ABSTRACT This study aims to redefine the role of green buildings (GB) in sustainable urban design by integrating deep learning (DL)–based façade segmentation and Internet of Things (IoT)–driven environmental information. The study's specific objectives are to (i) create a precise DL framework for extracting façade characteristics, and (ii) combine these visual insights with IoT‐based air quality metrics to comprehend their collective impact on healthy and sustainable urban settings. A U‐Net segmentation architecture is utilized to analyze façade properties, while addressing class imbalance challenges in the open‐access dataset through six data augmentation procedures. Two DL models are proposed—Model‐I (baseline) and Model‐II (enhanced)—and evaluated on both the original façade dataset and a newly generated edge‐enhanced dataset. Model‐II shows better performance, with a segmentation accuracy of 0.98. In addition to segmentation, the study examines how the architecture of a building's façade impacts air quality in cities and the performance of microclimates. The research demonstrates that integrating façade conditions with real‐time IoT air‐monitoring data can enhance ventilation, diminish pollutant exposure, and facilitate climate‐resilient design through optimized façade layouts. The results show that using DL‐enabled facade analysis, along with feedback from the environment, can help architects and city planners create buildings and cities that are energy‐efficient, low‐carbon, and beneficial to people's health.
Garima Verma (Thu,) studied this question.