论文部分内容阅读
针对现有导盲方式的不足,设计一种基于双目视觉方法的可穿戴式导盲机器人。该机器人借助左右两个图像传感器实现障碍物定位和场景三维重建,并将提取出的信息以振动和语音形式反馈给穿戴者。从硬件和软件角度分析了可穿戴式导盲机器人的技术路线,介绍了双目视觉的原理与算法基本步骤,分析了基于神经网络法的多传感器信息融合算法,最后建立了人机适应性的评价指标体系,为该可穿戴机器人的实际设计奠定了基础。
Aiming at the shortcomings of the existing blind guiding method, a wearable blind guiding robot based on binocular vision method is designed. The robot realizes the obstacle location and the scene three-dimensional reconstruction with two left and right image sensors and feeds back the extracted information to the wearer in the form of vibration and voice. From the view of hardware and software, we analyzed the technical route of wearable head-guiding robots, introduced the principle of binocular vision and the basic steps of algorithm, analyzed the multi-sensor information fusion algorithm based on neural network method, and finally established a man-machine adaptive The evaluation index system lays the foundation for the practical design of the wearable robot.