plan.md

A10_changeDet

plan:

• for each of the bgt objects that indicates a portal
  • to do: which bgt objecten?
  • BGT: bgt.overbruggingsdeel
  • Kerngis: vKunstwerk_FeaturesToJSON >> lower("OBJECTTEKST" ) = 'viaduct'
    • When does A10 go under and when does it go over
  • test area: 117813.6,487266.1,117942.1,487370.7
  • test area wkt: wkt = 'POLYGON((117813.6 487266.1,117813.6 487370.7,117942.1 487370.7,117942.1 487266.1,117813.6 487266.1))'
• filter cars out for visualization and cleaning of data before clustering
  • to do: 2018 < data cleaning
  • to do: 2019 < visualization/data cleaning
• read the points from the ept-folder
  • where is viewer hosted?
  • available to whom?
  • to do: create potree tiles
  • to do: create cesium tiles

method 1

- cluster on Z
	 - Assumption: the two biggest clusters are the top and bottom of the portal	
 - CLUSTERS FOUND
 - onderste 10 % (ofzo) vinden, ruis eruit halen en vergelijken
 - 

method 2

- validate method for calculating doorrijhoogte
- calculeate doorrijhoogte for x portals
- compare over different years

vragen:

• Hoe definieren we de doorrijhoogte?

• willen we de doorijhoogte of alleen het verschil?

• compare two clusters statistically

  • specify

• write std, mean, etc to ???

ideen

• bovenste vlak wordt niet gevonden met ransac.
• Snij-lijn onderste twee trusvlakken (soort van diagonale vlakken)
  • kan bepalen hoe schijn hij is in type 1 (trus horizontaal of niet)