基于计算机视觉的路基填料智能检测技术研究OA

To explore the application of intelligent detection methods based on computer vision technology in roadbed filler quality control.The instance segmentation model"Mask R-CNN"was employed to automate the segmentation,particle size calculation,and gradation analysis of fillers.Additionally,drone technology combined with computer vision was used to identify oversized particles,with a real-time early-warning mechanism introduced.This study compared resistance and electromagnetic wave methods for detecting filler moisture content and used neural networks to correct detection results,thereby improving accuracy.The result shows that,the intelligent detection technology significantly enhances testing efficiency and accuracy,reduces manual intervention,and provides effective support for quality control in roadbed construction,which has promoted the digitalization and intelligence of the construction process.