稻谷品质的图像识别与快速检测

Image Dicerment and Rapid Detection for Rice Quality

设计开发了一套计算机图像处理系统RQES1.0 ,专门用于GB/T 17891- 1999中优质稻谷分级质量指标整精米率、垩白粒率、垩白度和粒型等参数的检测。该系统采用可视化界面 ,具备人机交互功能。与现行的国际检测法相比 ,具有客观性强、准确度高、快速便捷等特点。研究结果表明 ,利用机器视觉代替现行的目测检验是行之有效的评价方法 ,对全面提升我国优质稻谷检测水平 ,推进优质稻谷的生产与开发具有重要的作用。

A digital Image processing system RQES1.0 has been developed for the detection of quality indices in GB/T 17891-1999.The indices include ratio of whole white rice kernel,number of rice,ratio of kernel length and width,chalkiness,ets.Visual interface is adopted with man-machine interaction in the system,which showed ideal characteristics of objectiveness,accuracy and quickness.It was revealed that machine vision could be one of the most effective ways for the substitution of the current subjective detection method,and should be helpful for promoting the level of national rice quality evaluation.