Phương pháp phát hiện dị thường trên ảnh viễn thám quang học ứng dụng trong công tác tìm kiếm cứu nạn

Abstract

Research on human survival after a plane crash shows that victims are 10% less likely to survive if the rescue is delayed by more than 2 days, and the survival rate is up to 60% if the rescue is made timely within 8 hours (Xuân Đông, 2014). The same urgency also applies to maritime emergency situations or on land. Therefore, the time to find victims and rescue organizations is a decisive factor for the success of that campaign. To reduce the search time, an increasingly commonly used approach is to detect anomalies in high-resolution remote sensing images. In addition, the size of the missing person or object of interest is very small compared to the scene and is easily mixed with the terrain. Therefore, it is necessary to have methods to automatically locate objects that help improve the performance and speed of searching. In this paper, several methods of detecting anomalies in remote sensing images will be presented to solve the problem mentioned above.