DETECTION APPARATUS AND METHOD FOR OBJECTS IN THE IMAGE USING SPATIO-TEMPORAL GAUSSIAN MIXTURE BACKGROUND MODEL

摘要

PURPOSE: A device for detecting an object in an image by using a spatial-temporal Gaussian mixture background model and a method thereof are provided to efficiently perform background modeling by using a temporal Gaussian mixture modeling method in a static image and to perform background modeling by using a spatial Gaussian mixture modeling method in a dynamic image, thereby supplying improved effect in the background modeling of the dynamic image. CONSTITUTION: A Gaussian modeling application module(130) distinguishes each pixel of an input image. The Gaussian modeling application module performs background modeling for an unstable pixel corresponding to a spatial Gaussian modeling method and performs the background modeling for a stable pixel corresponding to a temporal Gaussian modeling method. A temporal Gaussian modeling module(140) and a spatial Gaussian modeling module(150) output a pixel value by performing the background modeling with the control of the Gaussian modeling application module. [Reference numerals] (110) Image input module; (120) Pixel value setting module; (130) Gaussian modeling application module; (140) Temporal Gaussian modeling module; (150) Spatial Gaussian modeling module; (160) Median filter module