无人机视角下的红外图像去模糊算法研究
DOI:
作者:
作者单位:

1.国家管网集团东部原油储运有限公司;2.西南石油大学理学院

作者简介:

通讯作者:

中图分类号:

TP751

基金项目:

国家管网揭榜挂帅项目 WZXGL202106 无人机视角下的潜在管道威胁事件识别技术


Research on Infrared Image Deblurring Algorithm from Drone's Perspective
Author:
Affiliation:

PipeChina Network Corporation Eastern Oil Storage and Transportation Co. Ltd

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    针对油气长输管道采用无人机巡检时所拍摄的红外图像去模糊问题,本文利用图像通道的先验知识提升模糊图像质量,分别基于双边滤波和非盲去模糊网络NBDN去除人工伪影的方法达到更佳的图像复原效果。首先,基于暗通道先验知识,在最大后验的优化框架中添加暗通道的L_0正则项;然后使用图像梯度的L_0正则项,代替图像像素的L_0正则项作为潜在图像的正则化约束,使用迭代交替估计图像模糊核和中间潜在图像;采用半二次分裂方法和查表法间接优化求解,估计中间潜在图像;采用双线性插值估计图像模糊核,通过对图像进行上下采样,构建图像金字塔,进而利用共轭梯度法直接优化求解。最后,利用估计的模糊核,使用基于超拉普拉斯先验的图像非盲去模糊方法得到潜在图像I_1;使用基于L_0正则化的非盲去模糊方法得到潜在图像I_0;计算估计的潜在图像I_1和I_0之间的差值映射,从I_1中减去双边滤波过滤后的差分图,得到最终的潜在图像I。将本文算法在低照度图像、含有饱和像素的图像、真实图像以及红外摄像图等图像数据上进行实验,相对于其他图像去模糊算法,实验结果表明本文提出的方法在多种模糊图像复原效果上,均具有较强的竞争力。

    Abstract:

    In this paper, we propose a method to enhance the quality of blurred infrared images captured during unmanned aerial vehicle (UAV) inspections of oil and gas pipelines. We address the issue of image deblurring by utilizing prior knowledge of image channels and employing bilateral filtering and the Non-Blind Deconvolution Network (NBDN) to remove artificial artifacts. Firstly, we incorporate the dark channel prior knowledge into a maximum a posteriori optimization framework by adding a dark channel regularization term. Then, instead of using regularization on image pixels, we utilize an regularization term based on image gradients as a regularization constraint for the latent image. We iteratively estimate the blur kernel and the intermediate latent image using alternating estimation techniques and indirect optimization methods such as semi-quadratic splitting and table lookup. The blur kernel is estimated using bilinear interpolation, and an image pyramid is constructed by upsampling and downsampling the image, which is then directly optimized using the conjugate gradient method. Finally, with the estimated blur kernel, we employ a non-blind deblurring method based on the super-Laplacian prior to obtain the latent image, and another non-blind deblurring method based on regularization to obtain the latent image . We calculate the difference map between the estimated latent images and and subtract the filtered difference map from using bilateral filtering to obtain the final latent image I.We conducted experiments on low-light images, images with saturated pixels, real images, and infrared camera images to evaluate the proposed algorithm. The experimental results demonstrated the competitive performance of our method compared to other image deblurring algorithms in terms of restoring various types of blurred images.

    参考文献
    相似文献
    引证文献
引用本文

曹旦夫,齐峰,谭冰,等. 无人机视角下的红外图像去模糊算法研究[J]. 科学技术与工程, , ():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-07-10
  • 最后修改日期:2023-11-11
  • 录用日期:2023-11-14
  • 在线发布日期:
  • 出版日期:
×
亟待确认版面费归属稿件,敬请作者关注