Infrared Target Tracking Algorithm Based on Bernoulli Filter and Support Vector Machine

Abstract

Infrared small target tracking algorithm is an important method to obtain the target states effectively. Support Vector Machine (SVM), a classic method in machine learning methods, is effective for target recognition. The Bernoulli filter is the optimal single-target filter for target tracking within the Bayesian framework. However, there is no related research on the joint application of the SVM and Bernoulli filter for infrared small target tracking. In this paper, the SVM and Bernoulli filter are jointly applied to the tracking and recognition of small targets in complex backgrounds. The research shows that, with the SVM, Bernoulli filter can indeed be applied to the actual multi-frame infrared target tracking scene to achieve effective tracking and recognition of infrared small targets.