Motiontrack News

well important to note that, while development is still "single-threaded", the program itself now can run multiple processes at once if the hardware supports it (cluster, smp machine, multiple cpus, ...)
possibly improving performace on these hardware.

also some minor bugs where fixed ;)

while only minor feature enhancements (sectorcheck is faster and outlines movement sectors now)
its important to note that the names of most executables have been altered to make them more unique, to prevent collission with other projects.
Users may have to alter them scripts accordingly ;)

and it had grown some neat features over the last half year if I may note it like that ;)

motiontrack can now automatically align and perspective correct 2 images to compare them.
This is beta code, so dont expect too much from it yet, but it has promising capabilities.

finally, 0.1.0 is out :-)

several things changed how they work from 0.0.x releases, so better read the documentation ;)

probably the last release before 0.1.0,
unless someone submits a bug report or important feature request ;)

originally i planned to release a bugfixed and optimized 0.0.9 as the 0.1.0, but then I just couldnt resist adding another feature (shrinking, enlarging images by factor 4) and now i decided, instead to name it 0.0.10

0.1.0 will get a greater improvement:
I am planning to implement a fuzzy pixel scaler, which will bring in tune-values to resize ( magnify and shrink will disappear, just give "resize" a positive or negative value ), blur, and maybe even enhance (the latter is a little bit more complicated)... read more

The speed-hack now works with ImageMagick, too,
using a seperate pixel buffer for all operations except loading and safing.
Motiontrack should now be practically usable with both libraries, GD and ImageMagick :-)
So we shall enter the bug exctinction phasenow, after that I may leap to the 0.1 release.

Whatever happens while using motiontrack, even a car crashing in front of your neighbours house, REPORT IT ;-)

Beta support for ImageMagick has been added, for all who had problems compiling libGD. Supports loads of image formats they say, -- that holds great potential I say :-)
And Motiontrack can handle all fileformats supportded by libGD now, not only PNG and JPG, but WBMP, GD, GD2, XPM and XBM, and maybe even GIFs! ;)
And it wont complain anymore if support for JPG or PNG fails, just be good and use some of the others :-)... read more

Thanks to Scott Hassan who submittewd a speed-improvement patch half a year ago, i recognized this as being an "important feature" and tweaked some code to be faster.

Corvus Corax

...well until i am through the AI and image recognition lecture here, and am able to improve things "a little"

or anyone else has a great idea and submits it, this way joining the project ;-)

...CvC

pattern detection has been implemented as first step on the way to object detection. 'asciiart' is the demo application which turns white on black line art into a nice ASCII char drawing.

The inclusion of several featured lead to a new version again.

almost a new project. internally.

the functionality is almost the same except the possibility to change the brightness via enlight or motiontrack -e
and the commandline-specifyable sector exclude list

but the internal objectoriented list structures are what rocks.
:-)

Corvus

I think 0.0.3 is the first widely useable version of motiontrack, since you can setup the motion sensitivity for your local needs and equipment now.

Yeah. I am proud of a new source tree.

The code has been scrambled around in several header and c files, which should make more sense now!
View it. And we have a 4th executable to do everything in 1 run. Called "motiontrack" of course!

Gimme Feedback if the calculation works with your cam! Please!

Today is a great day for mankind - or at least for me, since my first project made its way to sourceforge. I hope anyone might be able to use this :-)