摘要
针对常见车速测量方法存在测速准确性较低、自主性较差和灵活性不足等问题,建立了车载自主测速模型,将车辆的测速问题转化为序列图像的匹配问题;结合车辆运行状态的约束条件和序列图像的成像特点,依次从特征提取、相似性度量、搜索空间和搜索策略等4个方面对序列图像匹配方法进行分析,并设计序列图像匹配算法。仿真结果表明,采用基于二值图像的匹配方法和采用平均绝对误差作为相似性度量函数,保证了较低的运算复杂度,提高了运算效率;提出的图像匹配算法具有较高的匹配精度和匹配速度。最后通过载车试验验证了所提方法的整体有效性和可靠性。
In order to solve the problems of low precision, dependence and low mobility for common vehicle velocity measurement system, a vehicle-borne autonomous velocity measurement model is designed. Therefore, the problem of vehicle velocity measurement is transformed into sequence image matching. Considering the con strained parameters of vehicle movement and the characteristics of sequence images, sequence image matching method is analyzed in four aspects of feature extraction, similarity measurement, searching space and searching strategy. Image matching algorithm is also designed. The simulation result shows that the matching method base on binary images and similarity measurement of mean absolute error function ensure computation complexi- ty greatly reduced with a high computation efficiency. And the proposed matching algorithm exhibits a behavior with a high matching precision and speed. The result of vehicular test proves the effectiveness and reliability of the proposed method.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2015年第4期964-968,共5页
Systems Engineering and Electronics
基金
装备预研基金资助课题
关键词
车速测量
图像匹配
相似性度量
搜索策略
vehicle velocity measurement
image matching
similarity measurement
searching strategy