Improvement Algorithm of ViBe in Moving Target Detection
Aiming at the problem that Visual Background Extractor（ViBe） algorithm is not sensitive to light mutation and is mistakenly detected as foreground due to a large number of high-frequency disturbed objects in the background, an improved method is proposed based on the classical ViBe algorithm. The me...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | Chinese |
| Published: |
Harbin University of Science and Technology Publications
2022-02-01
|
| Series: | Journal of Harbin University of Science and Technology |
| Subjects: | |
| Online Access: | https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2052 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Aiming at the problem that Visual Background Extractor（ViBe） algorithm is not sensitive to light mutation and is mistakenly detected as foreground due to a large number of high-frequency disturbed objects in the background, an improved method is proposed based on the classical ViBe algorithm. The mean value of V component in HSV space of separated video frames is used as the method to judge the light mutation, and the threshold to judge the light mutation is obtained by experiment Δω=0.1 as the discrimination condition, if the illumination changes suddenly, the video frame with stable mutation is used as the initial frame for re modeling, so as to solve the problem of illumination change. Because the moving target will produce "ghost" during the initial modeling of the algorithm, the method of two consecutive frames of differential background model is used to remove the "ghost". And, come up with a concept that can judge the degree of perturbation, so that the matching radius of background modeling is dynamically changed with it. Experiments show that the percentage of wrong classifications of the improved ViBe algorithm is about 6% lower than that of the original algorithm and can achieve the effect of real-time detection. |
|---|---|
| ISSN: | 1007-2683 |