Au oma ic de ec ion o cle s and co ido s
wi hin he sands one pla eau
1.0
Kacpe Jancewicz
Depa men o Geomo phology
Uni e si y o W ocław
2024
Concep in oduc ion
INPUT DATA
Ai bo ne-LiDAR-based Digi al Ele a ion Model (high esolu ion g id o a leas 1 m x 1 m
esolu ion).
EXPECTED RESULT
G id –cells ep esen ing zones enclosed by a leas wo opposi e ock walls o heigh exceeding
ce ain h eshold alue.
ASSUMPTIONS
•So wa e en i onmen : A cGIS 10.8
•Analysis o ou pai s o di ec ions: N–S, NE–SW, E–W, SE–NW.
•Two c ucial a iables:
•Cle Wid h Th eshold alue (CWT) ha de ines he maximum wid h o de ec ed pi , cle
o co ido (measu ed in me es o numbe o cells)
•Dep h Th eshold alue (DT) ha de ines he minimum dep h o pi s, cle s o co ido s
(measu ed in me es)
So… how does i wo k? (p . 1)
Fi s ly, we need o ind maximum ele a ion wi hin a
ce ain neighbou hood o e e y DEM cell.
We use Focal S a is ics ool:
•Neighbo hood = IRREGULAR (ke nel ile mus be
p epa ed)
•S a is ics ype = MAXIMUM
The i egula shape o neighbou hood is de ined by
CWT and di ec ion (one o eigh alues: 0°, 45°, 90°,
135°, 180°, 225°, 270°o 315°) –see example
The calcula ion is done o each o eigh di ec ion
alues, esul ing eigh g ids: Hmax000, Hmax045, (...)
Hmax315.
Abo e: Example o ke nel ile designed o CWT
= 25 cells and di ec ion = 0°(N). Cen al cell
ma ked in g een, while i s neigbou hood is
ma ked in o ange.
How does i wo k? (p . 2)
Secondly, we use as e
calcula o and sub ac alues
o maximum ele a ion g ids
om ini ial DEM:
𝐷𝐸𝑀 − 𝐻𝑚𝑎𝑥000 = Δ𝐻000
𝐷𝐸𝑀 − 𝐻𝑚𝑎𝑥045 = Δ𝐻045
𝐷𝐸𝑀 − 𝐻𝑚𝑎𝑥090 = Δ𝐻090
(...)
𝐷𝐸𝑀 − 𝐻𝑚𝑎𝑥315 = Δ𝐻315
Example o he ΔH000 g id – he Szczeliniec
Wielki s udy si e (S ołowe M s, Poland)
How does i wo k? (p . 3)
Thi dly, we ha e o de ine he DT alue. Fo
ins ance, i we sea ch o cle s deepe han
1 me e , hen DT = -1 m.
Each ΔHxxx g id is eclassi ied as ollows:
<DT → 1
≥DT → 0
As a esul , we ge eigh g ids: 000_below
DT, 045_below DT, (...), 315_belowDT (see
igh o examples)
How does i wo k? (p . 4)
We’ e looking o land o ms consis ing o a leas wo
pa allel ock walls, so we sum g ids ep esen ing he
opposi e di ec ion alues:
000_below DT + 180_below DT = sum000_180
045_below DT + 225_below DT = sum045_225
090_below DT + 270_below DT = sum090_270
135_below DT + 315_below DT = sum135_315
Resul s should be eclassi ied:
0(no walls de ec ed) → NoDa a
1(one ock wall de ec ed) →NoDa a
2 ( wo pa allel walls de ec ed) → 1
A e eclassi ica ion we ge ou g ids: cle s000_180,
cle s045_225, cle s090_270, cle s135_315, which should
be mosaicked o new as e Cle s_ inal (mosaic ope a o =
MAXIMUM)
eclassi ica ion
WARNING!
CWT and DT alues should be de ined by use
indi idually o each s udied si e in o de o achie e
op imal esul s.
Cle s and pi s o
Szczeliniec Wielki
mesa
(CWT = 25 cells, DT = -2 m)
Tha ’s all! Choose you
s udy si e and gi e he
p ocedu e a y .
