Numerical Detection of Concrete Slump by Fusion of Target Segmentation and Image Classification Network

Abstract

In construction, concrete compatibility is an important comprehensive index to ensure the construction, and the concrete slump is an important criterion to judge concrete compatibility in the actual construction process. In this study, we propose to extract new data sets from concrete mixing video sequences and correlate the image features characterized in the concrete mixing and transportation process with the concrete performance features, starting from the concrete transportation process [1]. First, the UNet network model for semantic segmentation is used to identify and locate the concrete regions, and the localized concrete regions are zoomed in using interpolation; then the ResNet image classification network is used to determine the slump category of the processed concrete regions; finally, the results of the semantic segmentation network and the image classification network are fused to obtain the final concrete slump detection results. The experimental results demonstrate that the proposed method can guarantee real-time concrete slump detection with improved accuracy.

Image-based Deep Learning for the time-dependent prediction of fresh concrete properties

“…Ding and An (2018) show that also deep learning methods, here a combination of a CNN and a long short-term memory network (LSTM) based on image sequences, are applicable. Yang et al (2021) and Guo et al (2022) employ another combination of CNN and LSTM with image sequences to predict the slump value and slump flow value respectively the plastic viscosity, while (Gao and Yan, 2023) use semantic segmentation in combination with a residual neural network for single images for the prediction of the slump class. In Ponick et al (2022), a stereo camera set up is used to observe the mixing process of ultra sonic gel, a often employed surrogate for concrete.…”