An Algorithm of Occlusion Detection for the Surveillance Camera

As more and surveillance cameras are deployed in the Internet of Things, it takes work to ensure not occluded. An algorithm detecting whether camera is occluded proposed by comparing similarity images this paper. Firstly, background modeling method based on frame difference improved. The combination proposed, experimental results showed that can extract image video quickly accurately. Secondly, LBP (Local Binary Patterns) used compare between reference image. By changing window size setting an appropriate threshold, actual demands be satisfied. So, algorithms paper have high application value practical significance.