标签: performance opencv visual-odometry
我正在从运行Ubuntu 18.04的6核CPU上的OpenCV rgbd模块测试视觉里程表。我建立了启用了TBB,IPP,OpenMP和CUDA的OpenCV。在cv::rgbd::Odometry::compute,cv::rgbd::RgbdICPOdometry和cv::rgbd::ICPOdometry上调用cv::rgbd::RgbdOdometry的测试仅在单个内核上运行。是否可以通过使用多个内核以及GPU来加速它?
rgbd
cv::rgbd::Odometry::compute
cv::rgbd::RgbdICPOdometry
cv::rgbd::ICPOdometry
cv::rgbd::RgbdOdometry