Assessing Digital Surface Models by Verifying Shadows: A Sensor Network Approach

We propose to use wireless sensor networks to assess the accuracy and application of Digital Surface Models (DSM) for the study of shadowing and solar radiation over the built environment. Using the ability of sensor network data to provide information about solar radiation and predicting the exact time of the day that the Sun starts radiating a sensor, a comparative study and statistical analysis can be undertaken in order to evaluate the accuracy of the DSM for shadowing and radiation studies using image processing techniques. Two DSMs of the EPFL campus with different cell resolutions (1 meter and 0.5 meters), considering only information about ground, buildings with vertical walls and trees, are constructed step by step and employed. Three DSMs of the same campus with a cell resolution of 1 meter derived from raw LIDAR data and common interpolation techniques, such as Triangulated Irregular Network (TIN), kriging, Inverse Distance Weighting (IDW), are also used for comparison.