Jump to content
You must now use your email address to sign in [click for more info] ×

RogerWehage

Members
  • Posts

    9
  • Joined

  • Last visited

Posts posted by RogerWehage

  1. On 3/2/2017 at 4:12 PM, MEB said:

    Hi SmilingJ,

    Yes, only a few moments ago (after you started uploading) i realised they were more than i thought - although you said they were 32 (!) on your first post - i forgot about it when i post the request.

    Again, thank you so much for the effort. I will check them out tomorrow/this weekend.

    Is this a dead end or how do I find out what happened "tomorrow/this weekend?"

  2. I have a telescope and can photograph any number of nearly identical images in rapid succession. In each successive image the object I'm interested in may occupy only a few percent down to much less than one percent of the total image, in essence, a sub-image. Because of earth rotation and possible object movement through space, the desired object and sub-images may slowly drift and rotate across successive images. I want to use some or all of these sub-images in Affinity Photo to obtain an enhanced image of the object. The source and sub-images may also be captured and processed in RAW or DNG format, or similar high resolution formats that Affinity Photo can handle.

    The following is my idea of how this might be automated, possibly by programming and/or recording macros. I'll use a moon image as an example, but hopefully the process will apply to any object that does not distort as it translates and rotates, almost imperceptibly in successive images. The moon is about 382,000 km away, and its movement is nearly imperceptible to the naked eye. However, it can be seen to drift ever so slightly across successive digital images, but its size and shape remain constant. Because of such large distances, virtually everything observed in space does not distort but may slowly translate and rotate in successive images. Affinity Photo can easily handle these linear type image situations.

    It is well known that noisy, repetitive, linear signals can be greatly enhanced by linear averaging, if the noise is random. Small, clipped images from space, taken in rapid succession and aligned, act like noisy, repetitive, linear signals, and the resolution can also be greatly enhanced by averaging. But not all of the noise is random, pixelation being the main non-random noise component. I'll suggest how one might apply image interpolation at the right time to filter out pixelation noise without significantly degrading final image resolution.

    I would start by teaching Affinity Photo what object I want it to find by manually clipping out a sub-image from one image, intentionally leaving a small amount of space around the object on all four sides. Using this sub-image as a template, Affinity Photo would then locate and clip out that same sub-image in all other images. In the process it would align all sub-images, similar to what is done in the "New Focus Merge..." macro.

    At this point I could have Affinity Photo average the aligned sub-images into one final image, but that final image would look very pixelated when exported and expanded. I want to reverse the process. Let me instruct Affinity Photo to interpolate (by any one of its five methods) all clipped and aligned sub-images by #X where # is any reasonable number greater than one, such as 8 or 6.3. The larger the number the better, except that memory usage and processing time must be considered. Now, apply the averaging process to the magnified sub-images to obtain the final enhanced image.

    If Affinity Photo were instructed to generate many sub-images and expand them by interpolation, and there was no reason to save them for later use, then there would be no need to store all of them in memory during the averaging process. Exact averaging can be done by what I'll call the alpha or A method. Suppose I have five numbers, 3, 15, 27, 9, and 21 that average to 15, and suppose that the first four numbers have been averaged to 13.5. Now I want to use the A method to average 21 in with this previous average of 13.5 to get the final exact average 15. Let N=5, the total number of items averaged in this step. Then let A = 1/N and the new average is obtained from (1-A)*13.5 + A*21 = (4/5)*13.5 + (1/5)*21 = 15. This exact averaging method has two advantages: only two interpolated images have to be kept in memory at any time (the desired final image and the latest interpolated sub-image to be averaged in) and potential overflow associated with adding together many positive numbers and dividing is avoided. Furthermore, much higher precision RAW images could be used and the averaging accumulation process could be performed without losing anything until the final export process.

    Based on this progressive alpha averaging method, the process might even be made much better. Suppose I have captured 1000 images and have made them available to Affinity Photo in a single sub-directory. Do I need to tell Affinity Photo to process all 1000 of them or could I just let it run, generating and averaging in interpolated sub-image after interpolated sub-image until virtually no further final image enhancement is detected? Or maybe until I instruct it to stop because the zoomed-in, enhanced final image looks good enough to me or I can't detect any further improvements? Also, because Affinity Photo is now processing one interpolated sub-image at a time and accumulating a higher resolution average with each new interpolated sub-image, is it possible that this always up-to-date, high-resolution, running average image would make a better template for locating, transforming, and clipping the next successive sub-image? I think always using the most recent average as template may minimize drift biases.


    I would like to learn how to generate and record macros in Affinity Photo to carry out these processes or see it implemented as another Affinity Photo feature like "New HiRes Avg..." or something similar. Is this feasible and what are the next steps? What I described above doesn't look entirely like a "recording macros" process but getting down to the programming level. I have many years of programming experience in Fortran and C and less in Python, but I don't know what's used in Affinity Photo. I'm definitely not qualified to program Affinity Photo but I could work with others to help get this job done.

  3. 21 minutes ago, v_kyr said:

    Well there are a bunch of apps which can do this in the one or other way....

    Yeah, I probably have a couple dozen of them like Inkscape, but was looking for something right in Affinity Photo. I take images of small objects using a macro lens and may clip them to smaller size. Then I will figure out approximately how many mm the resulting clipped image will be on each side. I can paste in a saved tic mark image, rotate, place, and stretch it to get the correct number of tic marks along the edge, and clip off the excess.

  4. I want to draw a uniformly-spaced set of tic marks along two edges of an image that may have been clipped and interpolated to a different pixel density. I would like to specify the tic mark spacing in terms of number of pixels per tic mark or something similar. Currently I have Affinity Photo and suspect that Affinity Designer would more likely be able to do something like this. Can someone please explain how this might be done and which app can do it?

    I sort of answered my own question. Using Affinity Designer I drew a short horizontal line, duplicated the line at uniform intervals, selected the lot, and grouped them. I can duplicate, rotate, and stretch the group to set the line spacing to whatever I want. I can also ungroup, add or remove lines, and regroup. Maybe that's not the easiest way but it will do for now.

  5. I love hiking through the great southwest and taking many thousands of snaps, mostly of landscapes because my damn DSLR lens won't focus on small stuff. I recently threw my worn-out camera away and replaced it with a "toy" iPhone XR camera and three Moment lenses. Now I can take home some high resolution photos of small (even tiny) things that don't move while I'm shooting at them. That's better than nothing, and I don't have to carry around several thousand dollars worth of heavy photography equipment. If there is any interest I will describe in detail how I use CameraPixels on my iPhone XR to focus bracket up to 50 images and Affinity Photo to focus align and merge the images. There is a wonderful world of macro photography right at our fingertips, and it doesn't have to cost thousands of dollars.

    This first image below, about one cm across and one cm deep, was taken with a Moment macro lens attached to an iPhone XR mounted on a tripod. CameraPixels was used to focus bracket 40 images, and Affinity Photo was used to focus align and merge the images as well as remove a halo around the petals.

    The third image shows the sexual part of a bright red geranium flower taken with the iPhone XR and Moment macro lens. Moment claims a 10X magnification from this lens, and I've verified that with a ruler. Imagine looking at a three to five inch cluster of bright red geranium flowers and staring into a single flower that is about 3-5 cm wide. What you see in this third image is the reproductive part of one of those flowers; the image is about 2 cm wide. Look closely at the detail in this image. See the tiny piece of fiber stuck to the pollen on one of the stamens? The middle image was clipped out of the right image in Affinity Photo and expanded using Bicubic Interpolation. We can clearly see the individual pollen particles. This is over a 50X magnification!

    TinyYellowFlower2.jpg

    geranium6cropped4.jpg

    geranium6bc.jpg

  6. If this is all the better one can do it's a wonder anyone would purchase CameraPixels and Affinity Photo for the purpose of obtaining deep depth of field images. Focus bracketing in CameraPixels is an absolute wonder, and focus aligning and merging in Affinity Photo is straight forward, although it's not well enough publicized. The first image is a wild violet about one cm across and taken on an iPhone XR back camera and Moment macro lens. The second image, although not apparently detailed at first glance, is an amazing 50X+ image of tiny pollen particles taken with an iPhone XR back camera and Moment macro lens. #shotonmoment.

    TinyBlueViolet.jpg

    geranium6cropped4.jpg

×
×
  • Create New...

Important Information

Terms of Use | Privacy Policy | Guidelines | We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.