摘要
为了解决视觉引导的机器人加工系统中对目标进行快速、准确检测的问题,将人类视觉选择性注意机制引入机器视觉系统,提出了一种基于轮廓的视觉选择性注意计算模型(SECO模型)。该模型通过分离目标、提取边缘、感知轮廓、轮廓显著度竞争、注意焦点选择及转移等策略,实现对目标的检测。实验结果表明,该模型能够实现在复杂场景下对目标进行快速准确的检测,且具有很强的抗干扰能力和较高的计算效率,便于在机器视觉系统中直接应用。
To solve the problem of target detection, which is appeared in the visual guide robot processing system, the human visual selective attention mechanism was introduced into the machine vision system. A computational model of visual selective attention based on the contour was proposed. It realized the testing goal by using a strategy of separating object, extracting edge, perceiving contour, competing in the saliency of contour, choosing and transferring attention focus. Experimental results indicate that the model can be attained to the accuracy of target detection in complex scenes, possesses a strong anti--jamming capability and high computing efficiency, which has a potential of using in the machine vision system directly.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2008年第17期2024-2027,共4页
China Mechanical Engineering
基金
国家863高技术研究发展计划资助项目(2006AA04Z212)
