Search

Distributed solar photovoltaic (PV) harvesting is an effective way to collect solar energy in existing metropolitan cities. To facilitate the installation of PV modules at solar abundant locations, an accurate estimation of solar PV spatial potential is indispensable. Solar energy could be reflected on high-albedo building surfaces inside the urban canyon. However, using conventional ways to construct albedo datasets for different building surfaces is extremely labor-intense. In this work, we proposed to use semantic segmentation to identify façade materials from street view images. Due to the distinguishable features between materials in terms of the subtle texture and patterns rather than just their shapes and colors, identification requires more details from images, which makes multi-scale inference structure a promising solution. Compared with existing methods combining scales features at pixel-level, we proposed a novel Multi-Scale Contextual Attention Network (MSCA), using a Multi-Scale Object-Contextual Representation (OCR) block to exploit and combine contextual information from different scales in high dimensional layers. The experimental results show that the proposed model significantly outperforms the existing models, achieving a mean Intersection over Union (mIOU) of 70.23%. The results indicate that the MSCA can effectively obtain the materials information from street views and can be a reliable solution to providing urban albedo information for solar estimation. [ABSTRACT FROM AUTHOR]

This study aims to evaluate the effects of urban geometry on retrieval of emissivity and surface temperature in urban areas. An improved urban emissivity model based on sky view factor (IUEM-SVF) was further enhanced to consider all radiance contributions leaving the urban canopy, including (i) emission by all facets within an instantaneous field of view (IFOV); (ii) reflection by all facets of emission from surrounding facets; and (iii) propagation of emitted and reflected radiation with multiple reflections (scattering) within a complex 3D array of urban objects. The effective emissivity derived from IUEM-SVF was evaluated with a microscale radiative transfer and energy balance model: Temperatures of Urban Facets in 3-D (TUF-3D). IUEM-SVF performs well when urban facets have uniform emissivity and temperature; e.g., root mean square deviations (RMSD) are less than 0.005 when material emissivity is larger than 0.80 (ɛ ⩾ 0.80). However, when material emissivities are variable within the observed target, differences of effective emissivity between IUEM-SVF and TUF-3D become larger, e.g., RMSD of 0.010. When the effect of geometry is not considered and a mixed pixel emissivity is defined, the difference is even much larger (i.e. 0.02) and this difference increases with the decrease of sky view factor. Thus, the geometry effect should be considered in the determination of effective emissivity. Effective emissivity derived from IUEM-SVF was used to retrieve urban surface temperature from a nighttime ASTER thermal infrared image. Promising results were achieved in comparison with standard LST products retrieved with the Temperature and Emissivity Separation (TES) algorithm. IUEM-SVF shows promise as a means to improve the accuracy of urban surface temperature retrieval. The effect of thermal heterogeneity on the effective emissivity was also evaluated by TUF-3D, and results show that the thermal heterogeneity cannot be neglected since the RMSD between the effective emissivity based on TUF-3D and IUEM-SVF is relatively large. [ABSTRACT FROM AUTHOR]

This study presents a Single Channel Method using Urban Exitance Model (UEM-SCM) to retrieve land surface temperature (LST) from satellite data in an urbanized city, and evaluates the geometry effect on land surface temperature retrieval using single channel method and split-window algorithm. The UEM-SCM incorporates the effect of urban geometry and considers both reflection caused by the target pixel and its neighboring pixels. In order to evaluate the geometry effect, the retrieved LSTs with and without geometry effect were studied. Results show that the LSTs without geometry effect are generally higher than the LSTs with geometry effect. The temperature difference occurs because the material emissivity is always lower than the effective emissivity caused by multiple scattering and reflection in urban areas (cavity effect). The LST without geometry effect also cannot fully capture the variability and complexity of urban thermal patterns. The temperature difference between with and without the geometry effect can reach 2 K in built-up areas. A comparison was also conducted between LST retrieved by split-window algorithm with and without geometry effect. Results show that the LST retrieved by split-window algorithm without geometry effect has generally higher values than the one with the geometry effect, e.g. 1.1 K on average and 1.5–2 K in built-up areas. The geometry effect will be removed and mis-deemed as atmospheric effect when the split-window algorithm without geometry effect is applied in urban areas. The split-window algorithm with the geometry effect can be used to distinguish between geometry and atmospheric effect in further study. [ABSTRACT FROM AUTHOR]

Surface emissivity is a critical parameter for studying city-, meso-, and micro-scale climate and energy balance. The emissivity of complex surfaces e.g. a forest or an urban canopy is an effective surface property since it depends on both surface materials and geometry. This study presents a novel methodology for estimating effective emissivity using sky view factor retrieved from airborne Lidar data, building GIS data, and land use and land cover classification data. First, a high correlation between the effective emissivity retrieved from ASTER TIR bands 10–14 and the sky view factor was observed ( r 2 = 0.93, 0.99, 0.99, 0.97, 0.97). When the sky view factor decreases, the effective emissivity tends to increase, which is mainly due to multiple scattering (cavity effect), thus increases the effective emissivity. A simplified model which assumes that reflection and scattering only occurs within a single pixel was developed. Results showed that the correlations between the modeled and the spectral (band) emissivity retrieved from the ASTER multispectral TIR data (five spectral bands) are high ( r 2 = 0.93, 0.99, 0.98, 0.93, 0.97), and with low RMSE (0.019, 0.016, 0.012, 0.003 and 0.004 from band 10–14 respectively). The emissivity derived from this simplified model, however, tends to be overestimated in band 10–12. Thus, a refined urban emissivity model based on sky view factor (UEM-SVF) which considers the scattering and reflection from adjacent pixels was developed in this study. Results show a good agreement with ASTER spectral (band) emissivity: r 2 = 0.90, 0.98, 0.96, 0.94 and 0.96, and very low RMSE (0.006, 0.003, 0.004, 0.002 and 0.004). This study illustrates that the UEM-SVF can be useful for estimation of land surface emissivity of complex surfaces, and can further be used for accurate land surface temperature retrieval. [ABSTRACT FROM AUTHOR]

• Distinguish urban tree species with terrestrial hyperspectral camera (SPECIM IQ). • 450 bi-monthly images were acquired for 75 urban trees belonging to 19 tree species. • Hyperspectral phenology models were developed to explore seasonalities of trees. • A Deep Neural Network is used to identify tree species from hyperspectral imagery. Urban trees exhibit a wide range of ecosystem services that have long been unveiled and increasingly reported. The ability to map tree species and analyze tree health conditions would become vividly essential. Remote sensing techniques, especially hyperspectral imaging, are being evolved for species identification and vegetation monitoring from spectral reponse patterns. In this study, a hyperspectral library for urban tree species in Hong Kong was established comprising 75 urban trees belonging to 19 species. 450 bi-monthly images were acquired by a terrestrial hyperspectral camera (SPECIM-IQ) from November 2018 to October 2019. A Deep Neural Network classification model was developed to identify tree species from the hyperspectral imagery with an overall accuracy ranging from 85% to 96% among different seasons. Representative spectral reflectance curves of healthy and unhealthy conditions for each species were extracted and analyzed. The hyperspectral phenology models were developed to achieve high accuracy and optimization of data acquisition. The bi-monthly canopy signatures and vegetation indices revealed different seasonality patterns of evergreen and deciduous species in Hong Kong. We explored the utility of terrestrial hyperspectral remote sensing and Deep Neural Network for urban tree species identification and characterizing. This provides a unique baseline to understand hyperspectral characteristics and seasonality of urban tree species in Hong Kong that can also contribute to hyperspectral imaging and database development elsewhere in the world. [ABSTRACT FROM AUTHOR]