自动阈值法在生长环境茄子图像分割中的应用

在果蔬采摘机器人系统中,果实目标的智能化识别和定位问题一直是没有解决的问题。为此,研究生长环境中茄子图像的分割算法问题。通过对生长环境中茄子果实与周围环境颜色特征的试验与统计,得出了EXG颜色因子对于茄子果实分割最为有利的结论。采用自动阈值法对EXG灰度图像进行了分割。选择图像灰度的均值作为初始阈值的估算值,然后连续不断的进行迭代,直到取得满意的结果。试验表明,分割效率分别大于81%,平均用时为0.116s,能够满足果蔬采摘机器人对视觉系统的要求。

Application of Auto-thresholding Selection for eggplant Image Segmentation in the growth environment

The object's recognition and orientation are not solved for the fruit and vegetable picking robot.Image segmentation is studied for eggplant in the growth environment.With the experiment of statistic color characters,the EXG color mode is most available to segment the eggplants.The automatic threshold are adopted to segment the EXG gray images.The average value is regard as guesstimate of initialization threshold and iterative operation progress incessantly through to acquire satisfactory result.Experiments explain that the segmentation efficiency preponderate over 81% and the time consumed is 0.116s,meet the demand of the automatic sprayer.