AIBO image capture

Using the URBI Remote SDK, the image capture was finished yesterday midnight. I integrated the URBI Remote SDK to myLibrary and the applications (CamView, FindShapes). The switching between the OpenCV/URBI interface works flawlessly, thus I can speak real AIBO support now. I removed the deprecated R-Code+ interface to AIBO, because it is restricted to the ERS-210 model. The new interface is ERS-2x0/ERS-7 compatible.

Állatsimogató

Zsófit ma elvittem egy állatsimogatóba, ami a közelünkben van. Biciklivel mentünk oda, és párszor majdnem elkapott az eső, de sikerült megúsznunk. A simogatónál volt egy kis állatos játszótér is, ahol lehetett játszani az állatok között.

URBI integration

I added the URBI (Universal Real-time Behavior Interface) Remote SDK 1.0 to the repository. I was able to migrate this library liburbi-cpp and its dependency libjpeg-6c to cmake quite easily, but I have some obstacles that the simple image capture example for AIBO worked under Linux, but not under Windows. After some researches, I found, that the standard fopen("filename.jpg", "w") saved the binary content as "binary" in Linux, but as "text" with fwrite(). The effect was that the 0x0a bytes were replaced with 0x0d0a as CRLF instead of LF.

AIBO arrived

My ERS-7 is arrived to my home today. I did not choose her name, but I have tested, she works very well.

Új szerződés

Ma írtam alá az új szerződésemet a Nokiához. Egy nagyon jó lehetőséget kaptam, július 14-én fogok kezdeni egy korábbi ismerős nokiás menedzseremnél, a kilátások nagyon jól néznek ki. Nagyon kecsegtető fizetési ajánlatot is kaptam, amit nem tudtam visszautasítani, illetve a szakmai fejlődés szempontjából is kifejezetten kiemelkedő. Pozitívan tekintek a jövőbe, gyakorlatilag ez az első igazi előrejutásom a "karrieremben" azóta, hogy itt élek Finnországban.

I bought an AIBO

I've won the auction of an Sony's ERS-7M2 robot dog on eBay. I'm very happy to end up in this situation, because it was my dream to use a well-constracted robot for my research without the troubles of the robot-building. I can control it via wireless, thus it makes really a lots of things quite easy.
Now I must pay it and wait for receiving it from Swissland...

How to develop Cognitive Vision under Windows

Here are the steps, which are needed to develop/compile my project with Eclipse under Windows:

Download and install the windows binary version of the following items:

- cmake 2.6
- doxygen
- MinGW (Install the g++ compiler+MinGW Make option!)
- Gtk+Glade development package 2.10.11
- Gtkmm+Glademm development package 2.10.11
- OpenCV 1.0 (ChOpenCV package does not need!)
- Eclipse for C/C++ Developers
- Subclipse plug-in for Eclipse: Install the components Subclipse, JavaHL adapter.
- Optional: CMakeEditor Eclipse Plugin

Import the Cognitive Vision from svn into Eclipse:

- Start Eclipse
- Select File menu/New/Other/SVN/Checkout project from SVN
- Create New repository location->Url: https://aibo.svn.sourceforge.net/svnroot/aibo/trunk/CV->Finish
- Select C++/C++ project->Project Name: CV (you must see MinGW GCC in Toolchains)->Finish
- Close the Eclipse
- Start the Start Menu/CMake/CMake-gui-beta
- Select the CV location in the Eclipse's workspace and copy that location to the binaries build directory.
- Configure button->Choose generator Eclipse CDT4 - MinGW Makefiles+Use defaults->Generate button
- You should see in the bottom part somewhere that the g++ compiler/OpenCV/Glib/GTK/Glade are found properly.
- Close the CMake-gui-beta
- Open Eclipse->Project menu->Build all

Enjoy!