Identifying grid in image and removing it

I have images with some different objects on it and I have a specific backgound on it, it is a grid:

It can开发者_运维知识库 be colored to a different color, but still with that grid. I want to find shapes of objects on that image, if I do it, that grid will give me many noise as a result. Could you please suggest me how to get rid of that noise or grid from image, if every image has such a grid?

---

1a) If the grid is of similar color as your object behind the 'grid', you can use Hough Transform (to detect edges and remove edges at 45 degrees and -45 degrees).

1b) If the grid is of different color, you can just remove anything with that color range.

2) You will then need to apply a texture synthesis algorithm (which i assume, after removing the 45/-45 degree edges, you will see a tremendous amount of gaps) to fill the gaps.

Quick/Brief Reference:

[1] http://en.wikipedia.org/wiki/Hough_transform

[2] http://en.wikipedia.org/wiki/Texture_synthesis