Evaluation method for uniformity of steel slag concrete aggregate based on improved YOLOv8

• Published in: Journal of Building Engineering 2024-12, Vol.98, p.111046, Article 111046
• Main Authors: Ye, Guanting, Yuen, Ka-Veng, Jin, Qiang, Zhou, Manxu, Yin, Chao, Jiang, Qing, Zhao, Shangjun, Su, Wanzhong
• Format: Article
• Language: English
• Subjects: Aggregate segmentation; Instance segmentation; Steel slag concrete; Uniformity evaluation; YOLOv8
• ISSN: 2352-7102; 2352-7102
• DOI: 10.1016/j.jobe.2024.111046

The use of steel slag as a replacement for natural aggregates in concrete effectively alleviates the issue of natural resource scarcity. The uniformity of aggregates in steel slag concrete directly affects its performance and durability. Therefore, precise segmentation of aggregates in sedimentation images is crucial for evaluating their uniformity. This paper not only proposes an improved instance segmentation model for accurate segmentation of steel slag concrete aggregates but also introduces a new method for assessing the uniformity of aggregates when steel slag replaces natural aggregates. This method quantitatively analyses the area, replacement rate, and spatial distribution of the aggregates and conducts a comprehensive quantitative analysis using the comprehensive uniformity evaluation coefficient, K. This approach allows for the evaluation of the uniformity of aggregates in sedimentation images of steel slag concrete. Experimental results show that the proposed improved model achieved an mAP50 of 98.3 %. Furthermore, the aggregate uniformity assessment method based on the improved model can objectively and accurately evaluate the uniformity of steel slag concrete aggregates. •Introducing the Convnext V2-R architecture and the WS module, along with the EMA attention mechanism.•Proposing a method to evaluate the uniformity of steel slag concrete aggregates, quantitatively analyzing their uniformity.•The Precision, Recall, mAP50 of YOLOv8n-CSE reached 95.8%, 94.9% and 98.3%.