In this study, we focused on the recent spread of smartphones among the visually impaired and smartphones equipped with high-performance cameras including ToF(Time of Flight) cameras. By processing all data from photography to obstacle detection and notification with a smartphone, we realized a compact, easy-to-use system to detect obstacles. We devised two detection methods and combined them to compensate for their weak points, thereby improving the accuracy of detection. As a result, we were able to show that this system can reduce the number of accidents such as a visually impaired person falling off a station platform or colliding with a parked bicycle.