Abstract: With the rapid advancement of intelligent coal mining technologies, seismic monitoring while excavation has gradually become one of the core technologies for enhancing excavation safety and work efficiency.In traditional coal mine excavation processes, advance detection typically requires halting the excavation at the working face.This not only limits the real-time nature of detection but also causes work stoppages, severely impacting overall tunneling progress and mine safety management.We propose an innovative target detection method based on diffraction imaging results from seismic monitoring while excavation, combined with deep learning techniques to improve advanced seismic detection during excavation.By processing and imaging real-time monitoring data, this study integrates Generative Adversarial Networks(GAN) with the Faster R-CNN object detection model, leveraging the powerful features of deep learning to significantly enhance the accuracy and robustness of target detection.The research results indicate that the proposed method can efficiently and accurately identify potential geological anomaly zones within the coal mine working face, such as faults and other complex geological structures.This provides more real-time and precise safety warnings for mine excavation and enables the early detection of potential hazards while ensuring normal excavation at the working face.Practical application and validation in the Zhaozhuang No.2 Mine project fully demonstrate the significant advantages of this method in improving prediction accuracy and detection efficiency.It offers strong technical support for support and hazard prevention in mine tunneling faces, promotes the development of intelligent coal mine monitoring systems, and provides new solutions and technical safeguards for future mine safety management.