Moving Object Tracking Based on Infrared Thermal Imager and Vision Sensor

In this paper, we propose a moving object tracking method that fuses infrared thermal imager and vision data for solving the limitations of single mode image information. First, contour feature matching is realized by the proposed shape context and Hog feature on the basis of foreground detection by Gaussian Mixture Model. Second, the regularization and scaling characteristics are used to reorganize the corresponding relation and estimate the transformation. In order to extend match to the whole shape, a TPS transformation model is established. A regularization method is used to relax the interpolation requirements, and the estimation error is reduced by iteratively reorganizing corresponding relationship and estimating the transformation. Finally, moving objects are determined by adaptive background subtraction. In complex and changeable lighting and meteorological conditions, the application of vision sensor has certain limitations, and multi-sensor fusion can realize information complementarity, this method combines edge and contour feature information, with better fault tolerance, real-time and robustness than a single sensor.