Object tracking algorithm based on Bayes classification model and Haar-grey features on infrared camera

Abstract

Object tracking based on thermal cameras is a core issue in security monitoring systems. During operation, the size and shape of the object can change continuously, especially thermal imaging objects with significant noise and blurred borders, making it difficult to capture and track. This article proposes a new algorithm based on the Haar-grey features of the object and the Bayesian classification model to track objects on the infrared image background. Experimental results show the effectiveness of the proposed algorithm.