This study proposes a novel framework integrating Building Information Modeling (BIM) and Geographic Information Systems (GIS) with real-time crowd analytics from Closed-Circuit Television (CCTV) for quantitative walkability assessment. The framework extends open data standards (IFC and CityGML) to model infrastructural and pedestrian flow attributes comprehensively. A walkability scoring mechanism quantifies route quality based on accessibility, efficiency, and physical comfort, differentiating among pedestrian groups, such as individuals sensitive to weather conditions or carrying belongings. Implemented at the Hong Kong University of Science and Technology (HKUST), results indicate that the framework effectively captures variations in walkability scores due to directional differences (uphill vs. downhill), crowd conditions, and operational constraints like facility closures. Statistical tests confirm significant differences in walking costs across these scenarios with variations of up to 30%, demonstrating the framework’s robustness and practical utility for real-time, human-centric urban infrastructure planning.

Ensuring automatic safety inspection on dynamic construction sites requires reliable multi-entity visual interpretation and rule-based compliance checking, which conventional computer vision systems struggle to achieve due to limited semantic reasoning. Although modern vision-language models (VLMs) support end-to-end multimodal reasoning, generic VLMs lack fine-grained visual grounding for safety-critical inspection. This study proposes a transparent, object-centric VLM framework that (1) formalizes four visually grounded reasoning strategies-Enumerate Anchor Entities, Key-feature Identification, Sequential Inspection and Explicit Visual Grounding to guide VLMs toward evidence-based inspection; (2) auto-generates a self-instruct visual-grounding dataset for cold-start supervised fine-tuning, enabling internalization of inspection behaviors; (3) optimizes rule-level accuracy using Group Relative Policy Optimization with designed ComplianceReward function that reinforces reasoning trajectories leading to correct answers. Trained on Qwen3-VL-8B, the proposed agent achieved 0.916 violation recall and 0.904 overall accuracy in real-site evaluation, outperforming baseline models. The framework delivers scalable, interpretable, and domain-aligned multimodal reasoning for automated construction safety inspection.