Visual Surveillance of Human Activities: Background Subtraction Challenges and Methods
Thierry Bouwmans, Belmar García-García · 2019
This chapter offers a brief survey of the main visual surveillance applications and challenges of human activities where background subtraction is used to detect static or moving objects of interest. All these applications show the importance of moving-object detection in videos, as it is the first step that is followed by tracking, recognition, or behavior analysis. So, the foreground mask needs to be as precise as possible, as quickly as possible for an issue of real-time constraint. These different real-time applications present several specificities and need to deal with specific critical conditions due to the position of the cameras, the type of the environments, and the type of the moving objects. Because the environments are very different, the background model must handle different challenges following the application.