With the advancement of Artificial Intelligence Generated Content (AIGC) technology,the application of images in educational settings has emerged as a new research focus.As a pivotal medium for knowledge transmission,the clarity,texture details,color vibrancy,and overall image fidelity of images directly influence teaching efficacy.This study aims to modify the architecture of the diffusion model to achieve Super-Resolution (SR) reconstruction of images suffering from various quality degradation issues,and to evaluate the effectiveness of applying these SR-enhanced images across diverse teaching contexts.Initially,the study addresses the mismatch between image quality and teaching requirements by refining the diffusion model's structural design.Subsequently,both subjective and objective experiments are conducted to integrate SR-reconstructed images into real-world teaching environments.Finally,a comprehensive evaluation framework is constructed based on the experimental findings to substantiate the practical benefits of the reconstructed images.The results show that compared to images generated by traditional methods,the application of SR images generated by the modified model in teaching activities improves the average efficiency of knowledge transfer by approximately 22.9%,and reduces the time teachers spend on lesson preparation by about 15.6%.This study provides a theoretical foundation and practical insights for leveraging artificial intelligence to drive pedagogical innovation.