With the growth of waterway transportation,traditional channel inspection methods struggle to meet the need for efficiency and precision,especially in complex environments for navigation aids.In view of the above problem,this paper proposes an intelligent navigation aids recognition method based on unmanned aerial vehicle and artificial intelligence.The method integrates an improved YOLOv7 model,convolutional block attention module attention mechanism,and MobileViT lightweight structure to address low small target detection accuracy and high background interference in complex aquatic environments during the day.In the nighttime,hue-saturation-value color space segmentation and time-series analysis techniques effectively extract navigation aid light flicker cycles and mitigate multi-source light interference.The experimental results show that the improved model achieves an average precision of 0.99 and an average confidence of 0.97,while nighttime recognition accuracy reaches 97.1%,significantly outperforming traditional methods.Additionally,the exchangeable image file format-based pose information geometric mapping model enables high-precision conversion from image coordinates to geographic coordinates,enhancing location accuracy.The proposed method operates stably in dynamic environments and provides effective technical support for the intelligent operation and maintenance of inland waterway navigation aids,offering substantial engineering application value.