Abstract: In order to accurately detect the thickness of the synchronous grouting layer and its distribution in shield tunnels,this study relies on the Shanghai Rail Transit Municipal Line Airport Liaison Line project,and adopts the ground-penetrating radar to obtain the synchronous grouting information behind the shield wall.The collected data are pre-processed to perform the manual recognition and XGBoost intelligent recognition respectively.Combined with on-site excavation archaeological verification,the effectiveness of GPR system and XGBoost intelligent algorithm is analysed,and the grouting effect is evaluated.The results show that:the thickness of the grouting layer obtained by both manual recognition and intelligent recognition is within the range of from 28cm to 38cm.The XGBoost algorithm has the highest processing precision,with an accuracy rate of 96.3%,while comparing with the on-site archaeological test,it has a small error (2~3cm),and it is considered that it can be more accurately analysed for the grouting image sub-interface.The intelligent recognition results of shield grouting based on XGBoost can be fed back to guide the setting of synchronous grouting pumping pressure and slurry injection volume in subsequent projects,and provide refined guidance for secondary grouting.