Nedstat Belgium
Nedstat USA
Nedstat Belgium
Nedstat United Kingdom
Nedstat USA
The following elements brought me to this project.
The previous list may seem to be a nice collection of "things". But how are these "things" connected? Recently there was a message on the Connectix discussion list that explained how to calculate the exposure time for the Connectix B&W QuickCam. I realized that CCD integration times of more than 1 second could be used. See the B&W PC camera exposure calculation posting in the archives of the same discussion list. I also read a message that stated that the CCD inside the Connectix B&W QuickCam is a Texas Instruments TC255 device. This device is also used inside the SBIG ST-5 astronomical CCD camera (and other astronomical CCDs).
So, why not follow Friedt's example, and try to capture Comet Hale-Bopp?
Later on, I found out that this idea was not new, see Jeff Schwartz's page on the QuickCam and Comet Hyakutake.
Still later another astro-QC page became available later on: the AstroCam page by Walter Banks.
I guess other people also work on astrophotography with their QuickCam (B&W and/or color version). If you have a WWW page about it, mail me the URL now!
A drawback of the Connectix B&W QuickCam is that it is not cooled. That means that any QuickCam image recorded with a brightness setting greater than 225 shows speckles. A speckle looks like a star. This may lead to the conclusion that the QuickCam cannot be used for most astronomical purposes. Not true! I will explain how to get rid of the speckles below.
When preparing an astronomical observation run, one should not be afraid of the cold. Place the QuickCam outside for some time, well before the observations start, so that the camera can adapt to the environmental, ambient temperature. And, the colder it is outside, the better. For astronomical purposes, the temperature should be below 0 degrees Celsius.
Cooling CCD cameras is usually done with so-called Peltier elements. The CCD cameras as sold by SBIG, contain one or more such Peltier elements. One could think about disassembling the QuickCam, and adding a Peltier element. In my case, it took a very long time before opening my camera. If you too want to go beyond this point, you may refer to:
It is interesting to read the Texas Instruments specifications for the TC255 (PDF file, Adobe Acrobat reader needed!). From this product information sheet we learn that the TC255 operates between -10 and +45 degrees Celsius, so that one might think that the Connectix B&W QuickCam should not be cooled below -10 degrees Celsius! Nothing is less true: Meade uses the TC255 in its Pictors 208 and 216, and SBIG in its ST-5, while they cool well down to -20 degrees Celsius (depending on the ambient temperature). The TI data sheet also contains a highly interesting graph showing the CCD spectral responsivity at wavelengths between 400 and 1000 nm. By the way, data sheets of other Texas Instruments products (few of them also related to the TC255, others related to similar CCD elements) can be located through the TI data sheet locator.
The Connectix B&W QuickCam contains an infrared (= IR) filter. It is that greenish-bluish protective glass that you see in the small camera lens opening. Through this opening, and thus through this glass lens, passes all the radiation (light) towards the CCD element in the QuickCam. Advanced QuickCam users succeeded to remove the lens (often not without breaking the filter, so that this action becomes irreversible in many cases). It is reported that by removing the glass lens, the QuickCam becomes sensitive to infrared radiation.
I do not use the term light: light being the radiation from about 400nm to 700nm of the electromagnetic spectrum. Infrared is at longer wavelengths, beyond 700nm. One should realize that what the CCD element observes is in the near-infrared (from 700nm to about 1000nm), not in the thermal infrared (at about 12000nm) of the spectrum.
When the infrared filter is removed from the QuickCam, the camera can detect near-infrared radiation, such as that is used by infrared remote controls. It is also mentioned that the QuickCam is better for astronomical purposes without the infrared filter. That is true for several, if not most astronomical objects, but the camera is now open, and dust may enter the camera. It is also reported that the quality of daylight images is not substantially reduced by removing the filter.
I would suggest to indeed remove the filter to observe a comet. Before I did so, I still made remarkable images of Comet Hale-Bopp. For those who do want to remove the filter, I again refer to Hanno Mueller's explantion on disassembling the QuickCam.
In my case, I removed the IR filter by accident. While trying to hold the camera behind my telescope's lens, it fel on the ground, and splashed open. Now I could finally see the interior of my QuickCam! The camera still being connected to my PC, I saw that the electronics were all still functioning. I immediately started testing, and almost everything that I read about B&W QuickCams without the IR-filter passed the revue. And then, my telescope still being set-up now delivered the finest images of the moon that I ever made! For example:
After this, I built-in the round, "naked" Connectix B&W QuickCam PCB into plastic parts, added a part from a Barlow lens, so that the CCD can now be nicely and safely attached to my telescope, in the focuser, at the place where otherwise sits the lens. This lead to even more spectacular results.
If you want to see this picture of the moon, without any resizing, realize that the file is large (over 200 kBytes), but it is worth it.
The following picture shows, for comparison with a standard Connectix B&W QuickCam, my eye, as imaged from a distance of 50cm, with the same set-up, and processed the same image enhanced techniques.
Two important items:
When making astronomical CCD images, you should make a flat-field image. Such an image can be made during twilight. You make an image of a part of a clear sky, in which there is no structure (such as clouds, or stars); you can also make an image of a structureless wall. This is the flat-field image. This image serves to account for the sensor's differential response over the image. You subtract the flat-field from any image made during the observation later on, during the same night (with a simple subtraction algorithm, available in almost all image processing packages). The flat-field is a must when doing photometry (i.e. measuring absolute magnitudes of astronomical objects).
Then, during the observation run, you regularly make a dark image. That is an image with the CCD completely covered. This dark image too (just like for the flat-field) is subtracted from the target image(s). The dark image will remove speckles from the target image. This procedure particularly enhances images made with uncooled Connectix B&W QuickCam. Notice that dark images should be (theoretically) made before every target, since environmental (temperature) conditions change continually.
Practically, I found out that with the QuickCam, it is good to make at least 3, and even better 5 consecutive dark images (db1, db2, db3, …) before, and 3 (better 5) consecutive dark images (da1, da2, da3, …) after an astronomical image. During the (pre-) processing of the image, you make the mean dark image of db2 and db3 (and db4 and db5), leading to db, and the mean dark image of da2 and da3 (and da4 and da5), leading to da. You finally make the resulting dark image as the mean image of da and db. This resulting dark image is subtracted before any further processing of the astronomical image.
Although I did not yet work with a flat field image, I can imagine that following a same procedure as for the dark images, by computing the mean of two out of three flat images before and after the astronomical image would be preferable.
The underlying explanation for this could be that the Connectix QuickCam is not cooled, so that it is not at a constant temperature during the whole observation run. I also noticed that the first captured image of a series is often different from the following images. It could be that the electronics added to the CCD element is of a lower quality than that in astronomical CCD cameras (otherwise also using the Texas Instruments TC255), and that there is need for some relaxation time (i.e. during the first image of the series).
I mounted my 10x50 binocular on a tripod. I fixed my Connectix B&W QuickCam onto a part of the tripod (the head) that moves when I move the binocular. The Connectix B&W QuickCam is mounted so that it is continually positioned exactly behind one of the two lenses of the binocular. By doing so, I can use the other lens to point to the desired astronomical object.
During the test I successfully imaged the Orion nebula (magnitude 4), and was able to detect stars down to magnitude 6.8 (at least). Remember that I had not removed the IR filter from the QuickCam, and that I use a brightness setting of 254, corresponding to an exposure time of 2.62 seconds (operating in transfer mode 1, to obtain an image size of 320 by 240 pixels). I did not use a flat-field image. But I subtracted the indispensable dark image (and was surprised to see how the speckles disappeared, and the stars remained). I estimate that the total field of view is about 3 degrees. That is less than the 5 degree field of view that I have when observing visually with my binocular. This reduced observation field can be explained by accounting for the optical properties of the binocular, the distance between the last lens of the binocular and the CCD element in the QuickCam, but I do not elaborate on that subject here.
The first thing that I did after my Connectix B&W QuickCam accidentally splashed open, was making a new housing for it. Believe it or not, but I used parts of plastic feeding-bottles, which my kids do not need any more, since they are reaching the age of five (yes, you guessed it: twins, a boy, and a girl).
An evening of work, sawing through plastic bottles, screwing caps, cutting through soft plastic (to dampen the PCB movements), and putting everything together including a part of a Barlow lens, so that the construction nicely, firmly and safely fits into the focuser of my telescope, yielded a highly surprised wife, and splendid astronomical images (of the other stars in my life, besides my wife and the twins).
I was astonished about the amount of light that the telescope focused on the bare Connectix. The same procedure of acquiring dark images delivered superior quality, compared to working without dark images. Notice that my telescope is not equiped with a motor drive. Thus, I must keep the exposure times short. That is not a problem for the brighter objects, such as the moon, the brighter planets, the coma of comets, or the brighter variable stars.
Focusing is a big problem, just like for the (commercial) astronomical CCD cameras. I found that the following procedure is workable. Using a (rather) portable telescope on an equatorial mount, I always first visually point at the polar star to do the alignment. This star does not run out of the field of view, so that once in the field of view, one has all the time to put the CCD in place of the ocular. I reduce the camera's brightness stepwise, while adjusting the focus, so that the image of the polar star remains at its brightest. I continue until the image of the polar star almost disappears; when any change to the focus makes the polar star to disappear, the camera is correctly focused. During this time the camera has sufficiently cooled (adapted to the ambient temperature), and I make my first series of dark images. Then I swing the telescope to the desired astronomical object, adjust the brightness adequately, place the object at the edge of the field of view, and make a series of images, until the object disappears the opposite edge of the field. I turn the object back the other edge of the field of view by carefully turning the declination kknob, and make another series of images. Then comes a new series of dark images. Next I change the brightness (to obtain another exposure time), and first make a series of dark images for new brightness setting. A new series of images of the object, and so on.
While the setup with the binocular gave the brightest images of Comet Hale-Bopp, and the telescope gave the finest details of the coma, I also tested a setup with a standard camera lens. This gave an intermediate brightness. Do not expect that pointing to an astronomical object is easy with this setup! See the picture of my eye above to realize that this kind of setup images only a very small part of what the lens shows when attached to the photographic camera. But still, pointing to the desired object is a piece of cake, compared to the telescope setup.
I again used the tripod that I use for photographic work. On top I mounted a piece of art, once made by my uncle to place a home-made card-board telescope on an elementary mounting. Is it not remarkable that this masterpiece, made twenty or even thirty years ago by my uncle, who brought me to astronomy at the age of eight, now serves for my experiments? Thanks uncle Gaetan! I will pass you my original computer images as soon as possible, so that you view the work of your godchild on your ow computer…
There is a metal part, in which there is a whole that accepts the screw of the tripod. The metal part has a handle, and a system to modify the altitude (which I both do not use for the present purposes). To the metal part, a board part is attached, and on top of that another, thicker wooden plank, which is filed so that my uncle's telescope exactly fit into it. This is perfect to fix the camera in its plastic housing with an elastic. To prevent further "accidents", I attached a safety cord between the camera and the tripod. The camera is mounted onto the same wooden plank as the camera. The fact the diameter of the plastic housing of the camera almost corresponds to the diameter of the lens, makes them nicely lined up. The lens is focused to infinity. Focusing the system is done by slightly shifting the lens and/or the camera.
I worked on this project while Comet Hale-Bopp was already leaving northern hemisphere observers. Simply put: I could not obtain an image of the spectrum of Comet Hale-Bopp. Too bad! So, I pointed to other objects: Mars and Spica.
Here is the recipe. Use a setup with a binocular, or with a photographic lens. Obtain a prism (get it for example from a zenith prism, sold for refractor telescopes). Place the prism just in front of the objective, and completely cover the remaining part of the lens. Rotate the prism in such a direction that the spectrum moves perpendicularly to the comet with its tail, due to the movement of the stars.
After the successful image of the Orion nebula (still before the "accidental" removal of the original IR-filter and lens), I was extremely eager to capture Comet Hale-Bopp. After two mornings with mist, I finally saw the comet on Monday 10 March 1997. In fact, this was also my first visual observation of the comet. A splendid view!
I did three observation runs:
The third observation run is almost unusable due to too strong twilight.
First of all I want to share with you my excitement of that morning, by showing an animated GIF for the second observation run (5:05:32 to 5:05:58 UT). One can see the round observation field (with already some influence of the twilight, so that the background was no longer completely dark).
Nine images from a total of ten were selected: one image behaved strangely, and was excluded. The dark image was removed from each of the nine images. Then I played a bit with the histogram, as to improve the brightness and contrast. The nine images were combined with Microsoft's GIF Animator. Each image is displayed for one second. The sequences loops indefinitely. This results in the movement of the comet through the 3 degree observation field.
Next I tried out some image processing techniques.
The subtraction of the dark was performed on all images. A first technique is to arithmetically add images of an observation run. I also tried to multiply two consecutive images. This dramatically enhances the contrast. One may apply some edge enhancing, and sharpening techniques (or Laplace transformations). This revealed some finer structures in the tail of the comet. Using a combined addition of three consecutive images, and sharpening called my attention to some fine lines in or near the coma. I also used an addition of three images, conversion to a negative image, and a colorization. This put into evidence the curved structures near (or in) the coma. One might suppose that this is in fact a spiral, finding its origin somewhere near the nucleus. This is of course an interpretation, and the structure later on simply seemed to be an artifact, produced by dust on the lenses of the binocular.
One day after the "accidental" removal of the IR-filter, the sky was clear again, and I mounted my bare camera at my binocular, and pointed at Hale-Bopp. The following moving GIF shows the result.
This image was obtained 30 March at 19:42 UT, so just before perihelion, when Hale-Bopp was at its best. The much more spectacular view is not due to the brighter comet, but mostly due to the removal of the IR-filter.
This first image is made on 11 April 1997, 19h57m UT, with the QuickCam without any filters or lenses, except the main and secondary mirrors of the 4.5 inch F/8 unguided Newton telescope.
The same image embossed to show the spiral structure, followed by the sum two, three and four consecutive (in time) embossed images. A second shell becomes visible.
When looking at the above images, the more experienced astrophotographer will probably apply also some edge detection or directional filters to the original images. Such a edge detection or a directional filter will show, just like embossing, the finer structures in th ecoma of the comet. An example of an edge detection filter is the Laplacian filter. An example of a directional filter is the so-called Sobel filter. This gives
where from left to right: original image, Laplacian filter applied, Sobel filter applied.
In fact, the Sobel filter operates in a single direction (e.g. East to West), and can be applied in four directions (North to South, East to West, South to North, and West to East). These four resulting images can then be added, as to obtain a semi-undirectional filter. This gives respectively:
One can also construct a Sobel filter which is twice as strong as the original one. The finest details seem to be displayed when the resulting images are converted to negative images. This gives for a series of images for one observation run on 11 April 1997, 19h57m UT.
| Image | Original | Sobel in four directions | Double Sobel in four directions |
| Image 01 |
|
| 
|
| Image 02 | 
| 
| 
|
| Image 04 | 
| 
| 
|
| Image 05 | 
| 
| 
|
| Image 07 | 
| 
| 
|
| Image 10 | 
| 
| 
|
| Image 13 | 
| 
| 
|
| Images 02+04+05+07 | 
| 
| 
|
For the above table I selected the best images of an observation run (lasting less than a minute!). The Sobel filter and its variant show lots of artefacts, so that each image is different. From the images I selected again the four best ones, and added them (images 2, 4, 5, and 7), in order to remove as many artefacts as possible.
With the technique and setup described in paragraph 4.4, I succeeded to image the spectrum of Mars and Spica.
These images prove that it should have been possible to catch Hale-Bopp, with a simple prism, a standard 50mm lens, and a QuickCam. Regretfully Hale-Bopp was too faint, and later on below the horizon, before I could point to it.
While on holiday in Reuven (Netherlands, near the Dutch-Belgian border, between venlo and Roermond).
