复杂光照条件下钢坯字符检测方法

摘要

To solve the problem of the detection of billet character in billet intelligent recognition system, we proposed a detection method of billet character based on multistage segmentation. The billet image was segmented recursively by maximum between - cluster variance. An effective projection character to determine the terminal condition of multistage segmentation was used, then, the interesting region was labeled and selected by clustering characteristics. The experiments show that the detection method has better accuracy and stability in complex scenes, solves the problems of adaptability and robustness in complex illumination conditions.%针对钢坯智能识别系统中字符检测问题,提出一个基于多次分割的钢坯字符检测方法.该算法通过计算最大类间方差,对处理图像做迭代分割；然后根据字符信息提出有效的投影特征作为确定多次分割的终止条件,利用聚类特征标记并筛选出所需要的兴趣区域.实验表明,该方法能在复杂场景中有更好的准确性和稳定性,解决了复杂光照条件下字符定位自适应性和鲁棒性的相关技术问题.