Abstract:
This thesis introduces a novel approach for calculating as-built quantity take-offs, for automated progress updation in road construction, by blending with drone technology, machine learning and artificial intelligence. It addresses the challenges of traditional construction surveying methods, known for their inaccuracies, high costs, and inefficiency. The study showcases an innovative approach using drone imagery combined with a Convolutional Neural Network (CNN) model, enhancing the precision and efficiency of construction monitoring. Central to the research is a CNN model trained on images of various road layers: asphalt, subbase, and subgrade. The model was tested in a practical setting, capturing drone images before and after each layer's application. These stereo images were processed using Agi-soft Meta-shape to generate dense point clouds, later used in Civil 3D to create surface models for accurate as-built quantity take-offs of the materials. A key innovation of this study is the model's capacity to differentiate between road layers, a new development in construction monitoring. The model effectively identifies layer types in images, streamlining the as-built quantity take-off report update process. As-Built Quantity take-offs by this method showed a close correlation (90%) with traditional manual calculations, marking a significant step toward automating construction monitoring. The thesis underscores the synergy of drone imaging, machine learning, and civil engineering software in revolutionizing construction project monitoring. This approach offers a more efficient, cost-effective, and accurate alternative to manual methods. The research provides a foundation for future advancements, suggesting the exploration of advanced machine learning models, high-quality drones with better GPS and 3D capabilities, and software integration at the coding level for automated data processing. This work significantly contributes to the construction industry, promoting more automated, efficient, and advanced construction practices.
