Dual Structured Light 3D using a 1D Sensor Jian Wang, Aswin C. Sankaranarayanan, Mohit Gupta, and Srinivasa G. Narasimhan Why Use a 1D Sensor for 3D Scanning? Wavelength Shorter Longer Visible Ultraviolet Infrared Structured Light Outside Visible Spectrum or Using Special Sensors Visible SWIR Penetrate Atmosphere Obscurants See through Smoke and Fire Visible SWIR Short-wave infrared (wavelength 0.9um~2.5um) Per-pixel Price Spectral band Detector Technology $/pix NIR/VIS/NUV Si ൏ ͳͲ SWIR InGaAs ͳͲ ଵ Dynamic Vision Sensor ͳͲ ଵ Case studies: 2D SWIR Camera ͳͲʹͶ ൈ ͳͲʹͶ Proposed: Optical Dual of Traditional Structured Light System Hardware Prototype and Results Objective lens Relay lens Cylindrical lens Line-sensor Projector 2D camera helper Cylindrical lens Line-sensor Hardware Setup Conventional Structured Light Proposed: Dual Structured Light (DualSL) Structured Light 3D Imaging The Digital Michelangelo Project Microsoft Kinect Depth map Pattern on scene Cost $100K 1D SWIR Camera ͳͲʹͶ ൈ ͳ Cost $100 2D DVS Camera ʹͶͲ ൈ ͳͺͲ Cost $5K Cost $100 1D DVS Camera ͳͲʹͶ ൈ ͳ Scene 2D Camera 2D Projector Scene 2D Camera 1D Projector Cylindrical lens Scene 2D Projector 1D Camera Cylindrical lens Replace the 2D projector by 1D projector + cylindrical lens 3D location obtained by ray-plane triangulation Switch the role of camera and projector Ray Diagrams of the Optics Design z x Objective lens Line- sensor z y Cylindrical lens Objective lens Line- sensor Cylindrical lens Without cylindrical lens With cylindrical lens Concave bowl Translucent wax Traditional SL with 2D sensor DualSL with 1D sensor 1.6mm mm Results 3D Meshes