In the recent NROL-76 SpaceX launch there are some really spectacular shots of the rocket blasting into the atmosphere and coming back down.

For example, at about the 14:17 mark there's a shot of the rocket stages separating when they are going well over 4 times the speed of sound and are 70km high and climbing:

SpaceX rocket separation

They obviously have a few cameras since the view is different at various times like 13:30 and 20:14.

My question is how do they so accurately track and record the rocket when it is so far away and moving so fast? What type of camera might they use and how is it operated?

I would think it's difficult to track so accurately manually so maybe there is a GPS tracking camera on actuators that can follow the rocket. Perhaps the footage is even taken from a plane or helicopter and not from the ground.

(Note that I'm not asking about the cameras mounted to the rocket itself, just the ones tracking it from a great distance.)

  • 6
    \$\begingroup\$ Cameras were tracking rocket launches like this long before GPS existed. \$\endgroup\$
    – Michael C
    Commented May 2, 2017 at 0:33
  • \$\begingroup\$ Most stars moves faster than any rocket we've made and further than any rocket has travelled, but aren't they very easy to "track"? \$\endgroup\$ Commented May 2, 2017 at 12:23
  • 4
    \$\begingroup\$ @user3528438 Tracking movements are measured in angular degrees. In that respect, no, stars don't move "faster" than any rocket we've ever made. They move approximately 15° per hour. These types of systems can routinely track objects moving 15° with respect to the camera in small fractions of seconds. \$\endgroup\$
    – Michael C
    Commented May 2, 2017 at 12:38

2 Answers 2


If it was launched at the Kennedy Space Center in Florida, there are ten of these Contraves-Goerz Kineto Tracking Mounts spread around the complex. Contraves-Goerz is now a part of the L-3 Brashear company. Multiple cameras are attached and all the cameras on each mount are precisely aimed on parallel optical axes. Other launch facilities use similar, if not identical, hardware, though perhaps in fewer total numbers per facility.


The earliest similar types of mounts used to track and film rocket launches were based on early naval antiaircraft gun mounts. Some of these early versions had spotting telescopes used to optically aim the contraption.

Today a lot of the tracking functions are automated or semi-automated, but back when it all first started the skill of the operator was responsible for keeping the camera array pointed in the right direction. Just as with those old anti-aircraft guns, the thing was aimed by using sets of control wheels, one set for altitude and one set for azimuth. These control wheels are similar to what one would see on large, crane mounted cinema cameras on a movie set during the same time period. The rocket tracking control wheels had both higher and lower ratio gearing that made both coarse and precise adjustments possible. A full 360° turn of the precision wheel might only move the mount a fraction of 1°!

Similar (near identical) systems are used to track and record other high speed flying objects. This would include military hardware during testing and developement (planes, missiles, even large artillery projectiles such as those launched by a tank or larger artillery guns, etc.)

You can read all about the (non-classified parts of the) workings and performance of these systems here.

This document from NASA explains how they are used at the Kennedy Space Center.

  • \$\begingroup\$ Might just add that the algorithms which allow an automated tracker, once locked onto a target in the camera's FOV, to stay locked are well-understood and have been around since at least the 60s. \$\endgroup\$ Commented May 2, 2017 at 12:33
  • \$\begingroup\$ @CarlWitthoft A lot of that type of info is in the two linked documents. There's no need to write a book here. \$\endgroup\$
    – Michael C
    Commented May 2, 2017 at 12:40
  • 9
    \$\begingroup\$ @CarlWitthoft Actually the systems are still quite manual. There's never been any real demand for advance automated tracking since we have the skilled operators who've been doing it for their entire careers. The cameras will continue along a straight (in polar coords.) trajectory if the controls are released momentarily, but basically it's just a guy with a joystick. (source: I work at NASA) also relevant: the gov. is more willing to keep civil servants around for semi-obsolete jobs rather than to buy new equipment. \$\endgroup\$
    – Aaron
    Commented May 2, 2017 at 15:25
  • \$\begingroup\$ I will add this mainly holds true to short and medium range trackers. there are always new things being tried out, but those are the workhorse observation stations \$\endgroup\$
    – Aaron
    Commented May 2, 2017 at 15:28
  • 12
    \$\begingroup\$ Talk about gear envy ;) \$\endgroup\$ Commented May 2, 2017 at 17:14

Michael Clark has some very good information which is mostly correct. These trackers are manually operated and the operators ARE THAT GOOD. Several forms of automation have been discussed and I have written tracker automation code for the KTMs. We have also looked into off the shelf automation from various vendors. Maybe one day, but you only get one shot at each launch! It will take a lot of testing before we put our faith in automation!

  • 2
    \$\begingroup\$ Please put a link to the source you mention and cite the important passages, so that we know what you mean, even in case Clark's website is offline. \$\endgroup\$
    – FarO
    Commented Feb 2, 2018 at 9:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.