Convolution filters are rather coarse devices used to sharpen or soften an image. You can use them on lunar and planetary images to help remove the softness caused by poor focus and bad seeing. You can also use them to build your own unsharp masks, although AIP4Win provides several ways to do this automatically. Essentially, these filters are useful building blocks of more complex operations.
For this tutorial, be sure that Auto Lo/HiStretch is selected as the default display mode in the Image Display Control window.
Step 1: Load an Image. From the Image Enhancement Tutorial subdirectory, load the image "moon.pa." This is an image taken with a Cookbook 245 CCD camera through a telescope with a tiny 6-millimeter aperture, so the image appears rather soft. We will use this image later with some more powerful techniques.
Step 2: Apply the Noise Filter. The Noise Filter is not, in a strict sense, a convolution filter, but rather a local median filter. It is useful in this case because a dark frame was not acquired with the image, and there are a few hot pixels that affect the processing. If you look around X = 80,Y = 180, you will see one of them. The Noise Filter can be useful in cleaning up images like these; but take care, it does remove fine detail.
Click the Enhance I Noise Filter... menu item and the Noise Filter window will appear. Leave the Deviation control set at 1 and click Apply. A new image will appear in which the hot pixels are gone.
Step 3: Apply the Crispen Filter. Click the titlebar of the noise filtered image to make sure it is active. Then click the Enhance I Convolution Filters I Crispen... menu item. A new image will appear in which the detail has been greatly enhanced. See how the contrast has increased and the detail in the craters has become more prominent.
Now try the crispen filter on the original "moon.pa" image by clicking its titlebar and invoking the crispen filter again. You will see a new image appear in which the lunar details are more prominent, as before; but also notice that the hot pixels in the image are enhanced along with the others. The filter enhances everything; it has no way of knowing which pixels are "pure" and which have been corrupted by noise. High-pass filters enhance noise.
Step 4: Apply the Sharpen Filter. Try the same steps you used in Step 2— but using the Sharpen Filter. Notice how the resulting image is much harsher. This is a fairly strong filter.
Step 5: Apply the Low-Pass Smooth and Blur Filters. Try the same steps using the Smooth and Blur filters and notice how the fine details in the image are suppressed.
When you are finished experimenting, close all the open images.
Was this article helpful?