• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>IEEE International Conference on Intelligent Transportation Systems >Fog detection for de-fogging of road driving images
【24h】

Fog detection for de-fogging of road driving images

机译：雾检测可消除道路行驶图像的雾

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Advanced driver assistance system (ADAS) can recognize traffic signals, vehicles, pedestrians, and so on all over the vehicle. However, because the ADAS is based on images taken in an outdoor environment, it is susceptible to ambient weather such as fog. Therefore, preprocessing such as de-fog and de-hazing techniques is required to prevent degradation of object recognition performance due to decreased visibility. But, if such a fog removal technique is applied in an environment where there is little or no fog, the visual quality may be deteriorated due to excessive contrast improvement. In this paper, we propose a fog detection algorithm to selectively apply de-fogging method only in the presence of fog. Experimental results show that in the actual images, the proposed algorithm shows an average of more than 97% fog detection accuracy, and improves subjective image quality of existing de-fogging algorithms.
机译：先进的驾驶员辅助系统（ADAS）可以识别整个车辆的交通信号，车辆，行人等。但是，由于ADAS基于在室外环境中拍摄的图像，因此易受雾等环境天气的影响。因此，需要进行诸如除雾和除雾技术的预处理，以防止由于可见度降低而导致的物体识别性能下降。但是，如果将这种除雾技术应用于几乎没有或没有雾的环境中，则由于对比度的过度提高而可能导致视觉质量下降。在本文中，我们提出了一种雾检测算法，以仅在有雾的情况下选择性地应用除雾方法。实验结果表明，该算法在实际图像中的平均雾检测精度达到97％以上，提高了现有除雾算法的主观图像质量。

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