Narrowband filtering is an excellent tool for obtaining useful data of weak signals in noisy environments.
Working for a small aerospace company, the problem of receiving weak telemetry signals is a regular challenge. Due to space and power constraints, telemetry output power is usually limited. To add to the challenge, the receiving environment is never the ideal RF sterile environment. Hence, narrowband filtering helps the ground receiving station pull-in the weak telemetry signals from the RF noisy locations.
Interestingly, the same situation occurs for those doing urban deep-sky astrophotography. Light pollution is analogous to RF noise and the optical signal of the object being photographed is analogous to the weak telemetry signals being received. The use of optical narrowband filters perform the same function to help differentiate the weak signal from the target being imaged.
My optical "antenna" is a wide-field astrograph with a fairly broad 530-mm focal length. This is coupled to a filter wheel to hold several optical filters and a dedicated cooled astro-CCD camera. This type of set-up is suited particularly well for imaging nebula. However, these nebula are typically very dim and are recorded (at least for me) in a light-polluted urban environment.
Emission nebula produce color that is characteristic of their chemical composition. The vast majority of these nebula contain hydrogen gas, which is typically ionized and produces a characteristic line-spectra emission. The most important emission line for astronomical imaging is the Hydrogen-alpha (Ha) line, at a wavelength of 656.28 nm. However, the signals from these objects can be very dim.
Narrow bandwidth optical filters allow this light to be efficiently captured with a high out-of-band rejection of the light pollution (which currently is mostly mercury and sodium vapor). This allows astrophotographers to image the hydrogen emission with good efficiency. The result is a monochrome image that contains the deep sky object, recorded in a single emission wavelength.
If multi-color images are desired, then other filters and emission lines are required to be imaged. This is solved by imaging similar characteristic emission lines of other ionized gasses that are typically present in nebula. Two of the more common (though less common and much less intense than Ha) are the forbidden lines of doubly ionized oxygen (OIII) and Sulfur II (SII). See the entry in Wikipedia for more information.
OIII is imaged at 501 nm (and sometimes 496 nm, depending on the filter’s band pass) and SII at 672 nm. As mentioned, these emissions are typically much fainter than the Ha emissions. The OIII emission, in particular, isn’t as well isolated from light pollution spectra and also is affected by the moon’s broad spectrum of reflected light. This creates a particularly difficult challenge. Removing the element of light pollution by imaging at a darker environment is the ideal choice, but usually requires a drive of a couple hours, maybe once or twice a month. My goal is to frequently image faint targets from his urban backyard.
Initially, filters with a bandwidth of 7 nm to 8 nm were used. While this worked well for the Ha emission, the high background levels swamped both the OIII and SII signals, making them all but invisible. As in the RF analogy, a narrower filter should help reject more of the out-of-band signal. The 7- and 8-nm filters were replaced with 3-nm units, which provide a lower background level and thus better contrast for the OIII and SII signals. Some of my images (both Monochrome and color) using this filtering technique, can be seen on my website.
Optical filtering is but one challenge with deep sky imaging. Since the sky has an apparent motion, accurate tracking is required. Cameras have noise that limits the image’s background level. Atmospheric distortion (both high- and low-frequency) requires other techniques to assist with target tracking. Temperature changes and the thermal characteristics of telescope materials require other solutions to maintain the system focus. Each problem is solved separately, yet work in unison to aid in providing the desired result -- good data in the form of pretty pictures.
— Eric Chesak is the Director of Engineering and Manufacturing Technology at TMD USA.