The common functions of perception systems include lane detection, obstacle detection, and more. A camera-based obstacle detection perception system is used as an example to demonstrate the implementation of scene data augmentation based on RL. For camera-based perception systems, the input data consists of camera images; therefore, real world data scene augmentation can be achieved through various image processing tools.