Moving object detection in video sequence images based on an improved visual background extraction algorithm
- PDF / 4,848,687 Bytes
- 22 Pages / 439.37 x 666.142 pts Page_size
- 39 Downloads / 227 Views
Moving object detection in video sequence images based on an improved visual background extraction algorithm Junhui Zuo 1 & Zhenhong Jia 1 & Jie Yang 2 & Nikola Kasabov 3 Received: 13 June 2019 / Revised: 4 July 2020 / Accepted: 4 August 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Visual background extraction is a method for detecting moving objects in video sequence images. In the traditional visual background extraction algorithm, ghost phenomena and dynamic background interference exist in the detection results. In order to speed up ghost removal and suppress the interference of dynamic background, an improved visual background extraction algorithm is proposed. In this method, secondary judgment is added to eliminate ghost pixel interference in the process of spatial transmission of pixels, and the flicker degree of pixels is analyzed to suppress the interference of dynamic background pixels. In order to improve the detection effect, the edge of moving object is obtained by edge detection method, then filled and fused with the detected object. Finally, the detection of moving object is optimized by means of median filtering and mathematical morphology. The simulation results show that the improved algorithm accelerate ghost removal, effectively suppresses the noise interference caused by dynamic background, and improves the accuracy of moving object detection. Keywords Moving object detection . Visual background extraction . Ghost removal . Dynamic background . Edge detection
* Zhenhong Jia [email protected]
1
College of Information Science and Engineering, Xinjiang University, 666 Shengli Road, Tianshan District, Urumqi, China
2
Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University, 2-427 SEIEE Building, 800 Dong Chun Road, Shanghai, China
3
Knowledge Engineering and Discovery Research Institute, Auckland University of Technology, Private Bag 92006, Auckland, New Zealand
Multimedia Tools and Applications
1 Introduction Moving object detection is a research topic of interest in the field of computer vision. In recent years, in tandem with the development of video detection, the technology of extracting moving objects from video has improved and is widely used in image segmentation [17, 18], object tracking [7, 20], and behavior analysis [8, 14].The purpose of moving object detection is to segment moving objects accurately from video sequence images. Accurate detection of moving objects is also the basis for follow-up object tracking and behavior analysis. At present, the main moving object detection methods are the optical flow method [33], the interframe difference method [15, 28], and the background subtraction method [16, 24]. The various research objects require new research methods such as pedestrian detection [4], face detection [22], and motion blur detection [23]. In view of existing problems in moving object detection, researchers at home and abroad have put forward many new research methods. For salience attention based o
Data Loading...