ALBERT

Description: Research in change detection from optical satellite data is widely investigated as a support for visual image analysis. Most of the methods, however, are based on radiometric changes and are suffering from high false alarms rate due to irrelevant radiometric changes. Change detection based on the elevation difference between two dates, therefore, seems a good alternative to identify relevant changes, especially in a context of urban change detection. In the present work, we provide a fully automatic method of change detection based on a digital surface model (DSM) comparison. The processing flow includes the bundle block adjustment of all the available data as a preprocessing step, followed by an improved DSM generation scheme and a differential DSM analysis. The last two steps have been formulated as labeling problems and solved by an optimization method with a spatial regularization constraint. The solution of these labeling problems is computed with a generalized dynamic programming algorithm that is adapted according to the input data and the defined labels. The final DSMs reach a planimetric and altimetric resolution of about 1 m, allowing changes from $bf 20nbsphbox{m}^{2}$ to be detected. The results show that 33%–75% (respectively about 95%) of all changes (respectively, changes larger than $bf 100nbsphbox{m}^{2}$ ) are detected, depending on the employed regularization and the area. Moreover, the calculated kappa coefficient of the processing flow reaches up to 0.80, which emphasizes the method accuracy. All the above features lead to a significant gain compared to the classical visual image analysis.

Description: This paper investigates the potential effects of spectral, shape, textural, and height information and their combinations on the classification of urban areas using airborne hyperspectral data. Based on analysis of the spectral, shape, textural, and height characteristics of urban land covers, the first ten spectral principal components, eight shape components, one height component, and seven textural components were selected to examine their performance on the classification accuracy. Correlation analysis was conducted to exclude correlated components. A support vector machine (SVM) was employed to determine the significant components affecting the urban hyperspectral classification through comparison of the classification accuracy. Different combinations of these components were then tested to estimate their contributes. The classification results showed that all these components contribute to the result of urban land cover classification, but different land cover classes benefit from the inclusion of different components. The experiment further revealed the effect of significant components on the classification of urban land cover in terms of area, convexity, elongation, form factor, rectangular fit, roundness, textual factors, and mean relative height. It is suggested that the inclusion of shape, texture, and height, together with the spectral components, significantly improved the classification accuracy of urban land cover.

Description: This paper attempted to evaluate the effects of urban growth on the land surface temperature (LST) in Beijing, China. Firstly, the multi-temporal land use types were derived based on the images of Landsat 5 TM and Landsat 7 ETM+ from 1996 to 2010, and two types of land use change trajectories (existing urban trajectory and urban growth trajectory) were also established to category the land transformation type under urbanization. Then the algorithm developed by the Landsat project Science Office was adopted to retrieve LST, in which the digital number (DN) of Band6 was converted to the spectral radiance, followed by being converted to the brightness temperature under the assumption of unity emissivity, and corrected for emissivity $(varepsilon)$ to obtain the natural surface temperature. Finally, to evaluate the effects of urban growth on the LST change, two indices of deviation degree and LST change contribution were built and calculated for each trajectory. The results demonstrated that the average deviation degree of urban growth trajectories was much greater than that of the existing urban trajectory, which presented a more violent LST change. However, the effects of urban growth on the LST change of the study area were very limited because the smaller areas of the urban growth trajectories would generate a limited LST change contributions. The results could not only provide a new way and methodology to assess the relationship between urban growth and the LST change, but also support the management and planning in the urban area.

Description: The seasonal and spatial variability of surface heat fluxes is crucial to the understanding of urban heat island phenomenon and dynamics. To estimate energy fluxes, remote sensing derived biophysical variables need to be integrated with surface atmospheric parameters measured in meteorological stations or in situ field measurements. In this study, based on the two-source energy balance algorithm, we applied a method to estimate surface energy fluxes by combined use of multispectral ASTER images and routine meteorological data, and applied it to the City of Indianapolis, United States, aiming at in-depth understanding of the spatial patterns of energy fluxes. By computing the fluxes by land use and land cover (LULC) type, we further investigated the spatial variability of heat fluxes. Results show that the energy fluxes possessed a strong seasonality, with the highest net radiation in summer, followed by spring, fall and winter. Sensible heat flux tended to change largely with surface temperature, while latent heat was largely modulated by the change in vegetation abundance and vigor and the accompanying moisture condition. The fluctuation in all heat fluxes tended to be high in the summer months and low in the winter months. Sensible and latent heat fluxes showed a stronger spatial variability than net radiation and ground heat flux. The variations of net radiation among the land use/cover types were mainly attributable to surface albedo and temperature, while the within-class variability in the turbulent heat fluxes was more associated with the changes in vegetation, water bodies, and other surface factors.

Description: The morphological hit-or-miss transform (HMT) was found to be efficient for the detection of buildings in panchromatic bands of very high spatial resolution images. The use of multispectral information was judged to be necessary to improve the results. The application of morphological operators to multispectral images is problematic, as no universal strategy for ordering the multivalued pixels of these images has been widely adopted. In this paper, we propose a new method to detect building locations based on a recently developed concept for the HMT to handle noise, called percentage occupancy HMT (POHMT). The parameters for the POHMT were defined with the aid of the top-hat by reconstruction transformation. To eliminate irrelevant locations, we applied a vegetation mask and verified locations by their proximity to shadows. The novelty of the method consists in the proposed vector-based strategy that allows for the application of the POHMT to multispectral images in order to detect building locations. Moreover, an original technique to automatically define the parameters for the POHMT was proposed. The method was tested on subsets from a pan-sharpened Ikonos image and from raw GeoEye and WorldView-2 images. The experimental results are promising.

Description: In this paper, extensive real-data experiments for the investigation of the benefit of exploiting multiple aspects and multiple baselines for the reconstruction of urban surface models by synthetic aperture radar interferometry are documented. These experiments are carried out within a recently proposed reconstruction framework that allows the fusion of almost arbitrary configurations of multi-aspect multi-baseline InSAR data. The results based on airborne decimeter-resolution millimeterwave imagery prove and quantify that multiple baselines help to solve the phase ambiguity problem, while multiple aspects reduce the parts of the scene affected by radar shadowing. In addition, the inherent redundancy provides a significant improvement in the achievable reconstruction accuracy, which is evaluated relative to the reconstruction error common for conventional single-aspect single-baseline SAR interferometry.

Description: Monitoring and nowcasting of urban air temperatures are of high interest for prediction of heat stress in cities. Routine observation is so far limited by the complex coupling between atmosphere and land surface in urban areas, which makes estimation more difficult. In this study, we have investigated the capability of multitemporal land surface temperatures (LSTs) from the geostationary Spinning Enhanced Visible Infra-Red Imager instrument for estimation of urban air temperatures. The results are very promising with root-mean-square errors (RMSEs) of 1.5–1.8 K for six stations in Hamburg and explained variances of 97%–98%. Both the annual and diurnal cycles were well represented by the empirical models and the use of multitemporal data substantially increased the model performance. Further, the model was run in a forecast mode without actual LST information. Here, the best predictors reached RMSEs of 1.9–2.4 K and R $^{2}$ of 95%–97% for a 2-h forecast.

Description: The nDSM is commonly used in land cover classification with different heights. However, it is difficult to classify vegetation accurately from other land covers, such as buildings, with only the height information. Meanwhile, traditional remotely sensed imageries have difficulties discriminating different urban vegetation components such as trees and shrubs. Therefore, it is essential to combine both remote sensing imagery and height information obtained from Light Detection and Ranging (LiDAR) data for classification of detailed vegetation components. In this paper, a two-phase classification method is proposed to fuse the airborne LiDAR data and aerial photography imagery to obtain detailed urban vegetation classification map. The first step is to distinguish vegetation from buildings, bare ground, and shade. In this step, two different fusion approaches and two classification methods were used, and the result with the highest accuracy for vegetation classification was selected as the input of the second step. The second step is to output the classification map of vegetation class into vector polygons and utilizes them to separate the vegetation LiDAR points from the nonground points. Then tree, shrub, and lawn points can be easily classified from the vegetation points due to their different heights. The proposed method yielded a classification result with an overall accuracy of 83.39% and a kappa coefficient of 0.79. Moreover, the producer accuracies of vegetation class (tree, shrub, and lawn) are 95.20%, 61.66%, and 79.35%, respectively.

Description: Urban areas are often characterized by geometrically simple and repetitive patterns, in particular, in cases where settlements have been built-up from scratch in a well-planned manner, e.g., according to architectonic, economic, or sociopolitic constraints. This leads to preferred rectangular and regular alignment of objects like windows or balconies at façades for the majority of buildings in modern cities. In this paper, we show how this regularity can be exploited for the challenging task of synthetic aperture radar (SAR) scene description and demonstrate the applicability in case studies. We present a virtually parameter-free method to segment a persistent scatterer point cloud and fit optimal lattices to describe separate façades. Formulating the PS as nodes in a graph allows us to use spectral graph theory to distinguish lattices even when they are overlapping or disturbed due to layover. As a result, we obtain an object-based representation of the SAR data, which allows for many new applications in the field of building monitoring and change detection in urban areas.