智能科学与技术学报 ›› 2021, Vol. 3 ›› Issue (3): 280-293.doi: 10.11959/j.issn.2096-6652.202129

• 专刊:目标智能检测与识别 • 上一篇    

基于视觉显著性的车载单目相机自运动估计及前车尺度估计方法

艾明欣1, 刘铁1, 王净1, 丁佳丽1, 袁泽剑2, 尚媛园1   

  1. 1 首都师范大学信息工程学院,北京 100048
    2 西安交通大学人工智能学院,陕西 西安 710049
  • 修回日期:2021-05-18 出版日期:2021-09-01 发布日期:2021-09-01
  • 作者简介:艾明欣(1999− ),女,首都师范大学信息工程学院硕士生,主要研究方向为计算机视觉、模式识别等
    刘铁(1981− )男,博士,首都师范大学信息工程学院副教授,主要研究方向为计算机视觉、模式识别、多媒体计算、图像处理、大规模数据计算等
    王净(1997− ),女,首都师范大学信息工程学院硕士生,主要研究方向为计算机视觉、模式识别等
    丁佳丽(1996− ),女,首都师范大学信息工程学院硕士生,主要研究方向为计算机视觉、显著性物体检测等
    袁泽剑(1971− ),男,博士,西安交通大学人工智能学院教授,主要研究方向为图像处理、模式识别、计算机视觉、机器学习、概率机器人等
    尚媛园(1977− ),女,博士,首都师范大学信息工程学院教授,主要研究方向为图像处理、计算机视觉、嵌入式系统等
  • 基金资助:
    国家自然科学基金资助项目(61876112);国家自然科学基金资助项目(61976170);北京市自然科学基金资助项目(L201022)

A method based on visual saliency for vehicle-mounted monocular camera ego-motion estimation and vehicle scale estimation

Mingxin AI1, Tie LIU1, Jing WANG1, Jiali DING1, Zejian YUAN2, Yuanyuan SHANG1   

  1. 1 Information Engineering College, Capital Normal University, Beijing 100048, China
    2 College of Artificial Intelligence, Xi’an Jiaotong University, Xi’an 710049, China
  • Revised:2021-05-18 Online:2021-09-01 Published:2021-09-01
  • Supported by:
    The National Natural Science Foundation of China(61876112);The National Natural Science Foundation of China(61976170);The Natural Science Foundation of Beijing(L201022)

摘要:

提出一种基于视觉显著性的车载单目相机自运动估计及前车尺度估计方法。首先,针对车载相机自运动估计,通过视觉显著性计算方法检测并去除含有噪声的单目图像序列中的运动目标,同时考虑图像的纹理区域和平滑区域,利用加权显著图保留有用特征点,进而对车载相机进行鲁棒的自运动估计。其次,将前车距离转化为前车尺度估计问题,通过描述子匹配与李代数中正则化的强度匹配相结合的方法最小化损失函数,通过设计视觉注意力机制选择有纹理无遮挡的图像块,并对选定的图像块中的像素赋权以减轻被噪声破坏像素的影响,从而实现鲁棒、准确的尺度估计。最后,利用多个具有挑战性的数据集对所提方法进行分析验证。结果表明,单目相机自运动估计方法达到了基于立体相机方法的水平,前车尺度估计方法在充分发挥强鲁棒性优势的同时保证了预测精度。

关键词: 视觉显著性, 单目相机, 运动估计, 尺度估计

Abstract:

A method based on visual saliency for ego-motion estimation and scale estimation of the vehicle in front was proposed.Firstly, for the ego-motion estimation of vehicle-mounted camera, the visual saliency calculation method was used to detect and remove moving objects in the monocular image sequence containing noise.While considering the image texture and smooth region, the weighted saliency map was used to retain useful feature points, to improve the robustness of ego-motion estimation.Secondly, the distance of the vehicle in front was converted into a vehicle scale estimation, by integrating descriptor match and the strength of regularization match of the lie algebra to minimize loss function.The visual attention mechanism was used to get texture image block without shade, and the pixel in the image block weight to mitigate the effects of destroyed by noise pixel, so as to realize the robust and accurate scale estimation.Finally, several challenging datasets were used to analyze and verify the proposed method.The results show that the monocular camera ego-motion estimation method reaches the level of the stereo camera-based method, and the vehicle scale estimation method ensures the prediction accuracy while giving full play to the advantages of strong robustness.

Key words: visual saliency, monocular camera, motion estimation, scale estimation

中图分类号: 

No Suggested Reading articles found!