To realize the intelligent and accurate measurement of pavement surface potholes, an improved You Only Look Once version three (YOLOv3) object detection model combining data augmentation and structure optimization is proposed in this study. First, color adjustment was used to enhance the image contrast, and data augmentation was performed through geometric transformation. Pothole categories were subdivided into P1 and P2 on the basis of whether or not there was water. Then, the Residual Network (ResNet101) and complete IoU (CIoU) loss were used to optimize the structure of the YOLOv3 model, and the K-Means++ algorithm was used to cluster and modify the multiscale anchor sizes. Lastly, the robustness of the proposed model was assessed by generating adversarial examples. Experimental results demonstrated that the proposed model was significantly improved compared with the original YOLOv3 model; the detection mean average precision (mAP) was 89.3%, and the F1-score was 86.5%. On the attacked testing dataset, the overall mAP value reached 81.2% (−8.1%), which shows that this proposed model performed well on samples after random occlusion and adding noise interference, proving good robustness.

The application of reclaimed asphalt pavement (RAP) can help reduce resource waste and environmental pollution in road construction. However, so far, only a small percentage of RAP materials can be used in road construction. The key obstacles to the application of a recycled asphalt mixture (RAM) with high RAP content are the variability of RAP materials and the difficulty of fully rejuvenating aged asphalt. However, there is still a lack of research on the effect of the variability of RAP materials and recycled asphalt on the quality control of a RAM. Therefore, this study investigates the effects of sieve pretreatment of RAP material using 4.75 mm sieve mesh and the use of composite crumb rubber-modified asphalt (CCRMA) to reclaim aged asphalt on the road performance and frame variability of reclaimed asphalt mixtures. Therefore, this study investigates the effects of the fractionation process of RAP material using 4.75 mm sieve mesh and the use of CCRMA to reclaim aged asphalt on the road performance of a RAM. The results show that the fractionation process can effectively reduce the mitigation of RAP agglomeration and reduce the variability of gradation, which in turn reduces the variability of road performance. The incorporation of CCRMA can effectively improve the high-temperature stability performance and low-temperature cracking resistance. The dynamic stability and the fracture energy of the CRAM (RAM prepared using CCRMA) were four and one and a half times as large as that of the NAM (RAM prepared using base asphalt), respectively. The fractionation process of RAP material and the utilization of CCRMA could help reduce the variability of the RAM while improving the road performance of the RAM.