【作者】 尚艳芬；

【导师】 侯彩云；

【作者基本信息】 中国农业大学 ， 农产品加工及贮藏工程， 2004， 硕士

【摘要】 现行稻米品质的标准中，多数指标的检测尚处于采用人工感官评定或常规的理化方法检测阶段。由于方法的主观性强、准确性低、可重复性差、检测时间长，使标准的权威性和合理性大打折扣，并在一定程度上制约了稻米产业的发展。针对这一问题，本研究应用计算机图像处理技术和RGB色度学理论，研发了一套稻米品质评价系统(RQS1.0)。系统采用可视化界面，具备人机交互功能。在该系统中探讨了黄粒米、整精米率、稻米加工精度等指标自动检测的可能性。研究中，通过分析整精米与碎米粒形的差异，提出一种自动检测整精米率的方法；通过分析黄粒米与正常米的色度特征差异，提出了一种自动识别黄粒米的方法；通过分析稻米皮层与胚乳的色度差异，探讨了自动检测稻米加工精度的方法，取得了较好的结果。此外，RQS1.0还实现了垩白粒率、垩白度、长宽比等外观品质以及直链淀粉含量这一理化指标的自动检测的功能。研究结果表明，利用该系统进行上述指标的检测具有可行性，并在一定程度上弥补了现行国标方法的不足，为贯彻执行稻米收购的优质优价政策奠定了基础。更多还原

【Abstract】 Among the current indices of rice standard, the majority of them still stay evaluated by manual method or general chemical and physical means, which brings disadvantages such as strong subjectivity, low accuracy, poor repetitiveness, long time-consumption, and so on. These discount the authority and rationality of standard and restrict the development of rice industry to a certain extent. To compensate the weaknesses, an assessment system of rice quality (RQS1.0) is developed by employing computer digital image processing technology and RGB chromaticity theory. Visual interface is adopted with man-machine interaction in it. With the system, a study is made on the feasibility of automatic detection of yellow-colored rice, head rice and the milling degree of rice. After an analysis of the shape differences between head rice and broken rice, a method for automatic detection of head rice is advanced. Based on a research of the different color traits of yellow-colored rice and general rice, an approach of automatic detection of yellow-colored is presented. The method of automatically detecting the accuracy of rice processing is studied according to the diverse colors of bran peel and albumen of rice and a better result is obtained. Besides, RQ1.0 makes it possible to employ the automatic-detection indices of physical and chemical characteristics as the rate of poor-quality white rice and its seriousness, outer features like the proportion of length and width, and amylase content. The findings show it is feasible that using the system for a detection of the above indices which improve the deficiency of the present national detection method. The system builds a flat road for the enforcement of "good price, high quality" policy of rice purchase.更多还原