Sunday, January 22, 2012

Open CV

Im going to step through a simple example with OpenCV 2.3.1 Mega Pack on Visual Studio 2010 and Windows 7 x64. I got the run around from searching around on the web for the past couple of days, so Im sure everyone will benefit from this solution.

I followed several tutorials step by step, with some changes, and it finally works. You must make the changes to get OpenCV 2.3.1 working. I'll walk you through the whole tutorial, to guarantee you dont miss anything.

Step 1: Install OpenCV 2.0

Go to the SourceForge page of OpenCV and install OpenCV 2.3.1 Mega Pack for Windows. Install the package to 'C:\OpenCV2.3.1'. If you look in the directory, you'll see a folder called build. We need to use that compile OpenCV.

Step 2: Compile Everything

Before you start working with Visual Studio, you need to compile the code into a VS project. You can do that using CMake.

Once you install, start up the command line prompt. (Win+R, type cmd and press enter). Type cmake and you should be able to see something on screen.

Next, type in this line:
cd C:\OpenCV2.3.1
mkdir vs2010_build
cd vs2010_build
cmake -D:CMAKE_BUILD_TYPE=RELEASE C:\OpenCV2.3.1
You’ll see a lot of things happening. And after a while, the process would complete, and you'll have a Visual Studio Project for OpenCV.

Step 3: Compile the project
This one is simple. Double click the OpenCV project, and compile it in Visual Studio 2010 (or Express). It’ll take a lot of time. Compiling the samples and the entire library itself takes a real long time.


Step 4: Sample Project
Then start a new instance of Microsoft Visual Studio 2010 ( or Express).
  • File -> New -> Project
  • Name: 'C:\Tutorials\OpenCV\OpenCV_Hello'...'OK'...'Finish'
  • Use the code below:
// OpenCV_Hello.cpp : 
// Microsoft Visual C++ 2010 Express and OpenCV 2.3.1

#include "stdafx.h"

#include "cv.h"
#include "highgui.h"

int _tmain(int argc, _TCHAR* argv[])
{
        IplImage *img = cvLoadImage("lenna_small.png");
        cvNamedWindow("Image:",1);
        cvShowImage("Image:",img);

        cvWaitKey();
        cvDestroyWindow("Image:");
        cvReleaseImage(&img);

        return 0;
}
Now comes the important part. Pay attention to this, or you'll be dealing with the same errors I got.

Step 5: Configure Project Directories
  • In VS 2010, Go to 
    • Project -> OpenCV_Helloworld Properties...Configuration Properties -> VC++ Directories
  •  Executable Directories -> use the drop down menu -> ... add: 'C:\OpenCV2.3.1\vs2010_build\bin\Debug' 
  • Include Directories -> use the drop down menu -> ... add: 'C:\OpenCV2.3.1\include\opencv;' 
  • Library Directories -> use the drop down menu -> ... add: 'C:\OpenCV2.3.1\vs2010_build\lib\Debug' 
  • Source Directories -> use the drop down menu -> ... add: 'C:\OpenCV2.3.1\include\opencv;' 
  • Linker ->; Input -> Additional Dependencies... [Use the top left drop down menu to find each build, i.e. where it says Active (Debug)]
    • For Debug Builds -> use the drop down menu -> ... add: 'opencv_core231.lib; opencv_ml231.lib; opencv_highgui231.lib' 

    • For Release Builds -> use the drop down menu -> ... add: 'opencv_core231.lib; opencv_ml231.lib; opencv_highgui231.lib' 

    • Ignore all Default Libraries -> 'No'  


  • Linker -> General-> Additional Dependencies -> use the drop down menu -> ... add: 'C:\OpenCV2.3.1\include\opencv;C:\OpenCV2.3.1\build\x64\vc10\bin;'

Once you have these configurations, there is one last thing you need to do. This is a critical step, or the program wont compile.

  • Go to Build -> Configuration Manager -> Platform -> use the drop down menu to select x64 instead of x86.
  • Now Build the project, and it should compile pretty quick. Then just run the project via the Debug menu. And thats it! You should get an image box with your sample image.

In my search online for OpenCV tutorials and code, Ive come across quite a lot of previous work. An image crop of Lenna for OpenCV purposes can be found here. The following websites were directly used (including images and code) for this tutorial:

  1. http://www.aishack.in/2010/02/hello-world-with-images/
  2. http://www.aishack.in/2010/03/installing-and-configuring-opencv-2-0-on-windows
  3. http://opencv.willowgarage.com/wiki/VisualC%2B%2B_VS2010
  4. http://stackoverflow.com/questions/7011238/opencv-2-3-c-visual-studio-2010
  5. http://abdullahakay.blogspot.com/2011/08/opencv-23-linker-error-with-vs2010.html
  6. http://siddhantahuja.wordpress.com/2011/07/18/getting-started-with-opencv-2-3-in-microsoft-visual-studio-2010-in-windows-7-64-bit/
If you have no idea what OpenCV is or how to use it, I found a pretty nifty set of tutorials here to get you started on Windows.

If you want to run OpenCV using Python on Windows, this blog will show you the way. If you want to run OpenCV using Processing and Javascript, this blog will help; this is useful for running OpenCV on the Beagle Board xM and the BeagleBone. The OpenCV Java library is here.

One of the most useful tutorials that Ive found thus far has got to be the one by Sparkfun. Its not just an OpenCV tutorial, but it has a motion tracking pan/tilt camera with free code!

I know these sites were very beneficial for my education of OpenCV and for projects, and I hope they help you too.

Wednesday, January 18, 2012

Angstrom, Cloud 9, Oh My!

Well after a lot of troubleshooting with Windows, I finally managed to get the BeagleBone recognized on my Windows 7 computer (Vista wont see it at all no matter what I do). So as a storage device, it works. Next I tried to access it via serial via Putty, and that didnt seem to be working. Windows couldnt find the com port! So without the com port, I couldnt access Cloud 9 IDE to do python development, and try the sample code that turns an LED on and off.

Fortunately, I found the solution. Since all the drivers you install are not signed, you have to reboot your system, then before Windows boots (but after your BIOS screen shows), hit F8 a few times, and you'll load into the Windows boot options. Select 'Disable Unsigned Driver Enforcement' and continue. Once Windows loads up (and you login), the COM port should be visible in your hardware manager. Screenshot of F8 screen shown below (Vista and Win 7 have the same screen).



Picture acquired here.
Find out what number it is, then use Putty to login to the BeagleBoard, using the standard settings given on the BealgeBone site: '115200' bits per second, '8' data bits, 'None' parity, '1' stop bits, and 'none' flow control. That gets you into the Angstrom Linux shell. It takes a few minutes, but once it loads, login as root, and bam you're in.

One thing I noticed was once I logged in as root, I could now visit all the links that the BeagleBone Instructions talk about:

BeagleBone 101 presentation - http://192.168.7.2

This application is largely self explanatory. The source can be edited using the Cloud9 IDE. The application is 'bone101.js'.

GateOne - https://192.168.7.2

For documentation, please visit the on-line GateOne Documentation. Note: This installation might be a bit slow, but we are actively working on improving this with the author. 

Cloud9 IDE - http://192.168.7.2:3000

This development environment supports direct execution of JavaScript via Node.JS. Visit nodejs.org for information on programming in Node.JS. The IDE is pre-populated with the source and demos of the BoneScript project.

So, once you have access to Cloud 9 IDE, all you have to do is load the sample Blink.py program and then run it. You should see one of your LED's on the board start blinking. I havent figured out which pin is the other LED that comes on, but Im quite sure Im supposed to supply the LED. I'll post when Ive got that figured out.

I'll probably post a couple more tutorials on my progress with this board, as there is absolutely nothing on the internet to help anyone along (videos arent the same as instructional tutorials). So if anyone has any questions for me, please ask away, and I'll try to answer as best as I can.


Edit: If for some reason you have to turn off the computer, then remember F8 on the startup, and then once you're logged into Windows, restart the BeagleBone, by hitting the little reset button on the board. Give it some time (about 5 minutes) to reboot, and it should be good to go. You'll have to eject the BeagleBone drive again if you want to start development on the Cloud 9 IDE or on Angstrom.

Thursday, January 12, 2012

My BeagleBone Arrives!

Well my BeagleBone just arrived today, and Im quite excited to begin working with it. (photo obtained here)


This development platform is made by Texas Instruments, and features a whole host of features. Quote from Adafruit website:
At over 1.5 billion Dhrystone operations per second and vector floating point arithmetic operations, the BeagleBone is capable of not just interfacing to all of your robotics motor drivers, location or pressure sensors and 2D or 3D cameras, but also running OpenCV, OpenNI and other image collection and analysis software to recognize the objects around your robot and the gestures you might make to control it. Through HDMI, VGA or LCD expansion boards, it is capable of decoding and displaying mutliple video formats utilizing a completely open source software stack and synchronizing playback over Ethernet or USB with other BeagleBoards to create massive video walls. If what you are into is building 3D printers, then the BeagleBone has the extensive PWM capabilities, the on-chip Ethernet and the 3D rendering and manipulation capabilities all help you eliminate both your underpowered microcontroller-based controller board as well as that PC from your basement.
  • Board size: 3.4″ x 2.1″
  • Shipped with 2GB microSD card with the Angstrom Distribution with node.js and Cloud9 IDE
  • Single cable development environment with built-in FTDI-based serial/JTAG and on-board hub to give the same cable simultaneous access to a USB device port on the target processor
  • Industry standard 3.3V I/Os on the expansion headers with easy-to-use 0.1″ spacing
  • On-chip Ethernet, not off of USB
  • 256MB of DDR2
  • 700-MHz super-scalar ARM Cortex™-A8
  • Easier to clone thanks to larger pitch on BGA devices (0.8mm vs. 0.4mm), no package-on-package memories, standard DDR2 vs. LPDDR, integrated USB PHYs and more.
So, this will be the development board for my OpenCV Smart Security System. I found a whole host of security projects out there dealing with the Beagle Board, but only one project with security systems. However, mine ads facial recognition, motion tracking, camera tracking, and a whole list of other features. 


As far as my RFID work is concerned, Ive made slow progress. However, I have experimented with some examples, and have an idea how the RedBee functions now. I'll have to write some code to see if I can run it independent from the computer (RedBee to Arduino communications). That way I'll have an idea of how to integrate this into my project.