双背景模型自适应运动目标检测算法在交通监控系统中的应用研究

Application of Dual Background Model Adaptive Motion Target Detection Algorithm in Traffic Monitoring System

针对现有运动目标检测算法不能满足复杂场景需求,提出一种基于高斯混合模型和时间平均模型改进的双背景模型自适应运动目标检测算法。对视频图像背景进行简单背景和复杂背景自适应判别,并建立相应的背景模型。双背景模型获取的运动目标区域信息更完整、清晰。实验表明,与传统检测算法相比,新算法在去除区域孔洞、目标区域完整性具有较好性能和优越性。

Because the existing moving target detection algorithm cannot meet the complex scene requirements, an improved moving target detection algorithm based on Gaussian mixture model and time-average model is proposed. The algorithm performs simple background and complex background adaptive discrimination on the video image background, and establishes the corresponding background model. The moving target region information obtained by the dual background model is more complete and clearer. Experiments show that compared with the traditional detection algorithm, the new algorithm has better performance and advantages in removing regional holes and integrity of the target region.