Tuesday, 16 April 2013

An android based bot


- people seem to love this bot

- Android <---- BtBee -----> Arduino

- Use of Opencv in Android platform

   - object detection based on HSV values
   - HSV trainer
   - Bt comm with arduino ( use of BtBee )

- A video i made

 -




- planning to make updates to it....

  • feature detection
  • char recognition
  • template matching
  • face detection
  • face identification
  • voice to speech to cmd
  • robotic arm - for obj pick up
  • control over wifi

- EOF

Monday, 8 April 2013

Too Much of Life

For the past 2 weeks or so, there's been too much life not much linux.

My updates:

 -- installed tesseract OCR engine in Fedora  ( from repo )
     -  works on *.tiff images but not for png
     -  tried installing libpng, png dev libraries - nope - nothing
     -  even while "ocr"-ing some tiff images - fails horribly



-- working on Sudoku grabber app in Android
     -  inspired by [ http://www.aishack.in/2010/08/sudoku-grabber-with-opencv/ ]
     -  following this guide
        [ http://gaut.am/making-an-ocr-android-app-using-tesseract/ ]
         for porting tesseract engine to Android




-- learning VIM from
     -  www.derekwyatt.org
        [ http://www.derekwyatt.org/vim/vim-tutorial-videos/vim-novice-tutorial-videos/ ]


 - EOF

 
 

Thursday, 7 February 2013

Thy shall fly




planned to make my first flying bot next week - uses android smartphone as brain - uses camera to capture and stream live video - the video is viewed and bot is controlled from another android device (my tablet) via sshd - connected thro wifi - the body is hollow - simple light weight dc motor used - the plan is just rough - need to do research on design of flying bots - img proc can be done = fun - to be continued...

Tuesday, 5 February 2013

Android based HUD





i hav a really cool idea - a hud display using android phone - an app that uses camera to display the env before us in a transformed way such that even placed close to eyes we cld see that - finds all objs and faces in the frame - draws rectangles around them with an id num as1 2 3 - we use speech recog - so we just say the num and it zooms in and displays info bout that face or obj - the info is stored in local and remote db
synced @ free time - new objs and faces uploaded to remote in real time - 
info added @ remote and then synced...


how bout dat? - EOF

Saturday, 19 January 2013

a collage tool - f17






installed a collage tool in f17 - Fotowall - only tool available for free for linux - its a pretty cool app - simple, easy to use - never like gimp - installed in fedora thro the cmd - 

$ sudo yum install fotowall.x86_64

- EOF

Tuesday, 15 January 2013

more magic with WGET








code @ [http://paste.ubuntu.com/1536369/] - wanted to download android lectures @ that site 4 a very long time - normally use "download helper" firefox plugin - now that v hav "wget magic" - i used it in a bash script in a loop - grabs all the req videos - an added advantage - with "-c" option , download can resume even after being interrupted - more wget magic 2 come - :) - EOF



Monday, 14 January 2013

PythonCv in "pi"



pi - vnc server - fedora
running "wheezy" on pi - tightvncserver runs on pi - vncviewer on fedora (f17)


running contours.py
running a sample py script contours.py for opencv samples

someone posted a question in Raspberry pi fb group - about installing and usin the python interface of opencv in pi - i have installed and tested opencv in pi from both pre-built package and compiled from src c0de - never tried python interface of cv - now managed to install pythoncv on pi and tested a sample code - these are the steps 



1. Install opencv 
$ sudo apt-get install dev-opencv

2. Install python dependencies
$ sudo apt-get install python-numpy 
$ sudo apt-get install libpython2.6 python-dev python2.6-dev 

3. Install python-opencv
$ sudo apt-get install python-opencv

4. Go to samples/python/ folder of Opencv src code 
(and run a python script)
$ python contours.py


- EOF