PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
March 2015
175
different levels of geospatial accuracy for (1) highest accuracy work, (2) standard
mapping and GIS work, and (3) visualization and less accurate work.
V
ertical
A
ccuracy
S
tandards
for
E
levation
D
ata
Three new terms are introduced: (1) Non-vegetated Vertical Accuracy (NVA)
based on RMSE statistics in non-vegetated terrain; (2) Vegetated Vertical
Accuracy (VVA) based on 95
th
percentile statistics in vegetated terrain; and (3)
low confidence areas. The naming convention for each vertical accuracy class is
directly associated with the RMSE expected from the product in non-vegetated
terrain. Table 3 provides the vertical accuracy classes naming convention for
any digital elevation data. Horizontal accuracy requirements for elevation data
are specified and reported independent of the vertical accuracy requirements.
The Non-vegetated Vertical Accuracy at the 95% confidence level in non-
vegetated terrain (NVA) is approximated by multiplying the accuracy value
of the Vertical Accuracy Class (or RMSE
z)
by 1.9600. This calculation
includes survey checkpoints located in traditional open terrain (bare soil,
sand, rocks, and short grass) and urban terrain (asphalt and concrete
Table 2. Common Horizontal Accuracy Classes for Digital Orthoimagery
Horizontal Accuracy Class RMSE
x
and
RMSE
y
(cm)
RMSE
r
(cm)
Orthoimage Mosaic Seamline Maximum
Mismatch (cm)
Horizontal Accuracy at the 95% Confidence
Level (cm)
0.63
0.9
1.3
1.5
1.25
1.8
2.5
3.1
2.5
3.5
5.0
6.1
5.0
7.1
10.0
12.2
7.5
10.6
15.0
18.4
10.0
14.1
20.0
24.5
12.5
17.7
25.0
30.6
15.0
21.2
30.0
36.7
17.5
24.7
35.0
42.8
20.0
28.3
40.0
49.0
22.5
31.8
45.0
55.1
25.0
35.4
50.0
61.2
27.5
38.9
55.0
67.3
30.0
42.4
60.0
73.4
45.0
63.6
90.0
110.1
60.0
84.9
120.0
146.9
75.0
106.1
150.0
183.6
100.0
141.4
200.0
244.8
150.0
212.1
300.0
367.2
200.0
282.8
400.0
489.5
250.0
353.6
500.0
611.9
300.0
424.3
600.0
734.3
500.0
707.1
1000.0
1223.9
1000.0
1414.2
2000.0
2447.7
surfaces). The NVA, based on an RMSE
z
multiplier, should be used only in
non-vegetated terrain where elevation errors typically follow a normal error
distribution. RMSE
z
-based statistics should not be used to estimate vertical
accuracy in vegetated terrain or where elevation errors often do not follow
a normal distribution.
The Vegetated Vertical Accuracy at the 95% confidence level in vegetated
terrain (VVA) is computed as the 95
th
percentile of the absolute value of
vertical errors in all vegetated land cover categories combined, including
tall weeds and crops, brush lands, and fully forested areas. For all vertical
accuracy classes, the VVA standard is 3.0 times the accuracy value of the
Vertical Accuracy Class.
Both the RMSE
z
and 95
th
percentile methodologies specified above are
currently widely accepted in standard practice and have been proven to
work well for typical elevation data sets derived from current technologies.
However, both methodologies have limitations, particularly when the
number of check points is small. As more robust statistical methods are
developed and accepted, they will be added as new Annexes to supplement
and/or supersede these existing methodologies.
If the VVA standard cannot
be met, low confidence area
polygons shall be developed
and
explained
in
the
metadata. For elevation data
derived from imagery, the
low confidence areas would
include vegetated areas where
Table 3. Vertical Accuracy Standards for Digital Elevation Data
Vertical
Accuracy
Class
Absolute Accuracy
Relative Accuracy (where applicable)
RMSE
z
Non-
Vegetated
(cm)
NVA at 95%
Confidence
Level
(cm)
VVA at 95
th
Percentile
(cm)
Within- Swath
Hard Surface
Repeatability
(Max Diff)
(cm)
Swath-to-Swath
Non-Vegetated
Terrain
(RMSD
z
)
(cm)
Swath-to-Swath
Non-Vegetated
Terrain
(Max Diff)
(cm)
X
-cm
≤
X
≤1.96*
X
≤3.00*
X
≤0.60*
X
≤0.80*
X
≤1.60*
X
