A Quantitative Study of Eye-level,Three-Dimensional Environments Using Graphic Computation:A Case Study Based on Indoor and Outdoor Environmental Impact Assessments of New Buildings

ZHANG Yi; GAO Yu; ZHAO Wei

doi:10.3969/j.issn.1000-0232.2025.08.004

Abstract：Vision is the primary channel for human beings to acquire environmental information.Eye-level environmental shaping is an important research topic in space planning and design.Although traditional imaging-based methods have been extensively applied to the quantification of eye-level environmental features,they have many limitations since they are based on the two-dimensional projection principle,with significant perspective distortion,a lack of depth information,poor three-dimensional simulation,and individual semantic analysis.It is difficult to meet the needs of refined design and multi-dimensional evaluation.Hence,an eye-level,three-dimensional environmental quantification method based on graphic computation was proposed in this study.It established a spatial econometric system centered at planar angles,solid angles,and depth(volume)to quantify one-dimensional to three-dimensional features like linear size,visual proportions,and blocking relationships in the environment,respectively. Based on the properties of retinal spherical projections,the proposed method adopted a non-projection graphical computation strategy using polygon mesh data within Blender as the spatial data carrier and simulated the eye-level sampling process via a GPU-accelerated ray-tracing algorithm.Sampling points were uniformly distributed on a spherical coordinate system using a Fibonacci lattice parameterized by the golden angle,effectively avoiding the high density at polar regions and sparse distribution at the equator in traditional latitude-longitude sampling.Each point recorded spatial locations,angular features,and semantic information to construct a comprehensive eye-level spatial feature map.Multi-dimensional quantitative indicators were extracted through geometric calculations and semantic aggregation,supporting the quantitative analysis of various types of landscape elements in the environment. The proposed method was verified by analyzing the effects of new high-rise buildings on heritage courtyards and residential areas.A total of 4,565 observation points were set in the heritage scene,and the imaging-based method produced error levels ranging from-16.70％to 13.60％in visual proportion measurement compared to solid angles.The error levels could be amplified by more than 400-fold in the extremum area of the angle of altitude,but might produce underestimations of 36.33％at the horizontal view angle.The solid angle showed higher stability and reliability in visual proportion assessments.Combined with semantic information,the different effects of new buildings on heritage environments could be identified,mainly concentrating on northern and northwestern waterfront areas.The individual influencing features were significantly different.According to analysis of 5,696 observation points in the residential scene,the visual volume of units below the seventh floor of new buildings was greatly reduced,accompanied by a notable loss of preferred landscape elements from the 10th to 13th floor.High floors were mainly blocked by greenery and water.Significantly different eye-level features could be observed on different floors. According to the research results,the proposed method has remarkable advantages over traditional methods in terms of measurement accuracy,application ranges,and semantic identification ability.By introducing the three-dimensional spatial data,the simulation and analysis of eye-level features more closely resembled practical experiences,supporting the unified assessment of statuses and schemes.Multi-dimensional features like dihedral angle,solid angle,depth,and volume reflect the spatial complexity of environmental perceptions.The involvement of individual semantic labels improves recognition of single buildings and landscape elements and influences judgment,which can provide basic support to spatial interventions and fairness evaluations.The proposed method can be extensively applied to street quality assessments,heritage protection efforts,walkability studies,visual fairness analysis,and other fields.It can be a substitute for traditional indicators,including green view indexes and sky view factors.By combining semantic data and three-dimensional models,it is possible to connect subjective perceptions and objective spatial features,expanding the development paths of multi-dimensional,individualization,and design orientations in eye-level studies.