Intelligent Recognition of Traffic Safety Facilities Using DINO Algorithm in Deep Learning

In an era of rapid technological advancement, the deep integration of computer technology with various fields, including traffic engineering, has become increasingly profound. Traditional methods of facility inspection and management suffer from inefficiencies and incomplete information, challenges that are effectively addressed by the intelligent and automated capabilities of computer technology. This pioneering study employs the DINO deep learning algorithm, focusing on comprehensive facility inspection and proposing an efficient intelligent recognition model. The model achieves an outstanding mean Average Precision (mAP) of 80.6%. Simultaneously, the article compares the training results of DINO with its predecessors, DN-DETR and Deformable DETR, highlighting the significant performance advantage of DINO. In comparison to previous research, this study places greater emphasis on the identification of various types of traffic safety facilities, further significantly enhancing the efficiency of traffic facility management and road safety.