Smart Stick Detecting Obstacles Early Using Random Forest Algorithm

In everyday life, blind people always have to face more difficulty. In modern technology, we can design convenient devices to help them live more easily. This paper focuses on designing a product for them. Specifically, a smart cane will be designed to help blind people avoid objects. To identify and classify objects, we use the random forest algorithm. Our main contribution is to build a detection system and classify obstacles early using cheap sensors and cameras. The results show that the model achieves an accuracy of up to $\mathbf{9 7. 8 7 \%}$ at a distance of up to 94 cm. The results show that the device is promised feasibility when applied in practice.