媒学院,山东青岛266109

出　　处：《河南农业科学》2014年第9期168-172,共5页Journal of Henan Agricultural Sciences

基　　金：山东省自然科学基金项目(ZR2009GM006);青岛市公共领域科技支撑计划项目(10-3-3-17-nsh)

摘　　要：为实现玉米叶片姿态的自动测量,针对玉米叶片形态特征,提出了基于图像处理的玉米叶片姿态自动测量方法。首先采用迭代方法对叶片图像细化,提取骨架图像上的各像素点,然后采用最大距离法计算叶片的弯曲度,通过测量叶片最高点与叶尖的垂直距离计算下垂度,最后对2种测量方法进行测试和分析。结果表明,所提出的叶片姿态测量的2种方法对样本叶片的测试准确率分别为92.4%和88.2%,能够有效描述和量化不同玉米品种的叶片弯曲程度,方法简单易行,且测量过程不需要标度校准。In order to realize the automatic measurement of maize leaf posture,based on the image processing technology,the automatic measuring methods of maize leaf posture were proposed and accomplished by following steps.Firstly,the image segmentation was adopted to separate the leaf from background.Secondly,an iterative method was used to refine the image and obtain the skeleton image.Thirdly,two methods were proposed to describe the posture of the maize leaves, of which,the curvature degree was calculated by maximum distance method,and the sag degree was measured by computing the vertical distance between the peak and the endpoint of leaf.At last,the two measurement methods were compared from the aspects of DUS （distinctness, uniformity and stability ） test and sample test.The results showed that the two measuring methods established in this paper were feasible and easy to realize,with the testing accuracy of 92.4% and 88.2%,respectively,which could effectively describe and quantify the curvature degree of maize leaves.Futhermore,the two methods need no calibration.

关 键 词：玉米叶片姿态 图像处理 弯曲度 下垂度 DUS测试 

分 类 号：S513[农业科学—作物学]