一种自适应运动目标检测算法研究

提出了一种新的自适应目标检测算法,实现了快速对目标区域进行匹配更新,能够自适应精确检测出运动目标。通过改进的混合差分模型确定运动目标最大分布可能区域并融合混合高斯模型对此区域进行背景重建,再运用背景差分得到精确前景目标。实验仿真结果表明能够在复杂场景下克服环境噪声与背景扰动,降低了复杂度和计算开销,具有很好的鲁棒性。

A Study on SeIf-adaptive Moving Target Detection AIgorithm

A novel self-adaptive object detection algorithm is presented in this paper realize quick match updating of the target area so as to accurately detect the moving target in the self-adaptive way. In this approach, an improved hybrid difference model is used to determine to determine the maximal possible distribution area of the moving object, and the hybrid Gaussian model is integrated to rebuild the background of the area. Then, the background difference of this area is used to obtain an accurate foreground object. Experimental simulation results demonstrate that the proposed method can overcome ambient noise and background disturbance under complex conditions and reduce complexity and calculation cost, and has good robustness.