The Behavior Understanding of Moving Objects Based on Vision

The moving detection and understanding are two important parts of intelligent monitoring. Firstly, the initial background was created by mean algorithm, then the Self Adapted Background Subtraction Algorithm was used to detect the moving object, and the Gaussian Hermite Moment, the centroid and area feature were extracted, and the tracking algorithm of moving object was developed with the help of these features. The classification of moving object was realized with the help of the feature and the Adaptive Resonance Theory (ART). Finally, we can understand the behavior of moving objects based on the warning line, understand the moving objects based on ART, and understand the behavior of moving objects based on trajectories. The experiment shows that the algorithm was effective.