Application of UAV Technology in Smart Construction in Construction Monitoring

Abstract

The unmanned aerial vehicles (UAVs), construction sites may obtain detailed aerial imagery and used to create 3D models and track development in real time. UAVs capture aerial data, enabling creation of 3D models and real-time progress tracking. This improves both safety by allowing inspections of risky areas and efficiency by offering a comprehensive view of the entire site. In this research application of UAV technology in smart construction in construction monitoring is proposed. Initially, the images are collected from Drone-View Building Identification by Cross-View Visual Learning and Relative Spatial Estimation (DVBI-CVVL-RSE) dataset. Then, the collected data is fed to Pre-processing segment. In pre-processing stage, Unscented Trainable Kalman Filter (UTKF) used to remove the noise. Then pre-processed output is given to Multi-Objective Matched Synchrosqueezing Chirplet Transform (MOMSCT) is used to extracting the image features such as shape, colour, and texture. The extracting features are fed into the pseudo-hamiltonian neural networks (PHNN) used to forecast the class and location of construction building site in a full image or video. The UAV is using to capture the photo of construction site and Used to track in construction site work progress in real time. The weight parameters of MORARNN are optimized using Osprey Optimization Algorithm (OOA).Therefore, the suggested approach looked at using performance measures like accuracy, computation time, error rate, Fl-score, precision, recall, receiver operating characteristic curve (ROC), Sensitivity and specificity. The proposed PJH-CPI-CS-RBNN approach attains29%, 24.5% and 20% higher accuracy, 30%, 20% and 25.5% higher Precision and 27.5%, 25.5% and 22% higher sensitivity compared with existing methods such as Towards UAVs in construction: advancements, challenges, future directions for monitoring (VAVC-ACFDMT-RNN), Change detection in unmanned aerial vehicle imageries for progress monitoring of road construction (CD-UAVT-PMRC-ANN), and Application of deep learning with unmanned aerial vehicle on building maintenance (ADL-UAV-BM-CNN).By comparing other three existing methods, the proposed PJH-CPI-CS-RBNN method gives high accuracy models respectively.