Landslides and debris flow events triggered by substantial precipitation in typhoon events often cause enormous damage. It is crucial to rapidly acquire, process, and distribute the most updated information of the disaster areas, particularly optical images with high-spatial resolution. An unmanned aerial vehicle (UAV) provides an innovative approach to remote sensing that is much cheaper, safer, and more flexible for deployment in a small area, ranging from a few tens of square kilometers. This paper demonstrates the application of a low-cost UAV to rapid landslide assessment in difficult access and climatic conditions, as well as presents two examples of rapid response to the natural disaster events at Mudan and Manjhou triggered by Typhoon Nanmadol on August 29, 2011. Various approaches were successfully integrated and implemented in an automatic mission planning and image processing system to plan a flight mission and generate three levels of georeferenced products. Comparing with 14 check points identified on the 25-cm resolution aerial orthophoto, the accuracy of the orthophoto product can achieve a root mean square error of less than 2.81 pixels. All orthophotos are further processed to one seamless, color-balanced, and georeferenced mosaic that can be published on the free-access Google Earth within 24 h.