Measuring Speed by Video

[rcs2022]

Currently my eagletree is out, so I don't have any means to measure the speed of my savage except by taping my phone on it and using the phone GPS, which not that good of an idea, and accuracy dunno how accurate the phone will be. The speed radar gun i surmise is the most accurate way, second comes an accurate GPS, I was relying on my GPS module hooked up to my eagletree, should be quite accurate but had no means to verify.

Anyway, without any of the above available, thought of using a video camera to do so, here's how:

Placed the camera close to ground to be at the level of the savage, tried to do the runs perpendicular to the FOV of the camera, after recording the video, did some screen captures of successive frames as such:
















Combine alltogether in a photo editing software, merging the layers in one layer as so:



Use the ruler option and set vertical guides in the center of the wheels for some frames:





Knowing the length of the wheel base of my truck is 39.5cm, and the video is 30 frames per second, can calculate the total length traveled in the 5/30 seconds, and after some calculations....... i get 83 km/hr, should be accurate enough, the run was quite perpendicular to the cam.

Note my truck with the 1717 and XL2 is geared for 104.6 km/hr @3.7V

I'd like to test this theory versus my eagletree gps when i get it back working. Thought of sharing it in the meantime.

[_paralyzed_]

Just tape your phone to the truck, gps is pretty accurate.

[rcs2022]

Yeah I know, but strapping a 500usd or more phone to a rc truck at more than 60 kmhr; hmmm….

I'd rather stick to my eagletree, can get more data for less than 100usd. and it can stay there

But nonetheless was just something I was experimenting with, and feel its quite accurate.

[bruce750i]

I've tried this also with a high speed camera and 1 foot indicators in the background. I like your wheelbase method better for eliminating the skewed camera angles of the distance from the vehicle to background measuring device. I was a little scared of hitting my camera on the ground. I never did analyzed that video or run the formula to figure the speed. Myth Busters use this method all the time.

[brainanator]

Yes, this is done on mythbusters all the time. I don't know enough about cameras to know if the low-end stuff I have is a constant 30fps or if there are any jitters in the framerate. With such a small amount of samples, any deviation could lead to a lot of error in the end.
Assuming the camera is consistent, I think this method should be pretty solid!

[Thirdgen89GTA]

Your method is right on, but you are making it too complicated. You have no need to know the wheelbase of the vehicle as the wheelbase does not change. You only need to have a recognizable object marker for measurement. Then you just need to divide the Distance travelled by the elapsed time between A and B.

If it covered 50ft in .5 seconds that works out to 68mph.

Rather than use markers on the curb, use markers on the pavement directly. Two pieces of tape set apart by a known length is adequate. The lens distortion is still there, but because the close proxmity of the mark to the vehicle you don't need to worry about distortion, it becomes a moot point.

[millzee]

show the video at full speed and we'll tell ya if it's quick, lol.