In this work, we describe a framework to analyze UAV videos content. A multi-class image segmentation approach is proposed considering UAV videos specific characteristics. A static image segmentation is applied on each frame. After a preprocessing step on resulting segments, a SVM classifier is used to recognize regions. A Markov model is introduced to combine the results from the previous frames in order to improve the accuracy. The framework has been designed to be as flexible as possible with an eye to allow to insert holistic information into the model.
展开▼
