Advanced

This section explains how to generate your own tables for a city.

Follow these steps:

1. Choose a city / create the initial table

Configure initiate_table.py:

• Choose the city name, table name, and WGS84/UTM CRS for the city
• Make sure the PostgreSQL credentials are correct

Run the script:

python3 initiate_table.py

2. Run the master pipeline

python3 master_pipeline.py

By default it uses the following example. However, if the city should be changed, simply all the variables can be easily altered.

Example:

CITY="Cape Town" CITY_EPSG="32734" BBOX_WEST="18.3" BBOX_SOUTH="-34.1" BBOX_EAST="18.5" BBOX_NORTH="-33.8" INNER="5" OUTER="20" MLY_SCORE="0.9" python3 master_pipeline.py

3. Test different thresholds

By default, the original filter_output.py script has different thresholds for:

Option	Short	Description	Default
--contrast-threshold	-C	Minimum contrast score	0.35
--tone-mapping-threshold	-H	Minimum tone-mapping score	0.35
--tone-mapping-floor	—	Tone-mapping floor	0.8

After testing, we only found that -C is useful. The modified script filter_multiple.py uses for -C five different values and the other two variables the default thresholds.

python3 filter_multiple.py

4. Decide which threshold to use

There is no such a thing as the perfect threshold, that is why the user should decide on their own which one to use. Depending on the size of the desired table a smaller or larger threshold can be chosen. The script compare.py provides assistance.

python3 compare.py

5. New table & delete entries

After deciding for value, the final script can be used. It creates a new table from the original city table and keeps only the accepeted ids'. The new table is then called city_cvalue, e.g., cape_town_035.

python3 delete_entries.py -C 0.35 # example value