Assessment of models predicting the maximum surface temperature in an urban setting using VIS/NIR/IR imaging from UAVs
MetadataShow full item record
- Master's theses (IMT) 
Investigating surface temperature of the materials in urban settings is an important way of analyzing how urban heat islands can be mitigated. The present thesis investigates models for prediction of the maximum surface temperature of urban materials in Ymittos, Athens, based on image analysis of data collected in VIS, NIR and IR spectrum with a UAV in spring 2016 and summer 2017. The main aim of the thesis is to assess the potential and limitations of three models in predicting daily maximum surface temperature: model of apparent thermal inertia, linear regression of maximum surface temperature as a function of material’s absorptivity and parametrization of maximum surface temperature. The model of apparent thermal inertia captures temperature diﬀerences based on moisture, but more research is necessary to conclude how well it performs on dry surfaces. Linear regression model and parametrization of maximum surface temperature give promising results. Furthermore, it has been shown with statistical methods that asphalt, marble, vegetation and concrete can be identiﬁed based on ﬁve bands: red, blue, green, NIR and NDVI. Since two diﬀerent IR cameras were used in Athens in 2016 and 2017, Optris 640 and Flir V ue Pro, the study is also based on the data collected in laboratory experiments which makes it possible to compare their performance. It has been shown that the former is more reliable and accurate. Correction factor for the latter has been obtained.