The invention provides a road underground hidden danger detection method and system based on radar images and artificial intelligence. The road underground hidden danger detection method comprises the steps of S1, screening the B-SCAN sample pictures containing underground hidden danger targets from a three-dimensional ground penetrating radar database; S2, performing target detection marking on each target of the sample picture, and performing data enhancement processing to form an underground hidden danger target detection data set; S3, respectively training the two R-CNN target detection neural networks to obtain two target detection models with cavity, void, pipeline and manhole object detection capabilities; S4, reading the to-be-detected B-SCAN pictures of each channel acquired by the three-dimensional ground penetrating radar, performing multi-GPU parallel target detection by using the two target detection models, and generating two groups of reasoning results; and S5, fusing and outputting the two groups of reasoning results through model integration. The method and the system are high in anti-interference performance, rich in target information and high in accuracy.