CAJ | 학술논문

基于 RGB 颜色空间的阈值分割方法的优点是计算量小能满足实时性的要求,最大缺点是难以确定合适阈值,对光照变化很敏感.本文将对 RGB 阈值分割方法进行改进,为交通标志检测建立一种有效的颜色概率模型,以克服光照变化的影响,采用皮尔森相关系数法(Pearson Correlation Coefficient)实现交通标志的识别.实验结果表明本文算法具有效性和鲁棒性.
기우 RGB 안색공간적역치분할방법적우점시계산량소능만족실시성적요구,최대결점시난이학정합괄역치,대광조변화흔민감.본문장대 RGB 역치분할방법진행개진,위교통표지검측건립일충유효적안색개솔모형,이극복광조변화적영향,채용피이삼상관계수법(Pearson Correlation Coefficient)실현교통표지적식별.실험결과표명본문산법구유효성화로봉성.
The advantage of segmentation with thresholds on RGB space is the low computational complexity. It could fulfill the requirement of real-time performance. However, it is hard to define reasonable thresholds and it is sensitive to illumination changes. In this article, the method of segmentation with thresholds on RGB space is improved by constructing an effective color probability model for traffic sign detection. The novel method could get rid of the influence of illumination changes and implement traffic sign recognition with Pearson Correlation Coefficient. The experimental results show the feasibility and robustness of the proposed algorithm.