果树冠幅的检测机理研究  

Detection mechanism for fruiter crown diameter

作　　者：张富贵[1,2] 洪添胜[3] 肖磊[3] 陈海生[3] 陆永超[3] 

机构地区：[1]贵州大学机械工程学院 [2]贵州大学机械工程学院,贵阳550003 [3]华南农业大学南方农业机械与装备关键技术省部共建教育部重点实验室

出　　处：《农业工程学报》2008年第4期25-29,共5页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家自然科学基金资助项目(30270764,60601009);贵州省科技基金项目(黔科合J[2007]2185)

摘　　要：为实现果树仿形精确喷雾,需要检测果树冠幅,在分析果树冠幅检测机理的基础上,提出了隔行计算相对冠幅的方法。根据相对冠幅和成像距离之间的关联作用,建立了数据融合的系统模型。采用BP人工神经网络进行数据融合,并阐述了具体的数据融合处理过程。试验结果证明,相对冠幅和成像距离之间采用BP人工神经网络进行数据融合可行,通过数据融合处理,消除了成像距离对相对冠幅值检测的影响,融合值与相应的理论值之间的最大相对误差为1.14%。It is necessary to detect the fruiter crown diameter for fruiter profile modeling of precision spray. The method of interlaced calculation and the theory of relative crown diameter （RCD） were developed based on the detection mechanism for fruiter crown diameter. The system model about data fusion was established by the relationship between RCD and imaging distance. BP artificial neural network was used for data fusion and the data fusion process was explained. The experimentation results prove that it is feasible to use BP artificial neural network for data fusion between RCD and imaging distance. The influence of imaging distance on detecting RCD was eliminated by data fusion. The maximum fractional error between fusion value and theoretic value is 1.14%.

关 键 词：果树冠幅 检测机理 成像距离 数据融合 BP人工神经网络 

分 类 号：S491[农业科学—植物保护] S436.6