Scholarlink

<p>The safety of hydraulic tunnel lining is an important issue affecting the operation life of hydraulic tunnel, and the effective identification of tunnel lining crack image is an important link to ensure the accurate judgment of tunnel safety state. Due to the high turbidity and low illumination conditions of underwater hydraulic tunnels, the existing algorithms are difficult to avoid excessive image enhancement while retaining the true color and image details of the original cracks. Therefore, according to the characteristics that the brightness channel adjustment is independent of the image color in the image Lab mode, this paper optimizes the brightness component in the Lab channel, and proposes an image enhancement algorithm based on the improved Lab color space Retinex algorithm. The Retinex algorithm based on guided filtering is used to estimate the brightness value of the incident component of the image Lab channel, and the reflection component is obtained. Then, after gamma correction, the adaptive logarithmic mapping method is used to improve the image brightness contrast. Finally, the chromaticity and brightness components of the Lab channel are recombined to obtain the image after the crack details are effectively enhanced. Based on the crack image data obtained from the model test and the tunnel site, the algorithm proposed in this paper is compared and verified. The results show that the algorithm can better deal with the relationship between over-exposure and low-exposure regions in the original image, making the overall brightness more uniform. In the comparison of tunnel field image processing, the proposed enhancement algorithm improves the quality evaluation indexes of Brenner, Tenengrad, SMD2, EOG and AG by 4.90%, 4.73%, 6.35%, 20.9% and 1.81% respectively compared with the NBUet (Noise Basis Learning for Image Denoising with Subspace Projection) algorithm benchmark in this paper, and improves the image processing speed by 16.8%. The crack image enhancement algorithm proposed in this paper can provide some help for the effective identification of lining crack diseases of diversion tunnels and the safe operation of tunnels.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhanced Retinex for subaquatic tunnel cracks via lab space

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • Yunlong Tie,
  • Bo Han,
  • Chunlei Zhang,
  • Qiyue Zhang,
  • Shilei Zhang,
  • Jin Bai,
  • Haipeng Bi,
  • Meiling Li

摘要

The safety of hydraulic tunnel lining is an important issue affecting the operation life of hydraulic tunnel, and the effective identification of tunnel lining crack image is an important link to ensure the accurate judgment of tunnel safety state. Due to the high turbidity and low illumination conditions of underwater hydraulic tunnels, the existing algorithms are difficult to avoid excessive image enhancement while retaining the true color and image details of the original cracks. Therefore, according to the characteristics that the brightness channel adjustment is independent of the image color in the image Lab mode, this paper optimizes the brightness component in the Lab channel, and proposes an image enhancement algorithm based on the improved Lab color space Retinex algorithm. The Retinex algorithm based on guided filtering is used to estimate the brightness value of the incident component of the image Lab channel, and the reflection component is obtained. Then, after gamma correction, the adaptive logarithmic mapping method is used to improve the image brightness contrast. Finally, the chromaticity and brightness components of the Lab channel are recombined to obtain the image after the crack details are effectively enhanced. Based on the crack image data obtained from the model test and the tunnel site, the algorithm proposed in this paper is compared and verified. The results show that the algorithm can better deal with the relationship between over-exposure and low-exposure regions in the original image, making the overall brightness more uniform. In the comparison of tunnel field image processing, the proposed enhancement algorithm improves the quality evaluation indexes of Brenner, Tenengrad, SMD2, EOG and AG by 4.90%, 4.73%, 6.35%, 20.9% and 1.81% respectively compared with the NBUet (Noise Basis Learning for Image Denoising with Subspace Projection) algorithm benchmark in this paper, and improves the image processing speed by 16.8%. The crack image enhancement algorithm proposed in this paper can provide some help for the effective identification of lining crack diseases of diversion tunnels and the safe operation of tunnels.