Back to results list
Please use this identifier to cite or link to this item:
|Title:||A 3D GIS-based valuation system for assessing the scenic view in residential property valuations||Authors:||Huang, Juan||Advisors:||Shen, Qiping Geoffrey (BRE)
Wong, Siu Wai Ivy (BRE)
|Keywords:||Real property -- Valuation
Real property -- Valuation -- Data processing
Real estate business
|Issue Date:||2019||Publisher:||The Hong Kong Polytechnic University||Abstract:||The total value of the world's wealth in 2016 was estimated at US$265 trillion, of which over half was embedded in real estate. From among residential, commercial and industrial properties, and land, residential properties account for the largest proportion of valuations for taxation, purchases and investments. In Hong Kong, by the end of March 2016 the total assessments were 2,454,450 units, among which 1,951,533 were residential units, accounting for nearly 80% of total assessments. Therefore, the importance to both governments and individuals of accurate and reliable valuations of these residential properties cannot be overemphasized. Numerous appeals of valuation inaccuracy have been made over the years with some having been litigated and even considered by higher courts. The contention that valuation is as much an art as a science poses a major challenge for researchers to find ways of improving the accuracy of valuations. Accurate real estate valuation requires complete and accurate data, effective valuation models, and the proper management of resources. Although researchers have paid considerable attention to improving valuation methods, there has been little interest shown in improving the accuracy of property related data. Between the two types of characteristic data, namely objective data and subjective data, the subjective data, which mainly deal with the visual factors, are not easily quantifiable. Previous studies mainly employed dummy variables and viewshed as the proxy variables to describe the quality of the property view. However, these measures are prone to measurement errors, resulting in the inaccuracy of input data. Few studies have investigated improving the accuracy of quantifying views, as the most important visual subjective characteristic data, from the perspective of view quality. Furthermore, while previous studies used 3D modeling techniques to generate 3D city models and conduct viewshed analysis, most of them failed to describe the texture information of buildings and the distribution of vegetation and other obstacles, which significantly affects the credibility of their results. Therefore, how to model the reality to meet the level of detail (LoD) requirement of 3D city models and how to measure scenic view from the perspective of view quality for valuation purposes have remained unexplored. To address the above mentioned research gap, the aim of this research was to examine whether the valuation accuracy of residential property can be improved by using an integrated modeling approach to generate a 3D city model based on a geographic information system (GIS) to measure scenic view from the perspective of landscape visual quality, and from that to create a novel 3D GIS-based valuation system. The specific objectives of this research were to: (1) Generate a 3D city model with sufficient details to meet the requirement of residential property valuation purposes. (2) Propose a framework to support the assessment of landscape visual quality in urban settings. (3) Develop a 3D GIS-based valuation system for assessing scenic view in residential property valuations. (4) Verify and validate the reliability and effectiveness of the proposed 3D GIS-based valuation system in a real case.
The study reviewed previous studies related to measuring views for real estate valuation and assessing landscape visual quality, and summarized the research in this area. The level of detail (LoD) required of 3D city modeling for residential property valuation was analyzed. To meet such a LoD requirement, an integrated 3D modeling approach was proposed to produce a 3D city model. Accuracy assessment was conducted to further validate the integrated modeling approach. Based on a review of relevant literature, a framework for assessing urban landscape visual quality was proposed. A questionnaire survey of urban landscape preference was used to verify the framework. Based on the generated 3D city model, indicators of the framework were quantified. Relationships between these indicators and human urban landscape preferences were estimated and used to make predictions to further validate the proposed framework. A 3D GIS-based valuation system was then proposed by integrating the three modules: the module of generating a 3D city model, the module of quantifying scenic view, and the module of valuing residential properties. To validate the entire 3D GIS-based valuation system, which utilized the integrated modeling approach to generate a 3D city model and the proposed framework to measure urban landscape visual quality, a real case study was conducted using empirical and comparative analyses. The key findings obtained from this study are as follows. First, an integrated 3D modeling approach, which combined the unmanned aerial vehicle (UAV) modeling approach with the procedural modeling approach, was used to generate a 3D city model with LoD4 for further spatial analysis. Results of accuracy assessment show that both the horizontal error and the vertical error of the proposed modeling approach were around 30 cm. Second, a framework for assessing urban landscape visual quality was developed for measuring human landscape preferences. In this framework, nine key visual quality concepts were identified, including naturalness, coherence, disturbance, stewardship, historicity, complexity, visual scale, imageability, and ephemera. Based on the results of an urban landscape preference survey, relationships between the visual concepts and human landscape preferences were estimated. The entire framework was validated by the fact that no significant difference was found between the human scores and the machine scores predicted by the model. Third, a real case study using the 3D GIS-based valuation system to measure scenic view was conducted. Empirical results show that the 3D GIS-based valuation system succeeded in measuring scenic view from the perspective of landscape visual quality. By comparing the performance of the 3D GIS-based valuation system with the performance of conventional methods, it was concluded that the proposed 3D GIS-based valuation system has a higher prediction accuracy, fewer inference biases, and higher uniformity, and performs better overall than the two conventional methods. This study makes an original contribution to measuring scenic view for residential property valuation from both theoretical and practical perspectives. From the theoretical perspective, this study proposes an innovative method to generate a 3D city model with sufficient detail for real estate valuation purposes through integrating the UAV photogrammetric modeling approach and the procedural modeling approach. It also proposes to measure scenic view from the perspective of landscape visual quality and develops a framework to support the assessment of landscape visual quality in urban settings. From a practical perspective, the 3D GIS-based valuation system can increase the accuracy of measuring visual factors to further improve property valuation accuracy. This would benefit the public and private valuation institutes and agencies, as well as other stakeholders such as real estate developers, commercial banks and insurance companies, when making determinations about properties.
|Description:||xix, 183 pages : color illustrations
PolyU Library Call No.: [THS] LG51 .H577P BRE 2019 Huang
|URI:||http://hdl.handle.net/10397/80587||Rights:||All rights reserved.|
|Appears in Collections:||Thesis|
Show full item record
Files in This Item:
|991022210743803411_link.htm||For PolyU Users||167 B||HTML||View/Open|
|991022210743803411_pira.pdf||For All Users (Non-printable)||10.18 MB||Adobe PDF||View/Open|
Citations as of May 21, 2019
Citations as of May 21, 2019
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.