- ホーム
- > 洋書
- > ドイツ書
- > Mathematics, Sciences & Technology
- > Technology
- > popular works
Description
A robotic camera lens naturally warps reality. Before it can be used to drive a car, it must be forced to stare at a mathematically flawless glass grid. An autonomous vehicle speeding down a highway at 70 mph relies entirely on an array of external cameras to identify pedestrians and calculate distances. But camera lenses are physically imperfect; the glass naturally warps and distorts the image at the edges. To prevent the car's AI from hallucinating the size and location of objects, the camera must be mathematically corrected at the factory using Optical Calibration Targets.This book exposes the hyper-precise, highly lucrative B2B economy of photogrammetry. Calibration targets are massive, flawless glass or ceramic boards printed with absolute, micron-perfect geometric checkerboard patterns. When a robot looks at this board, its software instantly measures how the lens warps the perfect squares, writing an algorithm to permanently un-distort every image that camera will ever take.We explore the intense manufacturing environment of the companies that produce these grids. As global demand for factory robotics, drones, and self-driving cars skyrockets, the invisible suppliers who manufacture the "eye exams" for artificial intelligence are quietly securing an absolute technological monopoly.Learn how we teach the machines to see the truth. Understand the billion-dollar glass grids that keep robotic vision mathematically flawless.
