论文部分内容阅读
梨的果梗是否存在是分级的重要特征之一。通过计算机视觉系统摄取黄花梨图象，应用图象处理技术完成图象与背景的分割。针对使用细化及收缩膨胀算法识别果梗速度较慢，提出了一种快速算法。该法利用梨果梗直径小，选择不同大小的模板，判别图象中是否存在果梗，同时得到果梗头、底部与梨相交点的坐标，依据切线斜率信息，对果梗的完好性进行判断。试验结果表明，该算法可以１００％判断果梗是否存在，判断果梗是否完好的正确率达到９３％，判别速度提高４～６倍。而且该算法具有一定的鲁棒性，对旋转、移位不敏感。
The presence of pear peduncle is one of the most important features of grading. Through computer vision system uptake pear images, the application of image processing technology to complete the image and background segmentation. Aiming at the problem of using the refinement and contraction expansion algorithm to identify stems, the fast algorithm is proposed. The method uses the small diameter of the pear stems, the selection of different sizes of the template to determine whether the presence of fruit stems in the image, and the stems obtained at the bottom, pear and the bottom of the intersection point of coordinates, according to the tangent slope information on the integrity of the stems and stems judgment. The experimental results show that the algorithm can determine the presence of peduncle 100%, determine the correctness of pedicle integrity is 93%, and discriminate speed is increased by 4 ~ 6 times. Moreover, the algorithm has some robustness and is insensitive to rotation and displacement.