English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Objective:Computer vision-based up-to-date accurate damage classification andlocalization are of decisive importance for infrastructure monitoring, safety,and the serviceability of civil infrastructure. Current state-of-the-art deeplearning (DL)-based damage detection models, however, often lack superiorfeature extraction capability in complex and noisy environments, limiting thedevelopment of accurate and reliable object distinction. Method: To this end,we present DenseSPH-YOLOv5, a real-time DL-based high-performance damagedetection model where DenseNet blocks have been integrated with the backbone toimprove in preserving and reusing critical feature information. Additionally,convolutional block attention modules (CBAM) have been implemented to improveattention performance mechanisms for strong and discriminating deep spatialfeature extraction that results in superior detection under various challengingenvironments. Moreover, additional feature fusion layers and a Swin-TransformerPrediction Head (SPH) have been added leveraging advanced self-attentionmechanism for more efficient detection of multiscale object sizes andsimultaneously reducing the computational complexity. Results: Evaluating themodel performance in large-scale Road Damage Dataset (RDD-2018), at a detectionrate of 62.4 FPS, DenseSPH-YOLOv5 obtains a mean average precision (mAP) valueof 85.25 %, F1-score of 81.18 %, and precision (P) value of 89.51 %outperforming current state-of-the-art models. Significance: The presentresearch provides an effective and efficient damage localization modeladdressing the shortcoming of existing DL-based damage detection models byproviding highly accurate localized bounding box prediction. Current workconstitutes a step towards an accurate and robust automated damage detectionsystem in real-time in-field applications.

A Computer Vision Enabled damage detection model with improved YOLOv5 based on Transformer Prediction Head