16-311 Intro to Robotics s13

Homework 3

Aaron Anderson (aaanders)

Part 1

Part 2

    1. A 3x3 mask implementing the average operator is
    2. This operator is known as convolution.
    3. There has been information lost--first of all, notice that the resulting image is at least two pixels smaller in each dimension. Also, we cannot reconstruct the original image, since there are many different combinations of pixel values which could have the indicated average.
    4. We want to weight the center pixel (the actual pixel) most highly. We could either weight all 8 surrounding neighbors equally or put more weight on the pixels directly above, below, and to the sides of the center pixel and less weight on those neighbors which only touch on a corner. I will go with the latter, and propose a mask such as:
  1. Solution to the Martian problem
Back to home