Using natural language processing to identify and evaluate heritage value for buffer zone delineation in Hengdaohezi Town

Understanding the value of urban heritage and protecting it effectively has become an increasingly significant challenge for many heritage cities. However, current buffer zone methods often lack comprehensive assessment and merely impose rigid boundaries around built heritage. Here, based on natural language processing, we extracted 17 value themes across 4 value aspects from 960 conservation documents and established a heritage value evaluation system comprising 11 factors. The effectiveness of this approach was demonstrated through a case study of conservation practices in Hengdaohezi Town, China. By quantifying and scoring over 180 heritage attributes and generating value assessment maps, we delineated a conservation core zone covering 22.97 hectares and a development control area covering 34.28 hectares. This study systematically examines a broader range of heritage attributes and their values, offering more precise delineation criteria and reducing the subjectivity of previous conservation practices.