This theme is designed to implement the robot＇s appearance and practical functions. Apply PSPNet to detect the walkable plane and Yolo to detect obstacles, so that the robot has the autonomous obstacle avoidance function, informing more information about the environmental obstacles around the visually impaired, and apply CNN to locate indoor position with self-built indoor database.

本主題使用深度學習網路架構應用之導盲機器人，其可利用影像偵測及圖像分析進行式室內平面偵測、室內定位，並可將辨識結果以語音方式告知視障者，藉以提升視障者之行進安全。

This topic applies three deep learning network architectures (PSPNet, Yolo, and CNN) for plane detection, object detection, and indoor positioning, integrates the three aforementioned functions to achieve the guiding function. The network architecture can be applied to image recognition