SonotaCo.JP

[Pavdu]

I have one problem. Paths of some meteors are curved, and when I finish analyzing I have two trajectories for one meteor.

Can anybody help me with this?

[SonotaCo]

Hi Pavdu
UFOAnalyzer has many functions to improve the accuracy of recognition.
Though it is not perfect, as for meteors, methods below are effective and will bring correct result for most of the cases.

1. Analyze area tuning
At first, the analyze area is set automatically according to the brightness change of pixels.
But there may be noise or improper detected points that disrupt the recognition.
You can restricts the analyze area manually.
Narrower area will bring more accurate result.
"qm" in Mask Editor sheet is easiest tool to restrict the area. It is useful for many cases.

2. Detect level tuning
"ddl" in Main and Profile/Analyze sheet set the default detection threshold level of brightness change.
If you set ddl higher value, then the detected points will decrease and noises will be vanished.
It is effective when a straight object is recognized as "curve" or "flash".
If you set ddl lower value, then the detected points will increase ,lost points will appear and separated trajectory may be joined as one.
It is effective when a when a detected trajectory is too short or separated to multiple objects.

3. Leap length tuning
"leap" in Profile/Analyze sheet set the leap length(pixels) that should be joined.
If you set higher value like 100 then the far leaped trajectory may be joined as one. But it may increase the possibility of wrong connection of points.

4. Analyze time span tuning
"f1" in Profile/Analyze sheet set the first frame position where the analysis begins.
-1 is the default, it means the same timing of detection of UFOCapture(may be equivalent to head setting, 25 or 30).
If the earlier part of the trajectory is missing, set f1 as 10. It will recover the missing part.
-------
I think methods above may solve your problem,
but there is possibility of other reasons.
For example, if you are using fisheye lens and the plate adjustment is not correct, then the path would be recognized as truly curved.
Please let me see the actual sample if you cannot solve it.
-------
Cheers.

[Pavdu]

SonotaCo, thank you for your help. I, unfortunately, can't analyze this clip today but I have screenshot of that meteor. I wonder does this curved meteod trail maybe is caracteristic of meteor... Maybe it is not problem with lens or mask is not adjusted correctly. Is that possible?

This is screenshot of that meteor (and here it isn't analyzed). When I have analyzing this meteor the resalts were that this clip have two objects, both are meteors and both are recognised as J1_Cap.

I'm new in this, and i don't know english very well and that is one more problem, because I can't understand all in your english manual. And I don't speek japanese.

Do you maybe will write better manual (with more details) soon?

[SonotaCo]

[Pavdu]

OK, problem is solved by increasing of ddl to 11.
It is a true meteor and I will analyze another clip of the same meteor today, from another location.
That is the first meteor that will be analyzed seriously.