Image processing with android : Kanamaru Lab. in Kogakuin Univ.
Navigation
The omni-wheel robot with two web cameras controlled by Android phone

The omni-wheel robot with two web cameras is controlled by Nexus 5.
The mjpg movies from two cameras are streamed asynchronously,
and they are viewed on Nexus 5. If you use Google cardboard compatible viewer,
such as Taovisor, you can view the 3D view.
To control the omni-wheel robot, Odroid U3 and Arduino mini pro are used.

Target tracking with omni-wheel robot controlled by android smartphone

A omni-wheel robot with 3 wheels is controlled via bluetooth by android smartphone.
A circuit (RT-ADKmini) with PIC24F and bluetooth dongle is settled on the robot,
and it controls a motor according to the commands sent from the smartphone.

The smartphone receives a streaming movie from a Wifi camera (IP Webcam installed on android) settled on the robot.

In the target tracking mode, it tracks something based on the color obtained by image processing.

Controlling toy train (plarail) via Bluetooth with android smartphone

A toy train (plarail) is controlled via Bluetooth by android smartphone.
A circuit (RT-ADKmini) with PIC24F and Bluetooth dongle is settled on the toy train,
and it controls a motor according to the commands sent from the smartphone.
The smartphone receives a streaming movie from a Wifi camera (Ai-Ball) settled on the toy train,
and when it recognizes obstacles by image processing, it stops the train.

We use BT_DROID developed by Mr. hrdakinori for the communication via Bluetooth.

A robot arm which grabs objects recognized by image processing

French students who visited our laboratory for three months from May 2012.

An android phone recognizes the shape of object in the images sent from the Web camera via Wifi, in which we utilize OpenCV.
Then the phone controls the robot arm to grab the object.

The control commands are sent to PIC24f on RT-ADKmini via Bluetooth, using BT_DROID created by Mr. hrdakinori.

Target tracking with USB webcam and android (PandaBoard ES and Zeemote)

USB webcam with motors are controlled by android so that the target (typically, human face) is settled in the center of the display.
The image recognition is based on either FaceDetector of andorid or user-defined template matching, and performed by PandaBoard ES with android (GB: Gingerbread or ICS: Icecream Sandwich).
PandaBoard ES sends commands to motors through USB to serial converter.
This system can also be controlled by Zeemote (Bluetooth controller) manually.

Target tracking with USB webcam and android (Iconia Tab A500, ADK, and Zeemote)

USB webcam with motors are controlled by android so that the target (typically, human face) is settled in the center of the display.
The image recognition is based on either FaceDetector of andorid, and performed by Iconia Tab A500 with android.
Iconia Tab A500 sends commands to RT-ADKmini (ADK board) and then it moves motors through RS-485.
This system can also be controlled by Zeemote (Bluetooth controller) manually.