dt-woods commented 3 years ago

The Challenge

You have been given three ASC raster files that represent the red, green, and blue color bands of a satellite image; however, you will note that about 80% of the pixels in each band are missing (i.e., assigned NODATA_VALUE), there are "holes" present in random locations, and a concurrent strip is missing.

Your task, should you choose to accept it, is to create a gapless version of the image.

You may assume that the following hold true:

Each ASC raster file has the same CELLSIZE and spatiotemporal extent (i.e., they all represent the same area at the same time).
Pixel values represent red, green and blue color bands, which should only have values between 0-255; no values should be outside this range.
The XLLCENTER and YLLCENTER are not correct in the ASC headers. This is on purpose to hide the image's true location.

Learning Objectives

Working in a team
Working under a tight deadline
Critical thinking
Problem solving
Code/algorithm development
Collaboration using a Git repository

Instructions

Find a partner and start coding an algorithm that provides the missing values from the given datasets (i.e., replace all the NODATA_VALUE values with numeric values). Together, upload a single script that reads the data from file and writes a single gap-filled output. Come up with an explanation of what assumptions you made, methods you used/borrowed, and how well you think your algorithm performs. Upload your script only (no data files) to the "sdc" folder in our semester repository (tagging the appropriate issue with your commit message) and post your answers to this week's discussion.

Requirements

Work with your partner(s) to brainstorm an approach and develop your solution
Add a comment to this issue with your names
You may use any methods and any tools to produce a single gapless image (keep track of the resources you use)
Push your one solution to the sdc directory of our semester repository within 48 hours of this issue going live. Tag this issue with your commit message(s). Your solution should be:
- a Python script with your solution methods (initials_intials.py) or as much of a solution as you have (e.g., twd_jpk.py); your script should read the ASCII raster files, gap-fill the NODATA_VALUES, and save your answers to a valid ASCII raster format (e.g., twd_jpk_rgb.asc)
- bonus rather than create a single gap-filled version, create a gap-filled solution of all three raster bands
- bonus convert your gap-filled raster band(s) into a grayscale/color image format (e.g., initials_initials.png)
Please do not upload ASC raster solutions to the repository; they are too big and we can just run your code on the input files to see your output.
Answer the discussion board questions before our next class discussion

Evaluation

Score	Description
0	Did not submit
1	Does not meet the requirements of the assignment (e.g., methods don't match the assignment; something submitted but not what we wanted)
2	Partial answer (e.g., methods appear to answer the question, but not complete)
3	Almost everything; maybe a typo or an error in the Python code that's easily fixed or discussion questions not completely responded to or you didn't work in a team or you submitted your answer after the deadline
4	Meets all requirements
5	Meets and exceeds requirements (see bonuses)