动态场景中运动目标快速检测算法

Fast Moving Object Extraction Algorithm in Image Sequences

提出了一种基于边界灰度投影匹配的全局运动估计和运动目标提取算法.算法将边界灰度水平投影和垂直投影值作为匹配特征,较好地估计了全局运动参数.由于只需计算一维特征向量所以降低了全局运动估计的计算量.经过全局运动补偿后,可以运用传统的帧间差法得到运动目标.为了减少噪声的影响,准确提取目标,采用了高阶统计量的方法(HOS)来区分背景和运动目标.试验结果证明,所提出的方法在估计全局运动参数和提取运动目标方面有较好的鲁棒性.

A novel and effective approach to global motion estimation and moving object extraction was proposed. Firstly, the translational motion model was used because of the fact that complex motion could be decomposed as a sum of translational components. Then, the edge gray horizontal and vertical projections were used as the block matching feature for the motion vectors estimation in this application. The proposed algorithm reduced the motion estimation computations by calculating the one-dimensional vectors rather than the two-dimensional ones. Once the global motion was robustly estimated, relatively stationary background could be almost completely eliminated through inter-frame difference method. To achieve an accurate object extraction result, the higher-order statistics （HOS） algorithm was used to discriminate backgrounds and moving objects. Experimental results validate that the proposed method is an effective way for global motion estimation and object extraction.