An Adaptive Threshold Algorithm for Moving Object Segmentation
Connected region detection is usually used to obtain foreground regions from foreground image after moving object detection. In order to remove noise regions and retain true targets, a threshold that limits the circumference of foreground regions should b
- PDF / 1,988,480 Bytes
- 10 Pages / 439.37 x 666.14 pts Page_size
- 77 Downloads / 261 Views
stract. Connected region detection is usually used to obtain foreground regions from foreground image after moving object detection. In order to remove noise regions and retain true targets, a threshold that limits the circumference of foreground regions should be introduced. The method which uses the same threshold for all surveillance videos cannot handle scene changes. In this case, we propose an adaptive threshold algorithm for moving target segmentation. A strategy based on the combination of background modeling and Grabcut is presented to extract foreground objects and set an initial threshold. On the base of this, we can choose some foreground as samples and classify them by K-means clustering method. Finally, an appropriate threshold could be selected for moving object segmentation according to the classification result. Experimental results show that the proposed method has strong adaptability to various scenes and improves the accuracy oftarget segmentation. Keywords: Moving object segmentation · Adaptive threshold · K-means clustering · Image segmentation
1
Introduction
In recent years, moving object detection [1-2] has been an important topic in the field of computer vision. As an essential step of intelligent surveillance technology, its results are crucial for subsequent advanced processing, such as object classification, tracking and activity analysis. After the detection of moving targets we can obtain foreground images. In order to acquire the features of moving foreground objects further and facilitate subsequent tasks, it is necessary to separate motion areas from background. Since the results of detection is sensitive to sensors noise and background changes, pixels that belong to background are often falsely detected as foreground and, as a result, the performance of foreground object segmentation are heavily affected. To handle this problem, it is need to include a threshold to limit the circumference of foreground regions so that fake targets mixed with moving objects could be judged and rejected with the help of this threshold. The performance of threshold algorithm determines the accuracy of object segmentation. A fixed threshold manually selected is usually not applicable to all video scenes, so it is worth to study the algorithm of adaptive threshold. In this paper, we © Springer-Verlag Berlin Heidelberg 2015 H. Zha et al. (Eds.): CCCV 2015, Part I, CCIS 546, pp. 230–239, 2015. DOI: 10.1007/978-3-662-48558-3_23
An Adaptive Threshold Algorithm for Moving Object Segmentation
231
put forward an adaptive threshold algorithm to get motion areas apart, which mainly focus on three issues: how to select an initial threshold, how to classify foreground regions and how to adjust the threshold based on concrete situations. Among these problems, the selection of initial threshold is the foundation of other issues. It relates to the sample collection of K-means clustering method and therefore has a great effect on the quality of classification. To reasonably determine the initial threshold, a nove
Data Loading...
