Tuesday, October 23, 2007

top as CC or bottom as CC?

Prerequisite

  • For any edges of a CC
    • the inner edge (of shorter radius) is indexed as (1), outer (2)

Top:

  • Pros
    • more direct and intuitive for CC analysis
      • no Inf in the data range, and they are really "connected".
    • edges are valid depth data
      • when calculating depth of the groove bottom, the depth of the top is directly available.
  • Cons
    • the edges are not the boundaries of the same groove, making the forthcoming groove locating more complex.
      • to locate groove(k), we have to get CC(k+1).edge(1) and CC(k).edge(2)
      • with another orientation (e.g. 12 o'clock), it might change into CC(k+1).edge(2) and CC(k).edge(1)

Bottom

  • Pros
    • more direct and intuitive for later polar coord extraction for grooves.
      • always CC(k).edge(1), CC(k).edge(2)
  • Cons
    • the CC analysis is less intuitive, need to combine the sidewall and the bottom as a pseudo-CC.
    • no valid depth value on the edge
      • when addressing the depth of the bottom, the reference depth of the top should be derived from surrounding area. This require a branched navigation to find adjacent valid area,
        • At 10 o'clock, go from 5-10 o'clock for edge(2)
        • At 2 o'clock, go from 8-2 o'clock for edge(2)

No comments: