motion analysis with windows media files

[jjwill04]

I have an unusual question...

I have created windows media files from video I captured with a sony miniDV camcorder. The video is of zooplankton (small crustaceans) swimming in a small plastic cube. I need to track these critters as they swim so I can calculate their swimming speeds. Does anyone have suggestions on software that might be able to do this? Ideally, I could pause the video every few seconds and click on the zooplankton and the software could give me x and y coordinates. If those coordinates are in pixels, that is fine as long as I can calibrate it to a known scale (there is a 1-cm grid on the back of the plastic cube throughout all the footage that would help with this). By knowing the coordinates of the zooplankton at successive points in time, I can determine how far it has traveled and how fast it swam. Any suggestions? Thanks for your help!

[JohnnyMalaria]

Have a search for "Particle Imaging Velocimetry" or PIV.

e.g., http://www.coriolis-legi.org/CIVPROJECT/main.html

[drewzor]

you could use virtualdub to set through each frame and as long as you have your measurements for both x and y, you can easily measure distance traveled. Measure each 2cm, X^2=2cm^2+Hight traveled^2 this will give you distance traveled and v-dub will give you the time code. Add all the times and all the distances and use Speed=Distance/Time which will give you the best average speed. Note you will be dealing with Cm per second

[SLK001]

These methods only give you the vector in the X-Y plane. There is also a Z plane that you aren't gathering data on that will really mess up your velocities. That is, of course, unless you properly instruct your zooplankton to swim only in the X-Y planes.

[drewzor]

great point! I live in a 2D world, sorry??

[jjwill04]

When filming the zooplankton in the cube, I actually took two video recordings using two video cameras placed orthogonally. So, I can find x,y, and z coordinates. However, I have to analyze each tape individually--one to get x and y, and one to get z. drewzor, I'm not sure if I understand what you are saying about virtualdub. I need something that will allow me to click on the image of the zooplankton and then have the software will give me the x,y coordinates. I couldn't tell by what you wrote if virtualdub will do that for me when I click or if I would have to calculate the coordinates myself by tracking the image. Finding the coordinates myself if possible, but it is would be much more time consuming and much less accurate than being able to click on the critter every few seconds for a variety of reasons. Thanks for the help. Keep the advice coming!

[JohnnyMalaria]

I still recommend PIV - it is *ideally suited* for your application and it seems there is a freeware solution out there. i.e., calculation of velocity fields for particles moving in a fluid....

Alternatively:

Overlay individual frames to create a single, composite image. You will be able to easily 'see' the paths taken by each organism.

Or:

You could process each frame individually to get co-ordinates of objects. You can use free software such as NIH Image ( http://rsb.info.nih.gov/nih-image/ ). It will allow you to automatically identify objects and list their co-ordinates.