A logistics company wants to automatically identify and locate forklifts, pallets, and people within recorded footage from warehouse security cameras so it can flag safety violations. For each frame, the solution must return the type of each item and a bounding box showing where it appears. Which computer vision capability best fits this requirement?