作者：张亚静,邓烈,李民赞,赵瑞娇,何绍兰,易时来
作者单位：中国农业大学现代精细农业系统集成研究教育部重点实验室,中国农业科学院柑桔研究所
摘要：利用机器视觉技术可以快速、无损预测柑橘产量。采集了10幅生长中的柑橘果树照片,同时测量了每棵果树的柑橘产量。基于RGB颜色模型,对柑橘图像进行分割,柑橘与背景的分割条件为(R-B>100)且(R>G)。通过提取柑橘个数、柑橘总周长、柑橘总面积3个特征参数,分析了特征参数和柑橘单株产量之间的关系。实验证明,经过图像分析后得出的柑橘数与柑橘单株产量之间的相关系数最高,达到0.97,说明了利用图像分析方法预测柑橘产量具有良好的应用前景。
关键词：柑橘,机器视觉,图像分割,颜色特征,产量
Title: Estimation of Citrus Yield Based on Image Processing
Authors: Zhang Ya-jing,Deng Lie,Li Min-zan,Zhao Rui-jiao,He Shao-lan, Yi Shi-lai
Abstract: The estimation of citrus yield is important to citrus precision management.Fast non-destroyed measure could be achieved using machine vision technology.Ten images of citrus were used in this experiment and yield of each tree was measured in advance.The citrus image was segmented based on the RGB color model.The segment condition for citrus and background is:(R-B>100) &(R>G).The total number,the total perimeter and total area of citrus were computed from the segmented image.The relationship between each image parameter and citrus yield was analysis.The result of experiment shows that the correlation coefficient between the total citrus number and the yield of citrus is 0.97 and indicates that it is has good prospect in the future.
Key words: Citrus； Machine vision； Image segmentation； Color feature； Yield；
文献注录：张亚静,邓烈,李民赞,赵瑞娇,何绍兰,易时来. 基于图像处理的柑橘测产方法 [J]. 农业机械学报. 2009, 0S1: 97-99.
基金: 国家“863”高技术研究发展计划资助项目(2006AA10Z255);“十一五”国家科技支撑计划资助项目(2007BAD47B02-3)
报/刊名：《农业机械学报》，发表于2009 / 第 0S1 期。
文献类型： [J] （文献级别：核心刊物）
页码：第 97-99 页 / 共3 页