Missing Nix frames

[John Dolva]

Thank you Antti. A reasonable guess. As Frank hinted at "I was going to postulate about 105 Kph, but I didn't use any mathematical methods to arrive at that conclusion.". This is precicely it. The way I see it, one or two questions need to be made. (what are they) Then the mathematical method will yield a part of the overall formula(e) looked for here in single lens(or lens assembly) single photo in order to percieve (mathematically) a multi dimensional space wherein every object is correctly located in relation to the camera film surface and also the relation between objects within the photo. Then this can be universally applied.

Frank, your guesstimate, and the reasoning for it goes down the path that all guesses go > to where they are argued. This, (and the reasoning is sound (though I can think of counter-arguments, you're assuming things, it's not a guessing game.) (I know you know that ) still, I must say that while your guesstimate is surprisingly hot, it does not answer that which I think CAN be answered very accurately. It's the PROCESS of deriving the answer mathematically that will yield far more than the correct speed. What is needed to be known to precicely mathematically derive the speed?

ed : ps Antti, perhaps you percieved I drove somewhat rallystyle (of which the finns repeatedly reign supreme), ie ignore the tracks most drivers use and preempt and follow that which will be the shorter distance thus perhaps giving a 'sense' of speed?

fa John, There are several tools that come to mind that could help with the math side of things.

Here are a few things that come to mind that I would ask about:

What type of vehicle were you driving?

What brand/model of headlights are fitted in the vehicle, and what is their age?

What type of camera was used to film the drive?

Were any frame rate, aspect-ratio, or other format conversions involved?

How and where was the camera attached to the car?

It would be informative to know:

Where you were driving (precisely).

Here is where I'm going with these questions:

One method to ascertain distance vs. time, which should reveal speed, would be to research the headlights, their throw, and try to derive the relative distance that an object is initially illuminated before it passes out of the field of view of the camera. This could then be timed and a speed computed.

Second thought would be to go back to the location of the drive, and measure the distance between various roadsigns, guideposts, etc.

The information on the camera is needed in all cases to (hopefully) understand the optical behavior of the lens.

jd Frank, I think we're getting somewhere. Some q's I'll have to think about answering, like where I am or anything related to that. (However I don't hink it is critical). If you can let me know why, it would help. I can say now I held the camera outside the window, though that's prob not much help.

There is no deliberate conversion, which is why I dumped the 10 meg in toto as is from the memory stick.

PERSPECTIVE obeys certain laws. (rip one frame and check it out?)

IMO the following are important variables to quote (slightly out of context perhaps)

"measure the ( ed: >A< distance, which distance?) distance".

A known distance on a plane paralell to the camera lens(?), preferably, but not neccessarily, at the centre to avoid distortions, (however, knowing lens and 'the formula' or part thereof, one should be able to deduce parspectives, obliqueness, and hence 'the matrix' that forms unit nodes (in space and time (..."ascertain distance vs. time"...) in the area filmed and therefore answer many of the questions.) (Some will be answered in surprising ways, I think.)

"The information on the camera is needed in all cases to (hopefully) understand the optical behavior of the lens."

yes. So what would a formula (or flow chart) roughly look like, using proper variables and unknowns that need to be found out??? I'm hoping the >frame rate< won't be necessary! (though it is a stock standard default settings Canon powershot A70), which is too easy. (remember, there is a hope 'it' can be used in single photo's too).

x = ( ? )

fa The "where" question would be only to determine the tilt and inclination of the road, and the spacing between road markers, etc. That is really all that is needed.

Frame rate is required and critical, because the time variable would be derived from it.

Basically, we have to compute s = d/t where s is speed, d is distance, and t is time (assuming for a moment that no acceleration is involved.)

The computation of time is derived from the frame rate of the video (how many frames were shot, etc, etc.) That way, we can allow for different frame rate cameras, etc, etc.

The computation for distance is where things get quite tricky...

________________

OK, Frank. (frame rate is knowable) so we're zeroing in on a formula. It has components d and t.

t is relatively easy. (the cruise control is variable from set +/- a couple of kph's). t = (formula)?

d is the thing.

("The computation for distance is where things get quite tricky.." yup, that's it!)

I will give the exact (+/-) distance of anything, just ask, I'll head down the highway measure, and post it. My suggestion is that once one knows one distance beyond dispute of ANY one thing, then having the z axis of the film, one can step out from there and derive all distances. eg, if one knew the width of the central double line, one can then deduce all other distances, including the location of the camera in relation to them. (I think pref. in 3D*) : d = (formula)?

( eds' : continuity from prev. page ; add* )

- post scriptum :

s ( and keeping in mind x ) = (formula(e) for any 'd') / (formula(e) for 't') ???

for a single frame x = (formula?) this formula (IMO) is dependent on ' formula(e) for any 'd' '.

Here, single frame blurs in a 3D unit 'matrix', again derived from an understanding of perspectives, ie potentially knowables, may, just by knowing exposure time, lead to s. ???

(gotta go... to preempt start, will return with average width from outside lines of the double lines)

Edited February 13, 2009 by John Dolva

[John Dolva]

average over three points, some meters apart, :: 9 3/8" +/- 3/8"

[John Dolva]

[John Dolva]

A hopefully inspiring perspective. The grid between the green dots are uniform, the hanging cables (pink) give the vertical. The vanishing point is the centre of the photo (lens z axis)

[John Dolva]

Still looking for the overall formula

sample001, how wide is the upright, how high is the hand rail, what time is it?

[John Dolva]

Still looking for the overall formula

sample001, how wide is the upright, how high is the hand rail, what time is it?

bump

[John Dolva]

[J. Raymond Carroll]

Still looking for the overall formula

sample001, how wide is the upright, how high is the hand rail, what time is it?

I read this fascinating thread from start to finish a few weeks ago and just got back to it today. Am I correct in thinking that the current exercise is an attempt to gauge/estimate the speed of the limo at various stages as it descended down Elm Street?

[John Dolva]

Raymond, yes, that is a definite part of it.

The idea is that there may be clearly definable variables that deal with perspectives, lens/film surface factors, (hence lens distortions) and other variables to be identified so that perhaps a universal formula can be produced that exactly places the film surface and hence objects within the photo and their relationships to each other. When one takes a series, such as from a film, the time dimension becomes a factor. There one needs to know windown rates, frame exposure times, frame speed or the time between frames exposure. Then one may move on to such things as precise timings and film/photo synchronisations. So, it's a wor in progress with a hope that useful universal formulae can be developed where one simply enters the necessary variables and derive a result depending on which formula or combinations one uses for a range of purposes. I think it would be good if we can independently derive such and have it as a contribution to the image research with all the implications one may encounter.

[J. Raymond Carroll]

I think it would be good if we can independently derive such and have it as a contribution to the image research with all the implications one may encounter.

Thank you John and good luck to you and your colleagues on this project.

[John Dolva]

I think it would be good if we can independently derive such and have it as a contribution to the image research with all the implications one may encounter.

Thank you John and good luck to you and your colleagues on this project.

Raymond, I appreciate the comments and for helping to keep the topic alive. It's gonna take more than luck to get there.

[Frank Agbat]

John,

I haven't forgotten about this! I've been busy with work lately and just haven't had any quality, uninterrupted time...

[John Dolva]

It's OK Frank, I know what it's like. (and to the extent one can on a forum like this, I know you as a scholar and a gentleman.)

(I'm also patient, and the time allows for half germinated thoughts to be ruminated on and gradually developed, We've been at this for quite some time now haven't we, and it'll take more time.)

edit : typo

Edited March 6, 2009 by John Dolva

[John Dolva]

logging for future reference

http://local.wasp.uwa.edu.au/~pbourke/misc...lenscorrection/

addendum : plus this from link from above page. (Seeing as he's UWA (Uni of West Oz) he's only a couple of hours away. Might go and see if he has time for someone not swimming in alphabet soup.)

http://local.wasp.uwa.edu.au/~pbourke/allpages.html

Edited March 10, 2009 by John Dolva

[John Dolva]

interesting how the recent forum format changes of various kinds introduces discontinuity. 5 posts per page, not 14 anymore?