Flame colour analysis using Icy

[playf1....@gmail.com]

I am interested in locating software that will analyse a video file.

The video would be cropped to show only a flame changing colour.

A graph based display would be ideal, to show colour changes as a curve, according to oxygen mix.

Perhaps each frame would require 'averaging', due to the colour not being consistent.

Does Icy have any 'flame colour analysis' capabilities?

[Stephane]

Hi,

You have different image analysis tools in Icy and i guess you can easily analyze the color variation of a flame.
You can use for instance the "Intensity Profil" plugin to display the mean of a specified area (designed by a ROI) :

The plugin let you the possibility to export the result in a excel file.
I think that is a good start for what you are trying to do :)

Best,

- Stephane

[playf1....@gmail.com]

Stephane,

Thanks for that interesting response.

May I ask that you to go further in your explanation, by clarifying certain issues?

1. Cropping the video

By boxing, or masking, an area of the video frame..... this satisfies the cropping requirement.

2. On your example Excel Graph......... 'x' axis is the 'video frame number'

Is this correct?

IE. Each video frame has it's own 'intensity profile'?

From this, I presume your chart shows 1 second of activity @ 30 frames per second?

As you suggest..... this would be workable (assuming intensity average could be replaced by 'colour' average).

It may even be possible to insert the 'audio' time line in the excel chart, and stretch it to match the 'x' axis frame numbers.

Could this be possible?

We would then have eliminated all visual/video reference from the data processing.

Example

Increase fuel...... the curve falls to increasing towards yellow (numbers).

Increase air....... the curve rises to increasing towards blue (numbers).

In both examples, the sound would fall and rise, against the video frame numbers ('x' axis).

We would in any method used, then require a further time line chart, showing the change of the combustion mixture.

Probably 2 further time lines....... one showing changing air........ the other showing changing fuel.

This is probably not relevant to icy, but I include this observation for intellectual benefit. :) 

Stephane..... Is this how you envisaged the system operating?

Questions

Assuming that all this is possible NOW using Icy, given that we were averaging the 'intensity' of a flame, in each video frame.....

..... can we also do this for colour NOW?

OR.... does this require a minor tweak..... or major programming?

Colour?

We think of colour, as a change in wave length/height/frequency.

Given that we are NOT USING spectrum analysis........ clearly computers interpret colour in a different manner..... or am I wrong?

When the computer interprets the light falling upon the ccd chip...... it allocates each pixel with some numbers, representing what.... brightness and colour?

Are we interested in brightness?

OR

Do we separate the brightness number, and concentrate on 'colour'.

Considerations:

For those who may find this topic interesting, but have never thought of 'combustion colours'.

Be careful when searching google images for flame temperature colours..... there are many colour charts.... some showing yellow as hot, and blue cold...... others showing the opposite.

For us, we know that a yellow flame is poor combustion (so cooler), while a blue flame is good combustion (therefore hotter).

Thanks again for your interest.

:)