Description
1.1 Submission package
Within the homework assignment package, there should be a submission-package.zip, which contains the directory structure and empty les for you to get started. Please edit the contents within the le for code, and then create a zip archive with the le name submission-package.zip, and submit it. Do not use other archive formats such as rar or tar.gz.
All assignments should be submitted electronically. Hand written reports are not ac-cepted. You can, however, include scanned pages in your report. For example, if you are not comfortable with writing equations, you can include a scanned copy.
1.2 Code
All assignments should be in Python 3. Codes that fail to run on Python 3 will receive 20% deduction on the nal score. In other words, do not use Python 2.7.
For this assignment, you should not need to create additional les. Fill in the skeleton les in the submission package. Do not change the name of these scripts. Your code should run as speci ed by the assignment specs by simply doing
python solution.py
in the command line. Do not expect additional command line arguments as input.
It is strongly encouraged to follow PEP8. It makes your code much more readable, and less room for mistakes. There are many open source tools available to automatically do this for you.
1.3 Delayed submission
In case you think you will not meet the deadline due to network speed or any other reasons, you can send an email with the SHA-256 hash of your .zip archive rst, and then submit your assignment through email later on. This will not be considered as a delay.
Delayed submissions are subject to 20% degradation per day. For example, an assignment submitted 1 minute after the deadline will receive 80% of the entire mark, even if it was perfect. Likewise, an assignment that was submitted one day and 1 minute after the deadline will receive 60%.
1.4 Use of open source code
Any library under any type of open source license is allowed for use, given full attribution. This attribution should include the name of the original author, the source from which the code was obtained, and indicate terms of the license. Note that using copyrighted material without an appropriate license is not permitted. Short snippets of code on public websites such as StackOver ow may be used without an explicit license, but proper attribution should be given even in such case. This means that if you embed a snippet into your own code, you should properly cite it through the comments, and also embed the full citation in a LICENSES le. However, if you include a full, unmodi ed source, which already contains the license within the source le, this is unnecessary. Please note that without proper attribution, it will be considered plagiarism.
In addition, as the assignments are intended for you to learn, (1) if the external code implements the core objective of the task, no points will be given; (2) code from other CSC205 students will count as plagiarism.
-
Aggregating multiple images to remove artifacts (2 marks)
Implement solution.py, speci cally the TODO parts in merge images. You only have to use np.nanmean and np.nanmedian to do each task. You can also test to run the mean version by doing python solution.py –method mean.
Here are some example outputs that opencv should give you. Simply press a key to keep on running the stitching process.
(a) Example matches (b) Example stitch
(c) Example matches (d) Example stitch
Figure 1: Reference results.
3
