Building Customized Cushion Seats Using Stereo Visions for Disabled

This paper presents a systematic approach to construct a grid-based 3D surface model for a customized cushion seat for people with disabilities. Initially, a 2D grid map is defined as the bottom surface of the cushion seat based on the required resolution. Then each point in the 2D grid map is projected onto one surface point in the 3D pixel-based point cloud. The height of each surface point is approximated using the nearest neighbor algorithm. A tool box is developed for the visualization the constructed 3D surface model. The height of the 3D surface computed using this tool box is compared with the manually designed cushion seat at Precision Rehab Manufacturing Inc. and found to be within tolerance.