Accuracy analysis of rebar quantity estimation and detection using random stereoscopic graph scalable convolutional diffusion quantum neural networks with Wombat Optimization Algorithm

초록

Accurate identification and estimation of rebar quantities in reinforced concrete structures are crucial for ensuring structural safety, maintenance, and repair planning. However, conventional automated methods often fail under noisy conditions. To address these challenges, this study proposes a Random Stereoscopic Graph Scalable Convolutional Diffusion Quantum Neural Network with Wombat Optimization Algorithm (RSGSConDQNNs + WOA) for robust rebar quantity estimation and detection. Raw GPR B-scan images from the Nebraska Bridge Deck GPR dataset are first enhanced using the Square Root Sage-Husa Adaptive Kalman Filter to suppress noise artifacts. Discriminative spatial features are then extracted using the Quaternion Quadratic-Phase Fourier Transform Domain. The suggested architecture combines Stereoscopic Scalable Quantum Convolutional Neural Networks and a Random Graph Diffusion Attention Network., while adaptive hyperparameter tuning is achieved using Wombat Optimization. Experimental results demonstrate a detection and estimation accuracy of 99.9%, outperforming conventional methods with improved reliability, scalability, and computational efficiency.

키워드