Robust Dynamic Object Recognition under Various Environments Using GMM

This research paper introduces a confidence measure scheme to detect the vehicles under various natural environmental scenes. In elaborate city traffic scenes, the GMM (Gaussian mixture model) is designed for vehicle detection. Using different climatic conditions, the experiments are conducted by Gaussian mixture model with parameter measurement. The measured parameters are compared with adaptive background learning algorithm. The experimental results shows that GMM excels adaptive method by staying the background model being uncontaminated from slow-relocating or briefly stopped cars.