谷物外观品质检测方法的研究  

Research on the Detection Method of Cereal Grains Appearance Quality

作　　者：王一丁[1] 李霁阳[1] 

机构地区：[1]北方工业大学电子信息工程学院,北京100144

出　　处：《计算机测量与控制》2017年第2期35-38,共4页Computer Measurement &Control

摘　　要：随着我国科学技术的发展,有许多传统的人力检测方法可以被机器视觉和图像处理代替;而且随着计算机的性价比越来越高,让机器视觉融入我们的生活也变为可能;因此,通过处理谷物外观图像来对一批谷物的质量进行评估,具有较高的理论价值和实际意义;利用了谷物长宽比的特点,可以有效地判断谷物的粒型和完整性,这种方法不仅适用于大米,对红豆、黑米、绿豆、燕麦、高粱米、红米等谷物也有着良好的识别效果;并且文章对流域算法进行了一定的改进,使流域分割算法可以应用在谷物外观品质检测中;实验结果表明,该算法及程序上的设计是可靠的、高效率的,可以利用机器视觉代替人眼检测。Using image processing technology to replace manual testing is an inevitable trend of automatic classification development. With the considerably increasing of the computers high cost-effective, it is possible to make practical application of machine vision technology. Accordingly, it has high theoretical value and practical significance by processing the cereal grain image to evaluate the quality of the cereal grains. In this essay we use the characteristics of the aspect ratio of cereal grain. It can effectively determine the size and the shape of grain and integrity. This method is not applicable only for the rice, red beans, black rice, mung beans, oats, sorghumrice, red rice, but other ce- real grains also have good reeognizahle results. In addition, there are some improvements of watershed algorithm in this essay, which can he applied in cereal grains appearance quality checking. As the result, the algorithm and the design of program are effective and reliable, which proves that machine vision can be used instead of human vision.

关 键 词：数字图像处理 机器视觉 谷物外观 流域分割 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]