A5000037660- Infrastructure Maintenance and Monitoring
- Asphalt Pavement Performance Evaluation
- Non-Destructive Testing Techniques
- Advanced Neural Network Applications
- Remote Sensing and LiDAR Applications
- Concrete Corrosion and Durability
- Traffic Prediction and Management Techniques
- Vehicle License Plate Recognition
- 3D Surveying and Cultural Heritage
- Concrete and Cement Materials Research
- Industrial Vision Systems and Defect Detection
- Geophysical Methods and Applications
- Anomaly Detection Techniques and Applications
- Image and Object Detection Techniques
- Innovative concrete reinforcement materials
Southwest Jiaotong University
2022-2024
Abstract Simultaneous pixel‐level detection of multiple distresses and surface design features on complex asphalt pavements is a critical challenge in intelligent pavement survey. This paper proposes deep‐learning model named ShuttleNet to provide an efficient solution for this by implementing robust semantic segmentation pavements. The proposed aims at repeating the encoding–decoding round freely or even endlessly such that contexts different resolution levels can be learned integrated many...
Accurate identification of cracks at the pixel level on intricate asphalt pavements represents a crucial challenge in domain intelligent pavement assessment. The current advanced deep-learning networks encounter limitations simultaneously capturing both global context and local features cracks, leading to discontinuous segmentation results suboptimal recovery details. This paper proposes robust architecture named Mix-Graph CrackNet present an efficacious solution for this challenge....
Pixel-level detection of expansion joints on complex pavements is significant for traffic safety and the structural integrity highway bridges. This paper proposed an improved HRNet-OCR, named as segmentation network (EJSNet), automated pixel-level asphalt pavement. Different from high-resolution (HRNet), EJSNet modifies residual structure first stage by conducting a Conv. + BN ReLU (convolution batch normalization rectified linear unit) operation each shortcut connection, which can avoid...
Automated pavement condition survey is of critical importance to road network management. There are three primary tasks involved in surveys, namely data collection, processing and evaluation. Artificial intelligence (AI) has achieved many breakthroughs almost every aspect modern technology over the past decade, undoubtedly offers a more robust approach automated survey. This article aims provide comprehensive review on collection systems, algorithms evaluation methods proposed between 2010...
Concurrently detecting multiple objects of interest will yield massive time savings in processing and enable a more streamlined unified detection system. The ShuttleNet is designed to repeat the encoding–decoding round freely or even endlessly, achieving prodigious successes terms simultaneous pavement distresses surface design features on asphalt pavements. This paper proposes an efficient improved architecture called ShuttleNetV2 for enhanced global modeling retrieving fine details...
The location and pixel-level information of the patch are all critical data for quantitative evaluation pavement conditions. However, obtaining both parch simultaneously is a challenge in intelligent surveys. This paper proposes deep-learning-based instance segmentation network (PISNet) that employs you only look once (YOLO)v5 as baseline adds semantic branch to provide an effective solution this challenge. proposed PISNet replaces original backbone CSPDarknet53 neck YOLOv5 with novel...
Accurate recognition and location of pavement manholes are great significance for maintenance. This paper proposes an improved You only look once X (YOLOX) automated detection on asphalt pavements. The proposed model improves the performance YOLOX in two respects. First, channel attention mechanism is introduced to enhance model's adaptive feature refinement; second, a microscale layer deployed extract more essential distinct features. experimental results impressive, with achieving F1 score...
Abstract At the present time, increasing attention is being paid to detection of road facilities, such as manhole covers— an important type facility which can have tangible impacts on driving safety and comfort. This paper proposes a robust method based modification Faster Region Convolutional Neural Network (Faster R-CNN) automatically detect pavement covers. We establish manually annotated image library that consists 1,245 cover images collected by 1 mm laser imaging system, implement...
Regular detection of pavement manhole conditions is significant for the integrity drainage system, and traffic safety. This paper proposes an improved Single Shot MultiBox Detector (SSD) automated under complex environments. The proposed method replaces “Base Network” (VGG16) original SSD with a novel feature extractor named Symmetrical Pyramid Network (SPN) to improve semantic features shallow layers localization deep layers. A modified Feature Selection Module (FSM) incorporated into SPN...
Pixel-level detection of sealed cracks on pavements is great significance for pavement maintenance. The Multi-fusion U-net network based proposed to detect cracks. multi-fusion module, dual attention mechanism, and Atrous Spatial Pyramid Pooling (ASPP) are designed efficiently capture the details increase receptive field. Trained with a dataset 2463 crack images, demonstrated outperform DANet. experimental results indicate that F-measure IOU 200 testing images 84.36% 0.7295 respectively....
This paper proposes a robust semantic segmentation algorithm named as Marking-DNet to implement pixel-level recognition of pavement markings. The proposed presents an improved encoder-decoder architecture based on DeepLabV3+. Different from DeepLabv3+, feature maps four different scales are imported the encoder into decoder Marking-DNet, resulting in more levels information exchange. Additionally, Object-Contextual Representation and Convolutional Block Attention Module both employed conduct...