嵌入式水稻氮素营养检测仪的研究与设计  

Research and design of the embedded detector for rice nitrogen nutrition

作　　者：孙铁波[1] 王卫兵[1] 刘春月[2] 

机构地区：[1]江苏食品药品职业技术学院,江苏淮安223003 [2]淮安信息职业技术学院,江苏淮安223003

出　　处：《中国农机化学报》2015年第3期257-261,285,共6页Journal of Chinese Agricultural Mechanization

基　　金：江苏省苏北科技发展计划项目(BN2012018);淮安市科技支撑计划(农业)项目(HGZ2011016)

摘　　要：针对水稻氮素营养实时检测的问题,设计了一种基于嵌入式技术的利用水稻叶片颜色特征进行氮素营养水平检测的系统。首先提取不同氮素营养水平下的水稻样本叶片图像的RGB颜色特征,并对各颜色特征值与水稻全株含氮量进行相关性分析。其中颜色特征G和2G-R-B分量的相关性高于其它,分别为-0.761 1和-0.742 1,进一步对叶片不同位置的G和2G-R-B分量与全株含氮量进行相关性分析得出叶片中部2G-R-B分量与全株含氮量间相关性最高,达0.765 6。以此为依据,结合Open CV图像处理技术和Qt4.5跨平台用户界面框架技术,最终实现系统的嵌入式设计。针对设定样本检测结果表明,系统对各营养水平稻株识别准确率均超过90%,且具有良好实时性与便携性,能够满足水稻氮素营养水平实时检测需要。A detector using the color character of the rice leaves based on embedded technology was designed to solve the problem of real-time detection for rice nitrogen nutrition. First, the RGB color feature of leaves image on different level of nitrogen nutrition was extracted, then the correlation be- tween eigenvalues from different color and total nitrogen content was analyzed. The correlation index of the quantity of color characteristics G and 2G- R-B is -0. 7611 and -0. 7421. The correlation between the quantity of color characteristics G and 2G-R-B on different position of the leaves with total ni- trogen content was analyzed, which showed that the correlation index of the quantity of color characteristics 2G-R-B on the middle part with total nitro- gen content is the highest, reached 0. 7656. Furthermore, based on the conclusion above, combing the image processing technology of OpenCV and cross platform framework technology of Qt4.5, the design of embedded system was completed. Test results according to sample indicated that the accu- racy of the identification on rice nutrition content is over 90%, and the embedded detector has good real-time performance, is portable and able to meet the requirement of real-time detection for rice nitrogen nutrition.

关 键 词：水稻 氮素营养 嵌入式 颜色特征 相关性 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置]