You are on page 1of 14

Zeroth Review

Real Time Motion


Sensing And
Robotic Control
Mr. Malaya Kumar Nath
Assistant Professor
Dept. of ECE

Nikhil Kumar K (EC12B1018)


Vignesh D (EC12B1029)

Ms. K. Bala Nivetha


Assistant Professor
Dept. of ECE

Objective

To design an Real Time Motion Sensing Algorithm


The data should be processed to control a Embedded
System based Robot

Materials and Methods

The development comprises two phases.

Phase - I - Algorithm Design

Mathwork's Matlab Implementation

OpenCV Implementation (Python - Raspberry Pi)

Materials and Methods

Phase II - Hardware Implementation

Raspberry PI with Camera PI

Atmel Atmega8 Microcontroller

Servo Motors

Mechanical Assembly

Visual Tracking
Wide Range of Applications such as
Motion Analysis, Active Recognition,
Surveillance and Human Computer
Interaction
The Main Challenge of Visual Tracking is
that the algorithm should be robust and
should be able to process the frames in
real time with Maximum Accuracy.

Algorithm Development - I

The first proposed Algorithm will be High Compression


Real Time Algorithm having very high processing speed
but less Accuracy.
The frames will be dissimilated into the Surrounding
and Object Matrices , followed by compression and
Classifier Updation
The same steps are performed for further frames,
comparing classifiers to detect the Object.

Algorithm Development - I

Algorithm Development II

The Second Proposed algorithm will be Machine


Learning based Spatio - Temporal Context.
By using the Spatial Weight Function, we obtain the
Focus of Attention , using which the Frame is Processed
to get the Frame of Attention.
Frequency Domain Analysis of focus is performed to
Sense the motion to obtain the Filter which is convoluted
with the Next frame to get object location.

Algorithm Development II

Basic Flow of the Proposed Machine Learning Based Algorithm

Hardware Implementation

The algorithm developed will be implemented in


Raspberry PI B Model, with Camera Pi Attached to it.
The GPIO Pins will them be Connected to Atmega8
Based Embedded System. Raspberry Pi acts as
Processing unit and Atmega8 acts as Controller Unit for
servo Mechanisms.

The Basic Block

Raspberry PI
Atmel Atmega8
Servo Motors

References

Real-Time Compressive Tracking by Kaihua Zhang,


LeiZhang and Ming-Hsuan Yang (IEEE)
Fast Visual Tracking by Kaihua Zhang, Lei Zhang,
Qingshan Liu, David Zhang and Ming-Hsuan Yang
(IEEE)

Thank you

You might also like