Pavement Condition Analysis and Decision Making based on Big Data and Machine Learning

This thesis presents an end-to-end, data-driven framework for network-level pavement management, linking distress measurement, condition forecasting, and maintenance prioritisation. It develops robust deep-learning methods for patch and crack detection with measurement-oriented outputs and improved reliability under field variability. A practical multi-source pipeline aligns condition, traffic, and climate data to consistent segment–time units, infers maintenance signals when records are missing, and supports forecasting under irregular survey intervals with efficient model updating. A web-based DSS integrates these outputs to produce interpretable, budget-feasible maintenance programmes.