Within the context of artificial intelligence-driven expansion of visual arts boundaries, the transformation of spatial structural information into visual representations constitutes a significant research focus in computational art studies. This study employs Depthmap spatial analysis software as the core computational tool for artistic research. By integrating methodologies from space syntax theory, visual narrative theory, and parametric generation theory, an experimental framework for image generation is established through a four-step pathway: spatial structure analysis, parameter extraction, semantic mapping, and artistic generation. The process is primarily driven by three key metrics derived from Depthmap: integration, connectivity, and visible area. This provides novel perspectives for the visualization of future urban concepts, the reconstruction of cultural heritage, and the implementation of immersive spatial installations. This study endeavors to integrate the structural logic of programming languages with the perceptual dimensions of artistic expression, thereby exploring novel possibilities for spatial language representation in computational art.