Vous êtes sur la page 1sur 1

Assignment 4

EE60062 Digital Image Processing


MM61503 Digital Image Processing and Applications
(LTP 3-1-0), Autumn 2014-15

Instructor(s): Prof. Pranab K. Dutta (EE), Dr. Debdoot Sheet (EE)
Type: Computer Based
Programming languages: Matlab (preferred), Octave (optional), Python (optional and good for you)
Start Date: 11 Aug. 2014, Monday
Deadline: 11 Aug. 2014, Monday
Submission type: Solve in Class, NR 323

Problem 1:

Dr. Katsushi from Petapixel Inc. is now facing another problem. In order to sense sparsely emitted
photons for imaging in very low illumination conditions, the integration time of the sensor is very high.
As such, subtle movements of the sensor or camera introduces motion blur in the images, and the
clarity is compromized. They have found out a way of compensating for this motion blur for which they
would need to estimate the direction in which the blur has occurred. Since they currently lack in-house
expertise to estimate the direction in which blur has occurred, they have approached you for your
assistance. Can you write down a utility for estimate the direction of blur at each pixel location?
Please estimate the motion direction between 0 - 180 at each pixel location and return a matrix of
the size of the original blurred image.

Problem 2:

We have shot a sequence of photographs in classroom of the same scene, where objects at different
distance are in-focus in each image. Can you write a utility to combine this sequence of images and
obtain an all-in-focus image?

Hint: Estimate local contrast / gradient / entropy or similar kinds of measures of sharpness of focus at
each pixel to search for in-focus regions in different images.

Vous aimerez peut-être aussi