Moving Objects Detection and Tracking in Infrared or Thermal Image

Moving objects detection and tracking in infrared images is an important goal in most of the practical applications of the thermo vision systems. For these thermo vision applications here is proposed to apply a cost function associated with the minimization of a global criterion for simultaneous estimation of the optical flow and detection of the moving objects in infrared images. The optical flow and moving objects detection and tracking in infrared images are modeled with an appropriate neural network. The thermo vision or infrared images, captured from thermo camera, are first partitioned in rectangular blocks. The blocks are described with a number of parameters placed in the corresponding feature vectors. It is proposed to apply as parameters of the blocks the following important in thermal images characteristics: the position, the gray level and the local motion information. It is chosen the classification of the feature vectors by considering the displaced frame difference, according to Bayesian theory of decision criterion and representing a metric in the parameter space. Key-Words: thermal images; infrared images; real time objects detection;