Abstract: Chalky rice detection is very important in human’s daily life.This paper presents a chalky rice detection method that is studied based on multi-color conversion and fuzzy threshold segmentation.First，a rice image is transformed twice by using a color model，and then，applying a fuzzy threshold segmentation algorithm，which has good performance for a bimodal gray-scale histogram，binaries the transformed image.As a result，the chalkiness in the chalky white rice can be accurately abstracted.This algorithm has greatly enhanced the accuracy of the detection.

Key words: chalky rice, color conversion, fuzzy threshold segmentation

摘要： 垩白米的检测在人们的日常生活中具有重大意义。提出了一种基于多重色彩转换和模糊阈值分割的垩白米检测方法。首先，对大米图像进行两次色彩模型的转换，然后根据模糊阈值分割在对具有双峰灰度图像中良好的分割效果对大米进行分割，使垩白米中的垩白信息准确地显现出来。该算法比传统检测算法在识别正确率上有很大的提高。

关键词: 垩白米, 色彩转换, 模糊阈值分割