CSCI 235 - Intelligent Robotics
Spring 2017
Project #6: Image Matching
Image Matching
To determine whether a robot recognizes a given image, we can calculate
the distance between a currently viewed image and a reference image.
A lower distance implies a higher similarity between the images. Two examples
of distance metrics are:
- Threshold each image into a binary image. Perform an exclusive-or
operation on the images. The number of bits in the resulting image is the
distance.
- Calculate the sum-of-squared-differences between each corresponding
pixel in each image. This sum is the distance.
Library
The library files are in modeselection.zip. This includes all of the library files from the previous assignments.
Sample Program
The sample program is in proj6.zip.
Assignment
Implement three programs that use image matching:
- The first program should employ at least two images. Program the robot
to drive between two locations. When it reaches a given location, it should
stop. When the ENTER button is pressed, it should turn around and drive to the
other location. This continues indefinitely.
- For the second and third programs, devise an interesting application that
makes use of image matching. You may use any technique we have employed so
far this semester.
Questions
- For each program, devise a metric for its performance on its task. How
well did each program perform? Feel free to experiment with different
parameter settings to optimize performance. Be sure to discuss the most useful
parameter settings in your report and presentation.
- How did the thresholded matcher perform in comparison with the SSD matcher?
For each program, you are welcome to use either one, but you should experiment
with both in at least one program to compare their efficacy.
- What combinations of techniques did you find most useful in
completing each of your tasks? Why?