改进的Otsu法在电梯门ROI提取中的应用
    点此下载全文
引用本文:支新鑫?覮,孙晓艳,张立材.改进的Otsu法在电梯门ROI提取中的应用[J].计算技术与自动化,2019,(1):167-172
摘要点击次数: 685
全文下载次数: 0
作者单位
支新鑫?覮,孙晓艳,张立材 (西安建筑科技大学 信息与控制工程学院陕西 西安 710055) 
中文摘要:针对传统电梯门安全保护方式的不足,开展了以阈值分割算法为核心的电梯门图像处理的研究。针对传统Otsu算法在目标区域具有低熵特点且背景区域较大的情况下无法得到理想分割阈值的问题,提出了一种基于波谷混合加权因子的全阈值分割算法,使得分割阈值处于电梯门灰度直方图谷底附近且更接近电梯门ROI(Region Of Interest,ROI)。同时结合电梯门图像特征缩小了算法的搜索范围。实验仿真表明,改进的算法相比于传统Otsu法和其他改进算法在电梯门图像的分割结果中,目标区域更接近于电梯门ROI,运算效率在同等仿真环境下提高2至6倍。
中文关键词:图像分割  图像熵  邻域频率信息  波谷加权  改进Otsu  
 
Application of an Improved Otsu Method in ROI Extraction of Elevator Door Images
Abstract:The study on elevator door image processing based on threshold segmentation algorithm was carried out to address the inadequacy of traditional elevator door protection methods. The traditional Otsu methods cannot obtained the satisfactory segmentation threshold values in some cases, such as the image entropy of targeted region is low and the background is relatively large. In order to solve this problems, a full-threshold segmentation algorithm based on mixing weighting valley-emphasis methods is proposed. And it ensured the segmentation threshold values near the valley of grey level histogram with low image entropy in elevator door image’s targeted region. In order to increase the segmentation efficiency, the image features were considered and the search scope of this method was derived and narrowed. Simulation results indicate that, as compared with the segmentation results of traditional Otsu method and other improved methods, the targeted region segmented by this methods was much closer to the Region Of Interest (ROI) in elevator door images. Additionally, calculation efficiency was improved by 2 to 6 times as compared with other valley-emphasis methods and neighborhood valley-emphasis methods under the same simulation environment.
keywords:image segmentation  image entropy  neighborhood frequency  information valley-emphasis method  improved Otsu
查看全文   查看/发表评论   下载pdf阅读器