摘要
针对摄像机在运动情况下运动目标检测问题的研究,提出了一种尺度不变特征变换图像分块特征匹配和局部差分算法相结合的运动目标检测方法。该方法不仅可以减少图像配准参数求解所需要的时间,弥补全局运动补偿对图像完整性的影响,并且可以在摄像机运动的情况下准确地检测出运动目标。实验证明了该方案的有效性。
To investigate the moving objects in the situation of camera moving, a method that combines the scale invariant feature transform block feature matching with the local differential algorithm is proposed. The method can not only reduce the cost time of image registration and compensate the global motion compensation, which effects on the integrity of the image, but also detect moving objects accurately. Finally, the experiment results can provide its effectiveness.
出处
《计量学报》
CSCD
北大核心
2015年第4期352-355,共4页
Acta Metrologica Sinica
基金
国家自然科学基金(61473248)
河北省自然科学基金(F2012203186)
关键词
计量学
目标检测
图像分块
尺度不变特征变换
图像配准
局部差分
Metrology
Object detection
Image blocking
Scale invariant feature transform
Image registration
Local differential algorithm