基于图像的大米品质检测的准确性问题的研究.由于垩白米与普通优质大米在像素灰度上差异极小,大米中经常混淆垩白米,传统的基于灰度的大米品质算法不能很好的区分混有垩白米部分的大米灰度差异,不能准确鉴定大米品质的问题.为了解决这个问题,提出了基于大米像素组织的大米品质鉴定方法.通过直线平滑分割、大米像素组织点性质识别的方法,克服了传统方法像素混淆后无法识别的弊端.实验表明,组织特征方法能够准确鉴定混合了垩白米的大米品质,并取得了比较好的效果.%Research the accurate detection of rice quality based on rice images. White rice often are confused with chalky rice, because of the tiny differences in pixel gray between the chalky rice han high quality rice, and the traditional quality detection method based on the grayscale cannot distinguish the gray differences of the rice mixed with the chalky rice very well. The quality detection method based on rice images was proposed in this paper, in order to solve the problem above. Through the linear smoothing segmentation and the method of nature recognition for the rice pixels organization points, it can overcome the disadvantages that traditional method cannot distinguish pixels after confusion. The experiments show that the method can identify the quality of the rice accurately after mixed with chalky rice, and has received good effects.
展开▼
