摘要
为农作物叶面积实时快速地精准自动测量提供参考,采用可跨平台部署的开发平台为基础,应用面向对象编程语言Java,设计基于移动终端的农作物叶面积测量系统。以农作物叶面为研究对象,以数字图像处理技术与基于Android系统移动终端的开发技术相结合为创新设计,采用灰度识别算法和Ostu阈值识别算法,从包含叶片图片文件中像素点的个数为突破口,应用叶片所占的像素与标记点模块所占的像素点的个数进行比较,计算较为准确的农作物叶面积。结果表明:农作物叶面积测量系统测量的叶面积具有导入图片或照相、确定标记点、确定叶片区域以及计算叶片面积的功能,能对选择的叶片区域进行标定和叶面积识别,并对该区域叶片进行叶面积的精确计算。系统突破大田实际测量环境的局限性,用户可用手机进行叶面积测量,具有测量方式灵活、准确快速、操作简单及便于携带等优点。
The crop leaf area measurement system based on mobile terminal was designed by using the development platform of cross-platform deployment and object-oriented programming language Java to provide a reference for real-time,rapid and precision automatic measurement of crop leaf area.The crop leaf area can be accurately calculated by using grayscale and Ostu threshold recognizer and comparing the image elements and mark points occupied by leaves with the image elements occupied by modules based on digital picture processing technique and Android system mobile terminal development technique.Result:The crop leaf area measurement system with functions of import picture or photograph,deterministic mark point,deterministic leaf area and calculating leaf area can calibrate the selected leaf area,recognize leaf area and accurately calculate leaf area of the region.The system with the advantages of flexible measurement mode,accurate,rapid,easy operation and portability breaks through the limitations of actual measurement environment in field and can be installed on mobile phones to measurement of leaf area.
作者
关海鸥
王璐
马晓丹
GUAN Hai′ou;WANG Lu;MA Xiaodan(College of Electrical and Information,Heilongjiang,Bayi Agricultural University,DaQing,Heilongjiang 163319, China)
出处
《贵州农业科学》
CAS
2018年第4期125-130,共6页
Guizhou Agricultural Sciences
基金
大庆市科技计划项目"基于图像处理技术的大豆叶部病害识别关键技术研究"(Zd-2017-28)
国家自然科学基金项目"大豆冠层三维重建及光照时空分布模型研究"(31601220)
中国博士后科学基金资助项目"农村低压电网生物触电暂态特性分析及快速识别机理研究"(2016M591559)
黑龙江省自然科学基金项目"果树冠层光照时空分布模型研究"(QC2016031)
黑龙江省农垦总局科技攻关项目"马铃薯疫病光谱特性及诊断技术研究与应用"(HNK125A-08-03)
关键词
农作物
叶面积测量
移动终端
灰度识别
Ostu阈值
crop
leaf area measurement
mobile terminal
grayscale recognition
Ostu threshold