Real-Time Obstacle Detection System in Indoor Environment for the Visually Impaired Using Microsoft Kinect Sensor

摘要: Any mobility aid for the visually impaired people should be able to accurately detect and warn about nearly obstacles. In this paper, we present a method support system obstacle in indoor environment based on Kinect sensor 3D-image processing. Color-Depth data of scene front user is collected using with standard framework 3D sensing OpenNI processed by PCL library extract accurate information The experiments have been performed dataset multiple scenarios different lighting conditions. Results showed that our four types obstacle: walls, doors, stairs, residual class covers loose obstacles floor. Precisely, walls floor are detected practically all cases, whereas doors 90.69% out 43 positive image samples. For step detection, correctly upstairs 97.33% 75 images while correct rate downstairs detection lower 89.47% from 38 images. Our further allows computation distance between