Apparatus and method for recovering images damaged by weather phenomena

1. An image recovery device, comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform:

receiving an input image;

detecting an image damaged by a weather phenomenon from the input image;

recovering the damaged image; and

outputting the recovered image,

wherein the recovering comprises recovering at least one of a contrast and color of the damaged image, and

wherein the recovering further comprises blending an original image and the recovered image by adjusting a ratio on a distance from a boundary of a preset region of interest (ROI).

2. The image recovery device of claim 1 , wherein the detecting comprises verifying at least one of a brightness value distribution and an edge component of the input image, and detecting the damaged image from the input image.

3. The image recovery device of claim 1 , wherein the detecting comprises verifying components of the input image, and detecting the damaged image from the input image when at least one condition for a threshold of each of the verified components is satisfied.

4. The image recovery device of claim 3 , wherein the detecting comprises detecting a specific image corresponding to the ROI in the input image as the damaged image when a number of edge components that have values equal to or greater than a constant threshold among edge components of the specific image, is equal to or less than a threshold for a number of the edge components.

5. The image recovery device of claim 3 , wherein the detecting comprises detecting a specific image corresponding to the ROI in the input image as the damaged image when a brightness value distribution of the specific image is equal to or less than a selected threshold.

6. The image recovery device of claim 3 , wherein the detecting comprises detecting a specific image corresponding to the ROI in the input image as the damaged image when an average brightness value of the specific image is equal to or greater than a selected threshold.

7. The image recovery device of claim 3 , wherein the detecting comprises detecting a specific image corresponding to the ROI in the input image as the damaged image when a standard deviation value of the specific image is included in a selected threshold range.

8. The image recovery device of claim 3 , wherein the detecting comprises detecting a specific image corresponding to the ROI in the input image as the damaged image when an airlight value of the specific image is equal to or greater than a selected threshold.

9. The image recovery device of claim 2 , wherein the detecting comprises detecting a specific image corresponding to the ROI in the input image as the damaged image when a color-difference signal vector magnitude of the specific image is equal to or less than a threshold.

10. The image recovery device of claim 2 , wherein the recovering further comprises expanding a brightness value distribution of the damaged image at a ratio equal to or greater than a constant ratio, and recovering the damaged image.

11. The image recovery device of claim 2 , wherein the recovering further comprises resetting a brightness value distribution of the damaged image through at least one of stretching and histogram equalization, and recovering the damaged image.

12. The image recovery device of claim 2 , wherein the recovering further comprises removing an effect of a weather environment based on modeling of the weather environment, and recovering the damaged image.

13. The image recovery device of claim 2 , wherein the recovering further comprises amplifying each of color components of the damaged image at a ratio equal to or greater than a constant ratio, and recovering the damaged image.

14. The image recovery device of claim 1 , wherein the detecting comprises verifying the preset ROI and determining whether a specific image corresponding to the preset ROI is damaged, and detecting the specific image as the damaged image when the specific image is determined to be damaged.

15. The image recovery device of claim 1 , wherein the recovering further comprises recovering the damaged image corresponding to the preset ROI when the preset ROI is detected to be damaged.

16. The image recovery device of claim 1 , wherein the recovering further comprises simultaneously recovering a contrast and color of the damaged image through at least one of stretching and histogram equalization.

17. The image recovery device of claim 1 , wherein the recovering further comprises sequentially recovering a contrast and color of the damaged image.

18. The image recovery device of claim 1 , wherein the recovering further comprises verifying a damage intensity of the damaged image, determining a recovery level of the damaged image based on the verified damage intensity, and recovering the damaged image based on the determined recovery level.

19. An image recovery method, comprising:

receiving an input image;

verifying at least one of a brightness value distribution and an edge component of the input image, and detecting an image damaged by a weather phenomenon from the input image;

recovering the damaged image; and

outputting the recovered image,

wherein the recovering comprises recovering at least one of a contrast and color of the damaged image, and

wherein the recovering further comprises blending an original image and the recovered image by adjusting a ratio based on a distance from a boundary of a preset region of interest (ROI).

20. A non-transitory computer readable recording medium storing a program to cause a computer to implement the method of claim 19 .