AI-Based Smart Cane for the Blind Using Raspberry Pi 3B

Abstract

The eye is one of the human sensory organs that plays an important role in human life. The main function of the eye is to record the surrounding conditions, so that it can provide information in everyday life. People who have visual impairments are called blind. This limitation is one of the causes of disruption in the daily activities of blind people. Therefore, in their daily activities, blind people use assistive devices, such as canes. In this study, the author will create a system that can later be implemented in a cane with features to facilitate the activities of blind people. This cane has a camera that captures object images, which are later processed by the Raspberry Pi 3B+ using the Faster R-CNN algorithm. Faster R-CNN is a deep learning method used to detect objects in images. In addition, this cane is equipped with several supporting sensors, including an HC-SR04 ultrasonic distance sensor and a water level sensor for detecting water puddles. The output of all systems is sound, transmitted via wireless earphones and a buzzer as an information source/alarm for the visually impaired. The results of this research can assist the visually impaired by providing an assistive device for daily activities.