中文摘要：

同时定位与建图（SLAM）是实现移动机器人真正自治的必要前提,视觉传感器由于能够提供丰富的环境信息而在SLAM研究中受到重视,本文从视觉传感器配置方式、视觉特征提取方法、视觉SLAM实现机制、地图表示类型以及环境对视觉SLAM的影响五个方面综述基于视觉传感器的同时定位与建图研究的发展现状,对已有的典型视觉SLAM方法进行分析和比较,并展望了未来的发展趋势.

英文摘要：

Simultaneous localization and mapping（SLAM）are necessary prerequisites to make mobile robot truly autonomous.The vision sensor attracts more and more attention in the SLAM research for its rich information of the environment.This paper summarizes the state of the art in the vision-based SLAM from the aspects such as the deployment of vision sensors,the extraction of visual features,the implementation mechanism of visual SLAM,the representation of the map and the effect of environ-ment on visual SLAM.Some classical visual SLAM methods are compared and analyzed.The research prospects are described.