Abstract: Rate of broken rice is an important factor in the measuring of appearance quality of rice.An automatic detection method of broken rice rate is obtained based on morphology principle，digital image processing and analyzing.Based on national standards，a detection method is developed which eliminates the noise of image and the bad illumination condition by morphology open operation，uses an improved Otsu algorithm to segment the image，and calculates rice number and area by region labeling algorithm.The experiment results show that the method is efficient and effective.

摘要： 碎米率是大米品质检测的一个重要因素。通过运用形态学理论及数字图像处理与分析技术来获取自动检测碎米率的方法。依照国家标准，提出了利用形态学开运算校正光照不均和去除图像中的噪声，运用改进的最大类间方差法来自动确定图像分割阈值，使用区域标记算法统计大米数量和面积。实验结果表明：该方法可以提高识别碎米的准确度和效率。