基于五帧差和二维Renyi熵的运动目标检测

针对传统帧间差分法存在的不足,提出改进五帧差分法和二维Renyi熵阈值分割法相融合的运动目标检测算法。该算法充分考虑视频序列帧间的时空相关性,利用改进五帧差分法对预处理后的视频图像序列进行差分运算,以提取运动目标的帧间时间相关性;将得到的差分序列使用二维Renyi熵阈值分割法分割处理,以获取运动像素与同帧周围像素的空间相关性。该算法避免了五帧差分法轮廓提取的空间信息丢失,克服了二维Renyi熵阈值分割法容易忽略运动像素的时间信息,从而提升了运动目标检测的完整性。实验结果表明,该算法能够更准确地检测出运动目标,具有较好的鲁棒性和有效性。

Moving object detection based on five frame difference and two dimensional Renyi entropy

Aiming at the existing problem of the traditional frame difference algorithm, a moving object detection algorithm based on improved five frame difference and two dimensional Renyi entropy is proposed in this paper. The algorithm considers fully the spatial temporal correlation between video frames, and extracts firstly the temporal correlation of moving target between frames by carrying on improved five frame difference operation for video image sequences, then obtains the spatial correlation between moving pixels and surrounding pixels by carrying on two dimensional Renyi entropy threshold segmentation for difference sequences. The algorithm avoids loss of the spatial information by five frames difference contour extraction, and overcomes the two dimensional Renyi entropy threshold segmentation easy to overlook the time information of moving pixels, so as to enhance the integrity of the moving target detection. The experimental results show that the algorithm can more accurately detect moving objects, and has better robustness and effectiveness.