基于微处理器的智能车图像采集与识别算法研究

摘要

Tracking anti-collision system for intelligent vehicle based on machine vision is not easy to be realized in embedded system because of its high requirement for information, computation speed and storage capacity. With the method of extracting machine vision image information equidistantly, images acquired only retain its minimum feature, which compressed the acquisition and lighten the stress of storage. A simplified threshold-based edge extraction operator is proposed based on the research of the traditional image of edge detection operator, which is helpful to increase the speed of computation. This method is approved to be efficient by simulation and be verified on the platform of smart model car.%基于机器视觉的智能车辆跟踪防撞系统,信息量大,计算速度、存储容量要求较高,不便在嵌入式系统中实现.通过对机器视觉图像信息的等间距抽取,压缩了图像采集量,保留了最小特征,减小了存储的压力；通过对传统图像边缘提取算子的研究,提出了简化的基于阈值的边缘提取算子,提高了计算速度.仿真研究表明:该方法简单有效,并在智能模型车平台上得到验证.