Moving Object Tracking Algorithm Based on Improved Gaussian Mixture Model
| Autor: | Long Hao, Zhang Shu-kui |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Pixel
Computer science business.industry Frame (networking) ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Mixture model Tracking (particle physics) Interference (wave propagation) Reduction (complexity) symbols.namesake Video tracking symbols Computer vision Artificial intelligence business Gaussian network model |
| Zdroj: | 2019 3rd International Conference on Electronic Information Technology and Computer Engineering (EITCE). |
| DOI: | 10.1109/eitce47263.2019.9094992 |
| Popis: | The moving object image frame has characteristics such as fast change of frame, large moving range, large difference between frames and rapid change of background image, which lead to tracking foreground misjudgment, tracking lag and even losing track objectives. In order to solve this problem, we propose a moving object tracking algorithm based on Gaussian mixture model. By setting up the mixed Gaussian model and updating the dynamic pixel parameters, we eliminate the effect of pixel reduction caused by rapid change of image frame, remove background image interference, and solves the problem of tracking lag. Simulation results show that the proposed algorithm can improve the accuracy and lag of moving objects tracking in complex background and achieve good tracking performance compared with traditional algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|