基于数字图像的马铃薯单株叶面积预测技术研究  

The Research on the Prediction Technology of the Area of Potato Leaf per Plant based on the Digital Image

作　　者：马文婷[1] 马达灵[1] 王晓燕[1] MA Wen-ting;MA Da-ling;WANG Xiao-yan(Dept. of Computing Science, Jining normal college, Inner Mongolia Wulanchabu 012000)

机构地区：[1]集宁师范学院计算机系,内蒙古乌兰察布012000

出　　处：《集宁师范学院学报》2017年第3期20-24,共5页Journal of Jining Normal University

基　　金：集宁师范学院校级课题"基于图像识别的马铃薯叶面积预测"(项目编号:jsky2016046)

摘　　要：本研究的目的是建立马铃薯单株叶面积的预测模型,为马铃薯叶面积无损、快捷测量提供技术支持。通过分时期采集植株数字图像和同步手工测量单株叶面积的方法,获得绿叶图像的总像素、单株实测叶面积。通过像素法计算绿叶的覆盖面积,利用覆盖面积与叶面积间的内在关系,建立基于数字图像的马铃薯单株叶面积预测模型,y=1821.3ln(x)-9567.9 R^2=0.56**(Y指单株马铃薯的叶面积,x表示叶片覆盖面积)。经验证,该模型预测精度较高。The purpose of this study is to establish the prediction model of the area of potato leaf per plant to offer thetechnique about the nondestructive and fast measurement of potato leaf area.By acquiring digital image in different periodsand manually measuring the leaf area per plant synchronously,the total leaf-image pixel of per plant and the measured leafarea of per plant are obtained.Covered area of the green leaves is calculated by the pixel method.The prediction model ofpotato leaf area for per plant is obtained by analyzing the relationship between covered area and leaf area.The equation wasy=1821.3ln(x)-9567.9R2=0.56**(y is the potato leaf area of per plant,x is its covered area).The accuracy of thisprediction model is higher by verification.

关 键 词：数字图像 像素法 马铃薯 单株叶面积 预测模型 

分 类 号：S126[农业科学—农业基础科学]