介绍了基于ImageJ软件处理数字图像辅助测量子粒大小的新方法。分别用图像处理和游标卡尺两种方法测量了70份玉米材料的子粒大小，图像处理法测量的相对误差均小于2%。对粒长、粒宽和粒厚成组数据进行t测验，两种测量方法间差异不显著。相关性分析表明，两种测量方法间呈极显著的线性相关。与游标卡尺法相比，ImageJ软件处理图像法更方便快速，可用于玉米等作物种子大小的实际测量工作。

Kernel size is one of the important economically characters of maize and traditional vernier caliper measurement is time-consuming. In this study, a new method was developped to measure seed size by ImageJ software. Totally, 70 maize materials were used and ImageJ method and vernier caliper method were employed to measure seed size, respectively. Relative error for the two methodes were less than 2%. Based on t-test for the data of seed length, widths and thicknesses under the two measurements, t_0.05(138) value are less than 1.977 3, indicating that there was no significant difference between the the two methods. According to correlation analysis, the correlative coefficients of the seed length, widths and thicknesses measured by the two methods were 0.994 2, 0.994 5 and 0.994 5, respectively, reaching to highly significant levels. As compared to the vernier method, the ImageJ method could be more convenience and could be used for measuring not only for maize seeds but also for other crops.

中国农业大学西部地区高校博士研究生访学项目(F-0704)、教育部创新团队项目