Thanks to the development of deep learning, the use of data-driven methods to detect pavement distresses has become an active research field. This research makes four contributions to address the problem of efficiently detecting cracks and sealed cracks in asphalt pavements. First, a dataset of pavement cracks and sealed cracks is created, which consists of 10,400 images obtained by a vehicle equipped with a highway condition monitor, with 202,840 labeled distress instances included in these pavement images. Second, we develop a dense and redundant crack annotation method based on the characteristics of the crack images. Compared with traditional annotation, the method we propose generates more object instances, and the localization is more accurate. Next, to achieve efficient crack detection, a semi-automatic crack annotation method is proposed, which reduces the working time by 80% compared with fully manual annotation. Finally, comparative experiments are conducted on our dataset using 13 currently prevailing object detection algorithms. The results show that dense and redundant annotation is effective; moreover, cracks and sealed cracks can be efficiently and accurately detected using the YOLOv5 series model and YOLOv5s is the most balanced model with an F1-score of 86.79% and an inference time of 14.8ms. The pavement crack and sealed crack dataset created in this study is publicly available.

The high-precision positioning of mileage piles on a digital map ensures accurate data for high-speed event releases, toll audits, and road condition monitoring. This paper introduces an efficient and automatic system for positioning highway mileage piles. The self-developed data acquisition system collects mileage pile images and road trajectory. Considering the limited variety of mileage piles, a Simplified-YOLOv5m (S-YOLOv5m) is proposed. Then a high-precision character detection network of S-YOLOv5m is proposed to enhance feature extraction and improve the accuracy of character detection on mileage piles. Thirdly, the end-to-end monocular distance measurement combines the target detection with the distance estimation, enabling simultaneous object detection and distance measurement. Fourthly, by combining with the geographical coordinates of the acquisition point, the direct solution to geodetic problems is applied to calculate the spatial coordinates of the mileage piles. Finally, the missing mileage piles are compensated for using the road curve and equal distance interpolation. The complete mileage piles and geographical information list of the inspection trajectory are output. Through a series of verification tests, the average positioning error of the mileage pile system is 1.265 m. The contribution of the automatic positioning system of mileage piles is to construct the relationship between mileage piles and geographical coordinates, match mileage piles with digital maps to realize the full-featured, real-scene, and high-dynamic management of road attributes.

As one of the common road diseases, the accurate detection of potholes during inspections can help to make timely maintenance measures, which will greatly save road maintenance costs and reduce the incidence of traffic accidents. In order to improve the accuracy and timeliness of pothole detection and to facilitate the development of disease maintenance, the YOLOv7 model is improved. Firstly, the Efficient convolution operator (DSConv) is introduced to reconstruct the Backbone and Head parts of YOLOv7 to reduce the computational effort of the original YOLOv7 model and improve the detection speed of the model. Secondly, the SE attention mechanism is incorporated into the model to improve the model's ability to extract features from potholes. Finally, the latest v3 version of Wise-IoU is introduced as the loss function of the improved model to reduce the impact caused by sample annotation. The accuracy and mAP of the improved model improved by 3.04% and 1.34% respectively compared to the original YOLOv7 model, and the number of FLOPS of the model decreased from 105.1 to 48.8. The results show that the proposed improved method can effectively improve the speed and accuracy of the YOLOv7 model in detecting potholes, and is advanced compared to the current mainstream target detection algorithms.

1. High-accuracy pavement condition evaluation is fundamental to structural health monitoring (SHM) for transportation infrastructure, directly influencing traffic safety, serviceability and sustai...

Due to the low foreground brightness and distorted background brightness of traffic backlight images, a technique of fusing LIME and improved histogram equalization is proposed to eliminate the backlight phenomenon in images. Firstly, the input image is segmented into two parts, foreground and background, using the maximum interclass variance method. Then LIME method is used globally for the backlit image to enhance the foreground brightness while maintaining the color distortion, and only the foreground part of the processed image is retained. Then the pixel value distribution of the background part is calculated separately, and the global histogram equalization results on the three RGB channels are mapped one by one to the corresponding limited interval, which improves the contrast of the background. Finally, the Canny operator is used to detect the black edges at the front background stitching, and three adaptive filtering templates are generated based on the black edges to perform step-by-step mean filtering on the black edges, eliminating the black edges and improving the visual quality of the image. The average metrics of the proposed method on the laboratory selfconstructed CHD_B dataset are, respectively, NIQE of 5.37, BRISQUE of 47.74, average gradient of 0.79, information entropy of 6.61, and average running time of 11.38s, which are synthetically better than the current backlight image enhancement methods and have better visual quality and visibility

There are two challenges presented in parsing road scenes from UAV images: the complexity of processing high-resolution images and the dependency on extensive manual annotations required by traditional supervised deep learning methods to train robust and accurate models. In this paper, a novel unsupervised road parsing framework that leverages advancements in vision language models with fundamental computer vision techniques is introduced to address these critical challenges. Our approach initiates with a vision language model that efficiently processes ultra-high resolution images to rapidly identify road regions of interest. Subsequent application of the vision foundation model, SAM, generates masks for these regions without requiring category information. A self-supervised learning network then processes these masked regions to extract feature representations, which are clustered using an unsupervised algorithm that assigns unique IDs to each feature cluster. The masked regions are combined with the corresponding IDs to generate initial pseudo-labels, which initiate an iterative self-training process for regular semantic segmentation. Remarkably, the proposed method achieves a mean Intersection over Union (mIoU) of 89.96% on the development dataset without any manual annotation, demonstrating extraordinary flexibility by surpassing the limitations of human-defined categories, and autonomously acquiring knowledge of new categories from the dataset itself.