With high-dimensional state spaces, visual reinforcement learning (RL) faces significant
challenges in exploitation and exploration, resulting in low sample efficiency and training …