Underwater Im Image Dehazing wit ith a Li Light Fie ield Camera - - PowerPoint PPT Presentation

โ–ถ
underwater im image dehazing wit ith a li light fie ield
SMART_READER_LITE
LIVE PREVIEW

Underwater Im Image Dehazing wit ith a Li Light Fie ield Camera - - PowerPoint PPT Presentation

May 19, 2023 108 likes โ€ข332 views

Underwater Im Image Dehazing wit ith a Li Light Fie ield Camera Katherine A. Skinner 1 and Matthew Johnson-Roberson 2 1 Robotics Program, University of Michigan, USA 2 Department of Naval Architecture and Marine Engineering, University of

slide-1
SLIDE 1

Underwater Im Image Dehazing wit ith a Li Light Fie ield Camera

Katherine A. Skinner1 and Matthew Johnson-Roberson2

1Robotics Program, University of Michigan, USA 2Department of Naval Architecture and Marine Engineering, University of Michigan, USA

slide-2
SLIDE 2

Motivation

2

Photo: Volodymyr Goinyk

slide-3
SLIDE 3

Motivation

3

Lizard Island, Australia Port Royal, Jamaica

slide-4
SLIDE 4

Motivation

4

Photo: Jaffe Lab

slide-5
SLIDE 5

Motivation

5

Photo: Jordt (2014)

slide-6
SLIDE 6

Light Field Cameras

6

  • Real-time RGB-D
  • Passive optical sensors
  • Compact form factor
  • Increased depth-of-field

Photo: Raytrix Photo: Lytro

slide-7
SLIDE 7

Real-time Underwater 3D Reconstruction

7

Katherine A. Skinner and Matthew Johnson-Roberson, "Towards real- time underwater 3D reconstruction with plenoptic cameras.โ€œ IROS, 2016.

slide-8
SLIDE 8

Real-time Underwater 3D Reconstruction

8

Katherine A. Skinner and Matthew Johnson-Roberson, "Towards real- time underwater 3D reconstruction with plenoptic cameras.โ€œ IROS, 2016.

slide-9
SLIDE 9

Pure vs. Turbid Water

9

Disparity map in turbid water Disparity map in pure water

slide-10
SLIDE 10

Underwater Image Restoration

10

Contribution: We present a pipeline for dehazing of underwater light field images incorporating a physical model of underwater light propagation.

slide-11
SLIDE 11

Underwater Light Field Image Dataset

11

Contribution: Provide a comprehensive light field dataset for underwater image dehazing. https://github.com/kskin/data

slide-12
SLIDE 12

Technical Approach

12

slide-13
SLIDE 13

Dehazing Model

13

๐ฝ(๐’š) = ๐พ(๐’š)๐‘“โˆ’๐œƒ๐‘’(๐’š) + ๐ต(1 โˆ’ ๐‘“โˆ’๐œƒ๐‘’(๐’š))

slide-14
SLIDE 14

Dehazing Model

14

๐ฝ(๐’š) = ๐พ ๐’š ๐‘ข(๐’š) + ๐ต(1 โˆ’ ๐‘ข ๐’š )

slide-15
SLIDE 15

Dehazing Model

15

Reference: Y. Li et al. โ€œNighttime haze removal with glow and multiple light colorsโ€. ICCV 2015. ๐ฝ(๐’š) = ๐พ ๐’š ๐‘ข(๐’š) + ๐ต(๐’š) 1 โˆ’ ๐‘ข ๐’š +๐‘€๐›ฝ (๐’š) โˆ— ๐‘ฉ๐‘ธ๐‘ป๐‘ฎ

  • Non-uniform illumination
  • Glow patterns
slide-16
SLIDE 16

Synthesizing Views

16

slide-17
SLIDE 17

Recovering Epipolar Images

17

slide-18
SLIDE 18

Results

18

In air Raw underwater Our result

slide-19
SLIDE 19

Results

19

Disparity map using raw underwater image Disparity map using corrected image

slide-20
SLIDE 20

Conclusions

  • Presented an underwater light field image dehazing pipeline that

incorporates a physical model of light propagation

  • Provided an underwater light field image dataset

20

slide-21
SLIDE 21

21 [1] Dansereau, et al. Light field image denoising using a linear 4D frequency-hyperfan all-in-focus filter. Computational Imaging, 34(2):15โ€“1, 2013.

  • Gather light field images of more complex scenes
  • Greater depth disparity
  • Varying turbidity
  • Further leverage light field structure
  • All-in-focus filter [1]
  • Consistent motion between sub-aperture images
  • Incorporate more components of physical model for underwater

light propagation

  • Recovering range in underwater imaging

Future Work

slide-22
SLIDE 22

Final Thoughts

22

  • Underwater light field camera โ€“ contact kskin@umich.edu
  • Deep learning approach to underwater image restoration

Honolulu, Hawaii