针对圆形目标物检测过程中常用模板匹配算法和广义霍夫变换(GHT)算法都存在实时性不高的缺点,特别是模板匹配法对目标物混乱和遮挡的鲁棒性不强的问题,本文根据图像金字塔分层思想提出一种改进的GHT算法用于圆形目标物的检测,实验结果证明了该算法在光强非线性变化、噪声干扰和遮挡等情况下鲁棒性都很好。%As the template matching algorithm and GHT algorithm have the disadvantage of real time application, especially the template matching algorithm is less robust when object under clutter and occlusion, an improved GHT algorithm according to image pyramid hierarchical ideal is proposed which is used for circle object detection. The experimental results demonstrate that the algorithm is robust under non-linear illumination changes, noise disturbances and clutter conditions.
展开▼