基于背景板比例尺的玉米果穗图像特征测量

玉米果穗图像特征的精确测量是实现玉米自动考种的关键技术，但玉米背景图像的复杂性，难以快速、准确测量出玉米的几何特征（玉米果穗的总长度、宽度、体积、秃尖等）。该文根据玉米灰度图像的分布特点，构造了三区域图像增强模型，采用形态学图像处理方法对增强后的图像进行分割，计算出背景板比例尺，以此为参考实现了精确计算玉米的几何参数。文中对计算结果进行了误差分析，表明该文提出的基于背景板比例尺的图像处理算法能够准确地测量出玉米图像中的几何特征值，准确率可达到96.01%，可实现玉米自动考种，替代传统的玉米手工考种，对于玉

Image features measurement of maize ear based on background plate scale

Geometric features measurement of maize ear is the key technology of maize cultivar automatic analysis. But it is difficult to measure the geometric features(total length, width, bald, volume, etc) of maize quickly and accurately for the complexity of the background image. In this paper , three regional image enhancement model was constructed according to the distribution characteristics of maize gray image. Then the morphological image processing methods were applied to segment the enhanced image and calculate the background plate scale. Based on the above background plate scale, the geometric feature parameters of maize could be computed accurately. In addition , an analysis was made in comparison with manual measure of maize feature. The comparison result showed that the image processing algorithm based on the background plate scale could measure the maize geometric feature accurately. The accuracy rate was above 96.01%. This algorithm raised in this paper can replace the traditional maize cultivar measure and realize the maize cultivar automatic analysis, which will play an important role in maize variety extension and breeding.