The document presents an algorithm for covering 3D objects with a set of camera views at a specified precision level. The algorithm first computes an initial set of camera views to cover the object's triangles. It then selects a subset of cameras that maximally covers the object and attempts to add cameras to cover remaining triangles. Experiments show the algorithm can cover test objects like a dinosaur and dog with 30-35 views at 99-99.8% coverage. The algorithm is also applicable to illumination, camera placement, range data validation, and compression.