Collecting satellite positional data using video frames and HNS_FITS

For satellite orbit determination.

2002-11-20 by Greg Roberts.

Scattered around the world, but predominantly in the northern hemisphere, there is a small group of amateur satellite trackers who specialize in the detection and tracking of satellites for which no orbital data is available from the official satellite tracking agencies. Most of these satellites- about 100 in number- are classified military satellites used for reconnaissance purposes of one type or another and range in size from Hubble Telescope sized satellites down to microsats the size of a hatbox.

Since the first satellite was launched in 1957 amateurs have tracked satellites by optical and radio means. The traditional way to track an optically visible satellite was to use a pair of binoculars and a stopwatch and determine the time at which a satellite was is a position, relative to the background stars, that could be easily identified. The position was then determined and, along with the time at this position, was sent to a computing centre which would then update or derive so called orbital elements which would enable the satellite to be tracked in the near future. By constantly getting new positional data the satellite could be tracked on a regular basis by anyone suitable equipped to see the satellite.

In the 1970's the policy of issuing data on all satellites launched was stopped and data for sensitive satellites was not available to the general public and thus grew a bunch of amateur "detectives" who made it their hobby to locate and maintain their own orbital data so that these satellites could be regularly tracked. They also used binoculars and stopwatches but with the fairly recent availability of low level light sensitive CCD cameras - similar to what one sees in supermarkets or banks, a new technique is starting to be used as it offers greater accuracy. A satellite pass is recorded on video tape using a sensitive camera and the time of each frame is stamped electronically on the picture. This has improved the timing accuracy considerably as ones personal error, or re-action time, is no longer applicable , so all that remains to be done is to identify the star position.

There are several ways of doing this and it can be quite a long process but with the availability of Han's HNS_FITS   program matters have been considerably improved. It is usually a simple matter to identify two stars in the video frame and thus derive the position of the satellite. HNS_FITS assumes a linear relationship in the field of view which is not 100% correct since one may have field curvature etc but with reasonable care the answers given by HNS_FITS are more than accurate enough for satellite orbit determination purposes. The time and positional data is sent to other members of the group and new orbital data is derived to enable continued tracking of the objects of interest.

I believe I am the first to make use of this CCD technology and the entire system, apart from the CCD chip, is homemade. The system is called CoSaTrak ( Computerised Satellite Tracking) and uses an altazimuth mounting fitted with stepper motors on the azimuth and elevation axis controlled by a freeware computer program written by my associate Willie Koorts. The prediction program to generate the data required was written by me and is called SeeCoSat and is also freeware. The CCD camera at the moment is known as the 1004x and has a 1/3 inch detector with a sensitivity of around 0.003 lux. With my favourite lens system - an 80mm focus f/2 lens equipped with a homemade focal reducer to give an effective f/ratio of f/1.6 I have a field of view of about 2.5 x 3.5 degrees and in my bright city lit sky - I live not far from the centre of Cape Town. I can see stars down to magnitude +9 with the normal camera speed of 1/50th second. The system has proved very efficient and several satellites that have been lost through lack of observations have been recovered as well as locating other satellites of interest. The system is under constant development and improvement. A webpage describing the early days of the CoSaTrak project may be found at

Due to pressure of work etc it is somewhat outdated but it will give an idea of what has been done and some of the software is available on this page.

Greg Roberts