基于小波图像预处理的运动目标检测方法

Moving Object Detection Methods Based on Wavelet Image Preprocessing Technique

目前,动态目标识别技术难以满足实时性要求的难点之一是需处理海量数据。针对机器人主动避让问题,选择小波变换中的成熟算法对图像进行压缩和分割处理,提出一种简单有效的算法,快速识别运动目标与机器人的相对位置关系,并指示机器人做出主动避让,经验证,该方法所测得的位置精度可控制在0.5 m内。

Currently, one of the difficulties for dynamic object recognition technology to meet real - time requirements is processing mass data. Considering the problem of robotic active collision avoidance, the paper chooses mature algorithm of wavelet transform to compress and cut images. It puts forward a simple and effective algorithm to quickly identify the rela- tive position relationship between moving object and robot, which can instruct robot to avoid collision actively. It proves that the position precision calculated with this method can be controlled within O. $ m.