Efficient Target Detection from Infrared Image Sequences Using the Sequential Monte Carlo Method

This paper presents an efficient target detection algorithm from a sequence of infrared images using the sequential Monte Carlo method (SMC). The algorithm employs an evolution process of the particles which correspond to the candidates of the target position and whose evolution is controlled by the weight of the target feature. Through the iterative process on the differential images, a valve of the particle set is proposed to decide if there is a target in the image, and the state of the particle set is used to position the target. The experimental results demonstrated that the algorithm can detect the target with sea-sky background effectively regardless of the existence of serious non-Gaussian noises. The experiments also showed real-time efficiency of the algorithm for target detection