Papers by Keyword: Background Updating

Abstract: The paper proposes a new method for moving objects detection based on fusion of three frames differencing and Gaussian Mixture Model (GMM). In the method, two images are obtained by three frames differencing, then the adaptive background are modeled and updated by GMM for each pixel in the two differencing images. Next, two differencing images are done logic "and" operation to get the shape of the moving object. Finally adopt the mathematical morphology operation to eliminate noise and the small areas of non-objects motion parts. The simulation results show that the proposed method can detect the objects effectively and real-time. So it can be applied in visual surveillance system effectively.

Abstract: Aimed at solving the problem of mold damage caused by a foreign body in the mold before mold clamping, this paper proposes a solution, which applies image processing technology such as background updating and the difference image algorithm to solve it. Not only can it judge whether there is a foreign body in the mold but it can also detect whether the product is perfect by comparing the foreground image with the background image at the appropriate time (before mold clamping or after mold opening) and by calculating the qualified rate of pixel in all ROIs (Region of Interest). To eliminate the influence of vibration and of changes in brightness in the surrounding environment on the detecting results, this paper utilizes near infrared illumination technology and the background updating algorithm. In addition, the ROI is set to improve the detecting speed and accuracy.

Abstract: Real-time segmentation of moving regions in image sequences is a fundamental step in video monitoring systems. This paper presents an improved motion detection algorithm in a dynamic scene based on change detection. The algorithm integrates the temporal differencing method and background subtraction method to achieve better performance. Background subtraction is a typical change detection approach to segment foreground, but the continuous or abrupt variations of lighting conditions that cause unexpected changes in intensities on the background reference image. So we combine the background subtraction’s result with temporal difference’s result. The foreground mask is segmented by both the methods of background subtraction and temporal differencing. Finally, a post-processing is applied on the obtained object mask to reduce regions and smooth the moving region boundary. Experimental results demonstrate that the proposed method can update the background exactly and quickly along with the variation of illumination, and the moving objects can be extracted effectively.

Abstract: A novel approach to video detection, synchronous detection lines method, is firstly proposed in this paper. It is actually a way to divide the pixel sets of the video images. Then, an algorithm of three-layer probability model based on synchronous detection lines is presented to solve the problems of background updating and target tracking. The results of a practical system show that this approach has better effect to solve the problems of background updating and target tracking.