Advertisement
Guest User

Untitled

a guest
Jun 26th, 2017
66
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 1.52 KB | None | 0 0
  1. Heat equation and other methods to blur an image.
  2.  
  3. The main goal of the project is to test methods which blur images.
  4.  
  5. Why do we need to blur pictures? When a spacecraft shots pictures, it has to send it to the ground (ground segment) and therefore, pictures need to be smallest as possible to optimize the use of the bandwidth. Indeed, when one blurs an image, it reduces the quantity of information delivered by the picture and so it is easier to compress it.
  6. It can also be needed to fit the image in a small size, for example when building an album for customers.
  7.  
  8. The first method used is the heat equation. What is the heat equation? It is a very fundamental equation which describes the behavior of the heat: diffusion and convection. This equation is very useful and thus used in many domains.
  9. In our case, we use it to do an isotropic blur. Isotropic means that the blur will be uniform and shall not depend of the picture itself: strong variations in schemes or colors for example.
  10.  
  11. However, it is not a very good mean to keep essential information. Indeed, the most important parts in images (in a blurring process at least) are the shapes: contours. Therefore, the mean curvature driven motion seems a better method because it is anisotropic. Unlike to the heat equation, it works on gradients. Gradients are the derivatives of the function composed by the pixels of the image. A strong gradient means a big variation in the shape or colors. Hence, this method can detect contours and keep them, i.e. blurs the whole image except edges.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement