Final exam
You want to see your final exam?
Two options:
- session 1: Friday 9 Feb 11:00-12:00 in BG.West.550
- session 2: Tuesday 13 Feb 09:45-10:45 in 01+.Oost.860
hw04
If you ask me in a private message on discord, I’ll send you the breakdown for your team + the PDF that we annoted/commented on. Give me your student number and your team number. You can come to the 2 sessions above to discuss the results also.
For the marking, we read your report, tried to run your code with our input somewhere in Limburg.
Our aim was to test your code with a 175mX175m area, as follows:
+--------175m----------+ (190810.0, 313260.0)
| |
| |
| |
| |
| |
| 175m
| |
| |
| |
| |
| |
+----------------------+
(190635.0, 313085.0)
and then we compared to our ground truth and marked based on what we read in the report, GitHub data submitted, etc.
code working + GitHub + README.txt [1.0]
[0.50] – followed all the rules and commits are not one dump
[0.25] – if the code for the GF + interpolation can be run without much hassles it’s fine. I guess if they explain where to change the values for the input it’s okay, but better if it’s a param to pass to the python script
[0.25] – if how to run the VF is explained and care was taken for us to potentially run the code it’s 0.25
GFTIN implementation [3.0]
[the original aim was not to thin the input but in the end I accepted to run it wiht maybe 33% only]
[1.5] – did they implement correctly the algo? we can compare to the ground truth we have here
[0.75] – description of the algo and how they implemented it, discussion about params tweaking (I expect a bit of trial-and-errors at least)
[0.75] – do they discuss quality of their results, did they try to validate their results and does it work (with their own datasets)
Laplace interpolation [2.0]
[1.0] – does Laplace work and creates a smooth terrain (you can use startinpy implementation)
[0.5] – do they deliver a complete 500mx500m grid (1000pxX1000px) with realistic values on the edges?
[0.5] – description and asssesment of results
Extraction vegetation [3.0]
This was very difficult to mark exactly (because running the code was complex often, and we thus couldn’t exactly compare), so we made those 5 categories:
- [3.0/3.0] – outstanding efforts and results to obtain the vegetation. Pit-free results are obtained.
- [2.5/3.0] – pretty good but a few hiccups (pit-free attempts)
- [2.0/3.0] – it’s an honest effort and it works but some objects (eg cars, lampposts, bus stops) are still there. and they didn’t really think of the pit-free
- [1.0/3.0] – an attempt was made and it’s not working super well
- [0.0/3.0] – they didn’t do nor delivered anything
CHM generation [1.0]
[0.50] – do we have many places wiht 0.0m (where there is no vegetation)
[0.25] – are there negative values? there shouldn’t
[0.25] – do they deliver a complete 500mx500m grid with realistic values on the edges