Moving Objects Detection Based on RGB Color Space and Background Subtraction
Long Chen · Jisuanji fangzhen · 2008
Moving objects detection and segment is a fundamental and critical problem in computer-vision field. Background sub-traction is a classical and widely used method and its difficulties are background reconstruction and update. So an algorithm for moving objects detection based on RGB color space and background subtraction is proposed in this paper. The background is first reconstructed using an image sequence of N frames. The range of RGB values at each pixel in an image sequence is calculated, and the middle values are considered as the RGB values of background. The leaning rate is used for background update. Then the fore-ground objects are obtained in RGB color space. Finally, morphologic operation and connected component measurement are applied to the results to eliminate the wrong detection. Experimental results show that the proposed algorithm combines the advantages of veracity and runtime, and fit for real time detection.