a very simple demo. adapted from CCV1.2, use AlexP’s CL NUI Platform driver, do threshold on the depth data. the rgb image is only for display

i’m sorry but in my country youtube and vimeo is blocked. so please visit http://v.youku.com/v_show/id_XMjM1MzE2MTY0.html to watch the video if you are interested.

attached is the relative codes.

1) install kinect’s driver
i’m using AlexP’s CL NUI Platform(version: CL-NUI-Platform-1.0.0.1121 (see the attachment below), it doesn’t work for the later version). After installation, run CLNUIDeviceTest.exe to see if kinect works correctly.

2) check out CCV from http://nuicode.svnrepository.com/svn/tbeta/trunk/tbeta/Windows
i’m using revision 210.

3) modify CCV following steps below:
3.1) put ofxKinectCLNUI under addons
3.2) put CLNUIDevice.dll under other\dlls
3.3) substitute ofxNCoreVision.h and ofxNCoreVision.cpp under addons\ofxNCore\src\Modules
3.4) substitute config.xml under apps\addonsExamples\VS2008\bin\data
3.5) substitute calibration.xml under apps\addonsExamples\VS2008\bin\data
3.6) substitute Community Core Vision.vcproj under apps\addonsExamples\VS2008

about the calibration, since we are doing kind of “air-touch”, i think it’s not necessary to do CCV’s original precise calibration point by point. so i manyally modified the calibration.xml, just map the four corners, i.e.
image screen
0,0 0,0
camWidth,0 1,0
camWidth,camHeight 1,1
0,camHeight 0,1

feedbacks and comments are welcome~

=======kinect-CCV with OpenNI driver========
since lots of people are using openni’s driver, below is the kinect-CCV with OpenNI’s driver.

1)install kinect driver
1.1)Uninstall all the old kinect drivers.
1.2)Install OpenNI: OpenNI-Bin-Win32-v1.0.0.25.exe download here：http://www.openni.org/downloadfiles/openni-binaries/21-stable
1.3)Install Sensor: SensorKinect-Win32-5.0.0.exe download here：http://dl.dbank.com/c059av7eg5
1.4)Install NITE:NITE-Win32-1.3.0.17.exe download here：http://www.openni.org/downloadfiles/openni-compliant-middleware-binaries/34-stable
CD-KEY:  0KOIk2JeIBYClPWVnMoRKn5cdY4=

the following steps are almost the same like the one with CLNUIPlatform’s driver.

2) check out CCV from http://nuicode.svnrepository.com/svn/tbeta/trunk/tbeta/Windows
i’m using revision 210.

3) modify CCV following steps below:
3.1) put ofxKinectOpenNI under addons
3.2) substitute ofxNCoreVision.h and ofxNCoreVision.cpp under addons\ofxNCore\src\Modules
3.3) substitute config.xml under apps\addonsExamples\VS2008\bin\data
3.4) substitute calibration.xml under apps\addonsExamples\VS2008\bin\data
3.5) substitute Community Core Vision.vcproj under apps\addonsExamples\VS2008

the attached file “kinect_openni_touch.rar” contains all the necessary file(src and exe). it should work if you directly put the exe with CCV1.2
enjoy~~~

its cool .....
you used the original CCV or you made some changes on it ?

Wow this is really awesome yaya… Could you share some on how you are getting the image from driver to CCV? Would love to get this in mainline build. There is a lot of potential here with mixing CCV+Kinect…

Cheers, Christian

This is great. Share with us.

woohooo finally!!!

the dark prince - 07 January 2011 10:19 AM

its cool .....
you used the original CCV or you made some changes on it ?

just some minor changes on CCV:
1) read image from kinect
2) threshold on the depth image to get a black/white image
3) copy black/white image into CCV’s processedImg
then the original CCV will take care all the other things.

Christian Moore - 07 January 2011 02:11 PM

Wow this is really awesome yaya… Could you share some on how you are getting the image from driver to CCV? Would love to get this in mainline build. There is a lot of potential here with mixing CCV+Kinect…

Cheers, Christian

wow, so pround that you appreciate this. i would definitely share the code. just need some time to clean the codes

@Yaya

Take your time, please let us know if you need any help testing or when your ready to release and we’ll help spread the word. I think it will be rather easy to create some controls to adjust the depth calibration like in this video: http://nuigc.com/kinectdepth

Is it ok if I put your kinect video on youtube? I will link to your original post and include video at top of this thread (Youku is rather slow in US)

A big thanks for taking the initiative on this I am sure it will help many members learn about Computer Vision with Depth Data. Its a honor to work with members like yourself and all the passionate people within the community.

PS Great avatar with panda

Christian Moore - 07 January 2011 11:24 PM

@Yaya

Take your time, please let us know if you need any help testing or when your ready to release and we’ll help spread the word. I think it will be rather easy to create some controls to adjust the depth calibration like in this video: http://nuigc.com/kinectdepth

Is it ok if I put your kinect video on youtube? I will link to your original post and include video at top of this thread (Youku is rather slow in US)

A big thanks for taking the initiative on this I am sure it will help many members learn about Computer Vision with Depth Data. Its a honor to work with members like yourself and all the passionate people within the community.



PS Great avatar with panda

well, youtube is blocked in my country. although i use VPN, it’s very very slow. so sadlly i can only watch the very first part of the video on http://nuigc.com/kinectdepth

about the depth calibration, do you mean letting the program adjust the depth threshold automatically? in my version i use two double values:kineticLowBound and kineticUpBound in the config file. i have some preliminary ideas about the automatic adjusting. i.e., in the initialization phase, people stand right before kinect and stick their two hands, the program increase the threshold from, say 1, to 1000, and record the value when there are two blobs(corresponding to the two hands), and also the value when there are only a big blob(corresponding to the whole body) and then use these two values to determine the final threshold. i think microsoft do this thing in XBox’s setup step.

p.s. feel free to put the video on youtube. i’m very happy to share with people

You can use proxy for you tube.
Can you put automatic and manual.

You are making big move.

filip911 - 08 January 2011 05:30 AM

You can use proxy for you tube.
Can you put automatic and manual.

You are making big move.

well, proxy is still slow…
anyway, you can try this beta version in my first post to see if there’s any bugs
automatic is the second step lol

The video looks great! I don’t have a kinect yet but I will definitely by one to try this out!

Do you think that is possible to run two versions of CCV at the same time with a PS3 eye and a kinect? The one with the PS3 eye could use TUIO UDP with accurate calibration to use the Multitouch vista driver and Windows 7 touch events. The one with Kinect would use FLASH XML for scaling, rotating and zooming functions in flash applications.

Regards

Fike - 10 January 2011 03:35 PM

The video looks great! I don’t have a kinect yet but I will definitely by one to try this out!

Do you think that is possible to run two versions of CCV at the same time with a PS3 eye and a kinect? The one with the PS3 eye could use TUIO UDP with accurate calibration to use the Multitouch vista driver and Windows 7 touch events. The one with Kinect would use FLASH XML for scaling, rotating and zooming functions in flash applications.

Regards

well, i don’t have a PS3, so i tried with kinect and a video file. one CCV run with kinect and sends FLASH, the other CCV runs with a video file and sends OSC. for this case, it works. so i think for you case, it should also work.

could you upload a compiled version also.

well, here is a compiled version for CCV with kinect

Thanks a lot! That is great news. I will buy a kinect asap to try it out and report the results!