video1.0<iframe src="https://www.loom.com/embed/9e2e225171b744cf80133016b9977a93" frameborder="0" width="1680" height="1260" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>12601680Loomhttps://www.loom.com12601680https://cdn.loom.com/sessions/thumbnails/9e2e225171b744cf80133016b9977a93-9c7a4f65131cd0c1.gif591.785Leaf Jiang explains how to build an ultra-wide baseline stereo vision camera that works in near-IR using an invisible illumination approach. It motivates near-IR operation at night with LEDs to avoid blinding drivers. The build uses IR-corrected long focal-length lenses (35 mm, 25 mm, and 16 mm) and measures intrinsic parameters with Image Engineering’s GEOCAL, using a pinhole model with K1 to K3 and P1 to P2 distortion coefficients. It describes removing the camera’s near-IR filter, then compensating for focal-plane shift by sanding C-mount threads by about 0.5 mm, mounting six cameras on a 52-inch bar with baselines from 0.5 to 0.8 m. It concludes with calibration steps for extrinsics, synchronization, and testing across day and night data collection. https://www.nodarsensor.com/hdk