首页> 外文会议>International technical meeting of the Satellite Division of the Institute of Navigation >Stereo-Vision Aided GNSS for Automotive Navigation in Challenging Environments
【24h】

Stereo-Vision Aided GNSS for Automotive Navigation in Challenging Environments

机译:在具有挑战性的环境中,用于汽车导航的立体视觉辅助GNSS

获取原文

摘要

Stereo vision has recently attracted increased interest forthe purpose of navigation, especially in areas where othermeans of navigation are either not available or notfeasible.This paper discusses models used for systematic errorcorrection, methods for feature tracking and ego-motionestimation and a model for integrating the vision systemego-motion estimation with GNSS measurements. A lensdistortion correction method using an existing algorithmand manually selected features of opportunity only isused, and the details of an algorithm for stereo cameracalibration using a least-squares estimator and manuallyselected features of opportunity are presented. The detailsof a least-squares based ego-motion and accuracyestimator are provided and supported by a covarianceanalysis of different feature point distributions. Althoughthe results don’t include the output of the systemcombination Kalman filter, its model is presented.The results include dead-reckoning performance of twodifferent vision system setups with low-cost webcams andhigh-quality GPS-synchronized cameras, compared to theoutput of two GNSS receivers as well as a referencetrajectory. Ego-motion estimations extracted from morethan 1400 images with a total travelled distance of morethan 500 m were computed and are shown in the results.
机译:立体视觉最近吸引了越来越多的关注 导航的目的,尤其是在其他地方 导航方式不可用或不可用 可行的。 本文讨论了用于系统误差的模型 校正,特征跟踪和自我运动的方法 估计和集成视觉系统的模型 GNSS测量的自我运动估计。镜头 使用现有算法的失真校正方法 而仅手动选择机会的功能是 以及立体摄像机算法的细节 使用最小二乘估计器进行校准并手动进行 介绍了机会的选定特征。细节 最小二乘法的自我运动和准确性 估计量由协方差提供和支持 分析不同的特征点分布。虽然 结果不包括系统的输出 结合卡尔曼滤波器,给出了其模型。 结果包括两个的推倒重击性能 低成本的网络摄像头可实现不同的视觉系统设置 高品质的GPS同步相机,相比之下 两个GNSS接收器的输出以及一个参考 弹道。从更多中提取出的自我运动估计 超过1400张图像,总行进距离超过 超过500 m进行了计算,并显示在结果中。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号