摘要
为了实现田间甘蔗出苗期的自动观测,本文基于2015~2016年广西柳城县甘蔗农业气象自动观测站实时采集的甘蔗生长图像,结合人工平行观测资料,分析甘蔗植株图像的突变性和渐变性特征,利用G/B、G/R特征指标和连通域标记方法分割植株图像及消除杂草,并基于植株整体性检测实现甘蔗出苗期自动识别。利用识别方法对2016~2018年广西扶绥县甘蔗出苗期进行检验,结果发现基于图像特征自动识别甘蔗出苗期是可行的;同时精度分析结果表明,与人工平行观测结果相比,出苗期起始和结束的时间3年分别相差6、5和3天,总体识别效果较理想。
In order to realize the automatic observation of sugarcane emergence stage in the field,the mutation and gradual change characteristics of sugarcane plant images were analyzed based on the sugarcane growth images collected from the sugarcane agrometeorological automatic observation station in Liucheng County,Guangxi Province from 2015 to 2016 and the artificial parallel observation data.Using G/B,G/R characteristic index and connected domain labeling method to segment plant image and eliminate weeds,and the sugarcane seedling emergence was realized automatically based on the plant integrity detection distinguish.The identification method was used to test the emergence period of sugarcane in Fusui County of Guangxi Province from 2016 to 2018,and the results showed that it was feasible to automatically identify the sugarcane seedling stage based on image features.Meanwhile,the accuracy analysis results showed that compared with the artificial parallel observation results,the starting and ending time of the emergence period were different by 6,5 and 3 days during three years,and the overall recognition effect was better.
作者
刘志平
匡昭敏
马瑞升
吴炫柯
王达
李莉
朱辛华
LIU Zhi-ping;KUANG Zhao-min;MA Rui-sheng;WU Xuan-ke;WANG Da;LI Li;ZHU Xin-hua(Guangxi Meteorological Sciences Institute/Sugarcane Meteorological Service Center,Nanning,Guangxi 530022;Liuzhou Meteorological Service,Liuzhou,Guangxi 545003;Fusui Meteorological Service,Chongzuo,Guangxi 532100;Meteorological Disasters Defense Technology Center of Guangxi,Nanning,Guangxi 530022)
出处
《甘蔗糖业》
2020年第5期41-46,共6页
Sugarcane and Canesugar
基金
广西自然科学基金项目“甘蔗主要生育期自动观测关键技术研究”(2017GXNSFBA198153)
公益性行业(气象)科研专项重点项目“蔗糖产量预测及气象灾害监测评估技术研究”(GYHY20140406030)
国家重点研发计划项目课题“热带与特色林果气象灾害监测预警技术与业务平台”(2019YFD1002203)。
关键词
甘蔗
出苗期
图像特征
覆盖度
自动识别
Sugarcane
Emergence stage
Image features
Coverage
Automatic recognition
作者简介
刘志平(1982-),男,工程师,主要从事于农业气象自动化观测研究与信息化应用,E-mail:liuzp02@163.com;通讯作者:匡昭敏(1968-),女,博士,高级工程师,主要从事农业气象研究与应用。
