图像匹配算法联合特征点检测在无障碍多媒体阅读中的应用  

Application of image matching algorithm combined with feature point detection in barrier-free multimediareading

作　　者：李丹霞[1] LI Danxia(Shangluo university,Shangluo Shaanxi 726000,China)

机构地区：[1]商洛学院,陕西商洛726000

出　　处：《自动化与仪器仪表》2024年第6期261-264,共4页Automation & Instrumentation

基　　金：陕西省图书馆学会2021年课题《数字人文视域下陕西红色文化保护与传承研究》(211028)。

摘　　要：视功能障碍不仅约束着该弱势群体的行动能力,还严重影响着他们的生活、学习和工作。为了让视障人群可以直观、自然和友好地进行多媒体阅读,研究基于图像匹配算法和特征点检测,设计了一款无障碍多媒体阅读器,并通过图像匹配实验和旋转不变性实验验证了该系统的性能。结果表明,图像匹配算法联合特征点检测收敛速度快,能够每秒匹配675对,正确比配对数与匹配准确率分别达到了179对和96.24%。在大幅进行旋转的情况下,该系统具有较好的旋转不变性,视障人群在进行无障碍多媒体阅读时,可以快速、准确地检索出目标书目,极大地提升了阅读体验。此次研究方便了视障人群的学习生活,同时促进了无障碍和谐社会的构建。Visual dysfunction not only restricts the action ability of the vulnerable group,but also seriously affects their life,study and work.In order to enable visually impaired people to read multimedia intuitively,naturally and friendly,a barrier-free multimedia reader was designed based on image matching algorithm and feature point detection,and the performance of the system was verified by image matching experiments and rotation invariance experiments.The results show that the combined feature point detection by image matching algorithm can quickly converge on the data,and can match 675 pairs per second,and the correct matching logarithm and matching accuracy are 179 pairs and 96.24%,respectively.In the case of large rotation,it has good rotation invariance,and visually impaired people can quickly and accurately retrieve the target bibliography when they read barrier-free multimedia,which greatly improves the reading experience.This study facilitates the study and life of the visually impaired and promotes the construction of a barrier-free and harmonious society.

关 键 词：图像匹配 多媒体阅读 无障碍 特征点检测 

分 类 号：F416.67[经济管理—产业经济] TP29[自动化与计算机技术—检测技术与自动化装置]