针对区域立体匹配计算量大实时性差的困难,分析了相关匹配算法的实际工作过程,采用消除冗余因子和M2C滤

波、多级分辨率匹配减小计算复杂度,对算法结构进行了改进和优化,并利用超线程和D65/EF技术对算法进行了加速,提出

了一种实时区域匹配算法。对算法进行实验,结果表明算法符合了视觉导航的准确性和实时性要求,并且对于提高其他区域

匹配算法实时性也具有重要借鉴意义。-Stereo matching for the regional real-time calculation of a large quantity of poor difficulties, analysis of the relevant matching algorithm the actual work process, used to eliminate redundancy factor and M2C filtering, multi-resolution matching to reduce the computational complexity of the algorithm structure to improve and optimize the and take advantage of Hyper-Threading technology and D65/EF accelerated algorithm, a regional real-time matching algorithm. Algorithm experiment results show that the algorithm is consistent with the accuracy of visual navigation and real-time requirements, and improve in other regions for real-time matching algorithm is also an important reference.