【摘要】 根据目标和背景颜色直方图的特点,针对异色背景干扰和近色背景干扰,提出了一种改进直方图映射和均值移动结合的目标跟踪算法,通过目标主分量提取和干扰分量鉴别,有效地剔除了背景干扰成分,提高了抗背景干扰能力;均值移动算法在生成灰度图中能快速准确定位目标位置。仿真实验结果证明,改进的直方图映射算法能有效地抑制背景干扰,甚至能抑制与目标色调相近的背景干扰,并验证了跟踪算法的实用性和有效性。更多还原

【Abstract】 Object tracking technique based on vision is of wide practical value. Color histogram has good stability and isn,t influenced by the variation of shape and scale of objects; meanwhile the algorithm for mean-shift can get local best key and is speedy and effective. So the algorithm for object tracking using improved histogram back-projection and mean-shift is widely used. One of the difficulties of object tracking is how to suppress the interference of background. According to the characteristic of object and the color histogram of background, this paper proposes an algorithm for object tracking using improved histogram back-projection and mean-shift which eliminates the interference elements of background by distilling the main elements of object and distinguishing interference elements and improves the ability of anti-jamming. Mean-shift rapidly locates the object in the value images. Contrastive experimental results show that the improved algorithm can suppress the interference of background which has different or similar color with tracked object and validate the practicability and validity of this algorithm.更多还原