Primer - planetary imaging with a Toucam

[MartinB 2,588]

Introduction to Planetary imaging with a Toucam

This is a rather lengthy article but it covers the whole process of planetary imaging from setting up the Toucam and scope to final processing using Registax. I hope to be able to spare you some of the pain and frustration I suffered getting to grips with the process. The tutorial contains much useful information and advice I gained from more experienced members of the forum.

What you will need: -

Hardware

Scope and mount (preferably motorised but not essential)

Toucam (or other ccd equipped cam)

1.25” adaptor

Laptop with USB connection (A desktop could suffice at a real pinch but very inconvenient). A higher end Pentium 2 will probably be adequate but hard disc space may be an issue with a larger avi being over 200MB. The processiing will be quicker on a more modern machine so you will need a way of transferring data over eg decent sized memory stick. Laptops seem quite happy out in the harsh dewy atmosphere of a starry night. Mine has had several frosty experiences and seems none the worse for it.

Preferably an infra red blocking filter but not necessary for lunar or monochrome shots.

Barlow lens – for planets you are aiming for an F number of at least 20. So if your scope is F10 you need a x2 barlow, if it is F5 you need a x4 etc. You can go up to F30 and beyond but you will need excellent seeing, you are stretching the optics further and not additing a great deal more.

Focal reducer – for longer focal length scopes you may wish to get a wider field view so that the whole of the moon is in the FOV. Atik make a relatively cheap x0.5 reducer

Software (all free)

K3ccdtools (version 1) http://www.pk3.org/Astro/index.htm?k3ccdtools.htm

Virtualdub http://www.virtualdub.org/

Registax (version 3) http://registax.astronomy.net/

Why a Toucam

The Phillips Toucam pro II is an upmarket web cam in that it actually has an 840K pixel CCD chip. Earlier versions with slightly lower resolution are still very good. By taking many frames per second over a period of 1-2 minutes the Toucam can easily capture over 1000 frames. Each single image will be corrupted by noise both from the camera itself and from the atmosphere. This noise is mostly random. By stacking the frames Registax is able to exclude random noise effects from the image but retain features which appear more regularly. This massively improves the quality of the final image.

This AVI video capacity makes the Toucam superior for bright solar system objects to much more expensive CCD cameras. The lower resolution is not a disadvantage for highly magnified planets since the limiting resolution is the image coming through the scope rather than the ability of the chip.

Capture Software

The Toucam comes with drivers, photo video and capture software. A suite called VR lounge. For video capture purposes this does work but is awkward and in my experience quite buggy. You are much better to start using K3ccdtools from the outset y(ou still need to load the Toucam drivers though). Version 1 is free and later versions offer little advantage for normal use. K3ccdtools allows you to name the capture file, subsequent files are then renamed sequentially and automatically. There is greater choice of frame rate and the whole interface is more convenient to use.

Setting up

The Toucam lens has to be unscrewed and replaced by a plastic adaptor which screws into the body of the cam – it is very easy to cross and strip the threads so be careful! The cam then drops into the scope just like an eyepiece.

Connect your cam to your laptop and then launch K3ccdtools. Don’t bother putting the cam in the scope at this point. You can set up the software prior to going out if you wish.

i) Select video capture from the top menu, check “show level meter”, “preview” and “Microsoft WDM Image Capture”, I don’t know if this last one varies depending on version of windows. This all enables the video window to be displayed. If enough light is getting into the cam you should be able to see it. Obviously you won’t be able to focus on anything now you have removed the lens.

ii) Still under video capture select “video format”. Choose 640 x 480 for resolution and 1420 for pixel depth. Apply.

iii) Now select video source. Under the “image controls” tab uncheck the “full auto” box. Select 10 frames per second. Move the brightness slider well over to the R. Gamma can be left low or off. Saturation half way. If filming the moon select black and white.

Under the “camera control” tab – leave the white balance on auto. Clear the Exposure auto box. Shutter speed 1/25th second and turn the gain up fairly high.

Apply the changes and return to the main screen.

Select “options” on the top menu bar then “settings”. Confirm that 10 fps is selected. Check enable capture limit and choose a time of 100 seconds. Make sure “capture audio” is left unchecked. You can select a directory for capture files. Select a file name and check incremental. This is extremely useful since it automatically renames each capture file e.g. select file name Saturn. Your first capture will be Saturn0000 next will be Saturn0001 etc. Preview rate 600ms.

Below the top menu are some small short cut items. Some of these are very useful. One is essential – the capture video sequence since this is the only way to start your capture.

You have now set up the software ready to identify your target on the chip. It will be highly overexposed which makes finding the object much easier when it is out of focus.

Getting the target on the chip

Lets assume you are filming a bright planet e.g. Saturn. The moon is less of a challenge because it is bigger and brighter. Find your planet and centre it in a higher and higher magnification eye piece. Once centred at high power it is time to insert the web cam (without barlow at this stage).

If you are using a polar aligned motorised mount the next bit is relatively easy, however without a tracking motor you are now in for a tricky time. The magnification for planets is high and the planet will move across the screen quite rapidly so you will have to be continuously working the RA slow motion control. A reasonable polar alignment is very helpful. When inserting the cam you should anticipate the amount the target will move across the screen and time it to be at the centre by the time you have put in the cam.

On inserting the cam you should see some sort of circular brightness indicating the presence of your planet. Bring it into focus remembering the direction of adjustment you have to make. Don’t be concerned that it looks very bright and fuzzy at this point. Once centred nicely insert the cam with the barlow (if working on a planet). Focus again.

Settings for Capture

It is now time to adjust the settings for capture of the image. Select video capture|video source or alternatively click on the little camera icon. Drag the Video Source window over to the right so that you can see your target whilst you make adjustments. Under image controls move the brightness down half way. For Saturn I leave a little bit of gamma on. Generally it is left low or off.

Select camera controls and start to turn the gain down. Clicking on the slider bar will adjust the setting in chunks, dragging the slider with the mouse will allow finer adjustment.

You now have to play with the brightness, gain and saturation. Some people use a faster shutter speed. It may be helpful for people with non motorised mounts who will have a lot of target movement, or if the seeing is poor to use 1/50 sec and more gain otherwise 1/25 is fine. Obviously the moon will require much faster shutter speeds. For imaging Saturn you should aim to lower the brightness and gain up to the point where ring detail is about to be lost. This point is quite important. Too dim and ring detail is lost, too bright and the disc can be burnt out. It will look quite dim but don’t worry, the stacking will brighten it up. Gain introduces random noise which Registax can remove therefore for Saturn I have found that taking the brightness down to about 40% and then adjusting the gain works quite well.

I have found it easiest to adjust the focus with the image slightly brighter than the final capture setting. Focussing is quite difficult but critical. Be patient and and make very fine adjustments. For Saturn you have a very nice focussing aid with Cassini. You will notice how the seeing effects the focus. In bad seeing the image will appear unfocussed most of the time and then every now and then it jumps into good focus. Just do the best you can!

Once the exposure and focus of the target seems about right I adjust the saturation. High saturation will produce a richer colour but if this is over cooked it can look unnatural. Too low a saturation and detail in the disc of Saturn is less evident. There is a lot that is down to personal preference. You should take a few captures using different settings to experiment.

One final adjustment to be made is the position of the target on the screen. The chip is likely to have one or two flecks of dust on it. These show up badly on the screen. Their presence is normal but you should try to locate the target away from them.

Note on frame rate

The Toucam uses a USB1 connection and has to compress the image when delivering it to the laptop. The more compression the more the image is degraded. The lower the frame rate of the capture the less compression is required. However this needs to be balanced against the benefit of a large number of frames. A higher frame rate will deliver more frames in a given time. Exposures can be shorter with less risk of the image being spoiled by rotation of the target (which becomes increasingly likely as capture times move beyond 90 seconds). In my view if seeing and tracking are good a lower frame rate of 5fps can be used since a high proportion of the frames are good quality. With bad seeing a higher rate of 10 or even 15 is preferable which will ultimately yield more useable frames. Again, experiment. So far I haven’t been impressed with improvements moving down to 5fps.

Capture

Click the capture video sequence icon and relax. Don’t worry if the image becomes jerky, this is normal. At the bottom of the screen will be an indication of the progress of the capture and will show the number of frames being dropped. It is quite normal for a significant number of frames to be lost. The information will be summarised at the end of the capture.

During capture with a non motorised mount stay as relaxed at possible and work hard to keep the image as stable as possible. This isn’t that easy since there is some delay in the image update during the capture process.

Even with a tracking mount there is likely to be a little drift. Make adjustments to ensure that the image doesn’t creep off screen or move over a dust blob.

That’s it!!! You have your image saved under a file name you chose earlier. Get on and take some more images experimenting with different combinations of gain and brightness, saturation level and frame rate. Just make one adjustment at time or you will end up confusing yourself.

Registax

Registax is a wonderful piece of generously donated software. It can be fully automated or you can tweak it endlessly. For this tutorial I am going to run through the automatic operation and bypass all the complexity. The most important manual operation you may wish to perform is frame selection. Registax permits this but I find it very clumsy. Virtual dub is much easier and quicker. I will cover this after since the first thing I want to do on getting in is to get a quick idea of what I have managed to get.

Launch Registax and click select. Open your chosen avi. You will see a white box on the screen which you can move around with the mouse. If you are using a colour image make sure colour is selected. Registax will prompt you about this anyway if it needs changing. Now move across and under processing check automatic. Select 1024 for the area (this slows down processing and if you have a slow computer and a steady image you could opt for 512 however if the image moves out to the edges of the screen your post processing will be limited). Leave the quality estimate as “classic”.

There is a slider bar below the image with the pointer over to the left. It will be pointing to the first frame and this is the image which is shown. You now have to create a reference image for alignment. This should be a good quality frame. Click on the slider bar and the pointer will move one frame at a time. Alternatively click on the frameset tab to the lower right of the screen. This will show all the frames selected. By highlighting the frame you move the pointer to that frame and it is shown on the screen. Find a high quality image that is as early in the frame set as possible. Once you have chosen that frame move the alignment box over the image. For a planet the box should comfortably contain it. For the moon select a feature such as a well defined crater. The smaller the box the quicker the alignment will be but the harder the software will find it to keep the image lined up so the more work you will have to do manually. If you have a lot of movement as a result of manual tracking choose a larger box. The size of the box is selected under alignment box. Try the different sizes to see which fits best. Don’t be tempted to click on a planetary feature in the hope that it will improve alignment accuracy, it won’t work well!

When happy left click the mouse with the box in place. Registax will perform something called a fast fourier transform. This means a lot to some people but not to me. You now need to adjust the quality. Under quality estimate you will see a “lowest quality” box. This is a percentage figure and indicates the percentage of images that will be rejected by registax when it does it’s quality analysis. Assuming you have around 1000 frames 80% is a reasonable figure. If you had good seeing and are a bit low on frames because you have been capturing at 5 fps you could lower this a little.

Now click the align button towards the top left – below the align tab (don’t confuse the align tab with the align button as I have done many times). Registax will now spring into action. You will see a quality graph, Registax moves through the sequence with the box attempting to keep pace with movement of the image. If the image moves too quickly the box won’t keep up and you will be asked if you wish to ignore the frame or add it manually. Always choose to add it manually since it will help Registax to catch up. You will need to click over the image again. Sometimes you have to repeat this each time it looses track. If the image goes off the screen you will need to remove these frames. This is best done using virtual dub.

Once it has run through the alignment it starts to process image against the reference frame and improve aligment. You can peer at the graphs or go and make a cup of coffee. It will repeat this process 2 or 3 times until it is happy. Once done it starts to stack. Don’t worry, the image you now have on your screen isn’t what you have achieved so far. I think it is the last frame in the sequence.

This is the exciting bit!! Once processing is complete a much brighter image appears on the screen which is your stacked image. You are automatically moved to the Wavelet section which is where the fun adjustments are done. Save your project file at this point. Don’t use the “save image” button but use the project file save at the top of the screen.

Don’t be alarmed if you image looks a bit blurred. This is quite usual.

RGB shift

You may notice a blue fringe at one side of your image and a red one at the other. On the right hand side of the screen you will see a tab RGB shift. Select this. An alignment box should be formed over your image. Click “estimate”. After a couple of processing runs the fringe should magically disappear.

Wavelet sliders

These sliders are over on the left hand side. There are 6 of them and they are Laplacian unsharp masks. I read that somewhere. They are very clever and look for points of change in contrast of the image. The top sliders look at coarser changes and the lower ones finer ones. You will notice that for a given movement of sliders the top ones will produce a more noticeable change, the lower ones are more subtle. The important point is that they are acting on different parts of the image so you can bring out more by using them in combination.

The key thing with the sliders is not to over use them. If any graininess develops move them back a little. Over use of the wavelets produces an unnatural looking image. Experiment – remember you have saved your project file so you can always reload that if you mess up your image.

Whilst on the wavelet screen you can also experiment with the brightness, contrast and histogram stretch. Once happy select the “final” tab which allows a few more adjustments.

Now you can save your image.

Manual frame selection

Registax makes an automatic selection of the best quality frames to use but this can often not match what we are looking for. Typically on Saturn we want to see a nice sharp Cassini and banding on the disc. Registax algorithms don’t necessarily think the same. Under manual processing with Registax it is possible to have a look at the frame selection and it sometimes looks pretty ropey. Despite this it seems to make a pretty good fist of the final image. It is possible to spend hours manually selecting frames for only very subtle improvements. However in conditions of poor seeing it is often worth a bit of effort. As I have said earlier, virtual dub is a much quicker tool to use than Registax.

Launch Virtual Dub and select you AVI. There are normally 2 frames containing your image. Just focus on the one on the left.

Underneath the image is a slider bar and this will indicate the number of frames. Below this is a set of video buttons. To the left is an obvious stop play button. To the right of these are 2 play buttons. I haven’t bothered to explore why there are 2 other that the fact that the software is designed for more involved editing of video that we need. To the right of these are a back to beginning button and then 2 buttons which look like fast reverse and fast forward buttons. Actually these advance the frame back or forward one at time. Then a go to end button. There are then some coloured video editing buttons which aren’t of interest to us. The final 2 buttons allow us to select a block of frames for deletion.

You can advance 1 frame at a time deleting the ones you don’t like (by pressing the delete button) or for a quick starter you can play the video to identify blocks for deletion. If seeing was poor you may be looking for the few clear blocks to leave! To select a block for deletion use the 2 buttons to the right as mentioned above. Use the left hand button to select the first frame to be blocked and then the right hand button to mark the end of the block. Press ctrl x to delete or click on edit delete.

You should be trying for 200 useable frames.

Once you have trimmed down your frame set you need to save the AVI.

Relaunch registax but this time having brought up your file and set the alignment box reduce the lowest quality box to 0% .

Registax seems to run through these Virtual Dub AVIs quite slowly, don’t be concerned, time for another cup of coffee.

Photoshop quick tips

These tips also work in photoshop elements and other software such as paintshop pro contain similar tools. Covering the range of tweaks possible in photoshop is well beyond me but here are a few quick suggestions that I have picked up.

Black sky – Enhance|adjust colour|replace colour. Use the dropper tool and click on a blank area of sky. Slide the lightness control right over to the left.

Colour cast – I haven’t found the cast corrector in photoshop, can be done manually under adjustment of hue and saturation using the master/r/g/b channel selector. In elements under adjust colour click on colour cast then click the dropper tool on an area of the image that should be (but isn’t) grey white or black. Click on the rings for images of Saturn.

Adjust rings and disc independently – on Saturn the rings tend to be under exposed and the disc over exposed. Also the rings can and disc can tolerate a different set of adjustments. Use the magnetic lasso tool to encircle and select the disc. Try working on increasing the banding detail by increasing contrast whilst lowering brightness. Then choose select|invert and any further changes will be applied changes to the rings. Again play with the brightness and contrast. You can also try a bit more sharpening.

Crop you image a little before resizing. Make sure resample image is selected or the image will simply pixelate. Do it in gradual increments and don't push it too far. As you resize the image becomes fuzzier. Often the best looking images are ones that are at their original size.

Go steady with photoshop, if in doubt underdo the changes. The image can end up looking very artificial.

This is my first Saturn. Using a Toucam and a 70mm Lidl refractor on a freezing cold windy night. Capturing and processing this image took more effort and pain than all my others put together! Notice that I didn't know about the RGB shift at this time.

This was taken with a Tal1 with good seeing. The mount isn't motorised. Manual selection of frames did produce some improvement

Using a Nexstar 8" SCT. Awful seeing. This image has had careful manual selection of frames and lots of work in photoshop. I spent hours trying to produce something out of the fuzz

Another with the NS8" this time with good seeing. All autoprocessed by registax and nothing done to it in photoshop

I hope this helps. Please let me know if I seem to have got anything wrong, omitted something vital or have confused anyone :?

Martin