Multisensor imagery analysis for mapping and assessment of 12 January 2010 earthquake-induced building damage in Port-au-Prince, Haiti

摘要

On 12 January 2010, a devastating earthquake of magnitude 7 struck the densely populated Haitian capital city of Port-au-Prince and several other towns along the bay of Port-au-Prince. The severely impacted areas lie in close proximity to the epicentre of the earthquake. The earthquake impacted more than 5 million people, of which more than 230,000 people lost their lives and approximately 1.2 million people were left homeless. In this study, a new multisensor imagery-based analysis approach is presented for detection, delineation, and mapping of collapsed and severely damaged buildings. The thrust of the approach is on improving the accuracy of pixel-based analysis performed on the post-earthquake Advanced Land Imager by utilizing very high spatial resolution WorldView-2 imagery acquired immediately after the earthquake. The study also highlights some of the challenges and constraints of traditional pixel-based analysis performed on medium-resolution imagery for detection, delineation, and mapping of earthquake-induced damage to buildings in a spectrally complex high-density urban environment.