| 数字图像分析的椪柑可溶性固形物含量检测 |
| |
| 引用本文： | 易时来,邓烈,何绍兰,张浩,郑永强.数字图像分析的椪柑可溶性固形物含量检测[J].农机化研究,2012,34(2):116-121. |
| |
| 作者姓名： | 易时来 邓烈 何绍兰 张浩 郑永强 |
| |
| 作者单位： | 1. 西南大学中国农业科学院柑桔研究所,重庆,400712
2. 中国科学院亚热带农业生态研究所,长沙,410125 |
| |
| 基金项目： | 重庆市自然科学基金项目，中央高校基本科研业务费专项资金项目，科技部科技人员服务企业专项资金项目 |
| |
| 摘 要： | 为了探索建立基于数字图像处理技术的柑桔果实品质非损伤探测技术,采用多光谱照相机MS3100获取不同成熟度椪柑果实样品的数码图像R,G,B通道颜色信息,结合实验室果实可溶性固形物含量(SSC,Solublesolids content)化学分析值,进行了椪柑果实SSC与颜色信息相关性分析和检测模型研究。结果表明,椪柑果实图像的5个颜色参数R,G-R,B-R,G/R和(G-R)/(G+R)与果实SSC值之间具有较好的相关性,校正模型相关系数均在0.83以上;模型验证结果表明,B-R值与果实SSC值所建立的模型预测效果最佳,决定系数为R2=0.651,是椪柑果实SSC检测的最佳图像颜色参数,一元二次方程SSC=-0.0001(B-R)2-0.0219(B-R)+9.601为其最优检测模型;神经网络模型可以容纳更多的相关波段参与柑桔SSC含量的估算,实测值与预测值的相关系数高于其他模型,而均方根误差(RMSE,Root mean square error)低于其他模型,表明利用计算机图像技术进行柑桔果实SSC的检测是可行的。
|
| 关 键 词： | 数字图像 椪柑 可溶性固形物含量 检测 |
Detection of Soluble Solids Content of Ponkan Using Digital Photography Analysis |
| |
| Yi Shilai,Deng Lie,He Shaolan,Zhang Hao,Zheng Yongqiang.Detection of Soluble Solids Content of Ponkan Using Digital Photography Analysis[J].Journal of Agricultural Mechanization Research,2012,34(2):116-121. |
| |
| Authors: | Yi Shilai Deng Lie He Shaolan Zhang Hao Zheng Yongqiang |
| |
| Institution: | 1(1.Citrus Research Institute,Southwest University-Chinese Academy of Agricultural Sciences,Chongqing 400712,China;2.Institute of Subtropical Agriculture,Chinese Academy of Sciences,Changsha 410125,China) |
| |
| Abstract: | Research on the detection model of soluble solids content of Ponkan(Citrus reticulata Blanco cv.Taiwan Ponkan) fruit by using computer vision technology and chemical analysis methods.Study on the correlation of R,G,B color channel image parameter information captured by MS3100 multi-spectral camera and the citrus fruits soluble solids content using chemical analysis in laboratory on different periods of collecting Ponkan samples.The experiment shows a good correlation existed in the five-image-color-parameters of R,G-R,B-R,G / R and(G-R) /(G+R) value and SSC,and with average correlation coefficient 0.83 of monitoring model.The validation results of monitoring model reveals the effective prediction model should be created by the image color parameter values of B-R and fruit SSC,whose determinational coefficient(R2) is 0.651.The image color parameters of B-R is the best in the monitoring of fruit SSC by comprehensive comparisons.The monitoring model is a quadratic equation and SSC=-0.0001(B-R)2-0.0219(B-R) +9.601.The neural network model can accommodate more related bands to participate the estimation of citrus SSC,and the correlation coefficient between the measured and predicted values is higher than the other models,while the root mean square error(RMSE) is lower than other models,indicated that it is feasible to detection citrus fruit SSC using computer graphics technology. |
| |
| Keywords: | digital photography ponkan soluble solids content(SSC) detection |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
