Leveraging Stable Diffusion for Monocular Depth Estimation via Image Semantic Encoding