The ERDAS DEMs have been analysed by variation in
the minimum and maximum template sizes and the
maximum parallax value, as shown in Tables 2 and 3.
Looking at the effect of the minimum template size, eo4
compared with eo5, and eo9 compared with eo10 reveals
very little systematic change between each set of results.
Comparing changes in maximum template size using eo1
With eo4, eo2 with eo5, and eo8 with eo10 there appears
to be no consistency. Marginally better results with the
smaller maximum template size with the smaller
minimum template size. However, the larger minimum
(and smaller maximum parallax) show the reverse. From
these results there seems to be little correlation between
the quality of result and the change in template size.
From using eo7, eo1 and eo6 (maximum parallax 3, 5
and 7) the value of 5 gives very slightly better results
than the value of 3 and both are better than those with
the value of 7. From eo8 and eo3 there is perhaps a
marginal improvement from using 5, however with eo5
and eo9 the maximum parallax value of 3 is slightly
better than 5. Changing the ERDAS parameters have
not brought about very consistent changes to height
values.
A few general observations from Tables 1 to 3 show that
there are a greater number of residuals which are
positive rather than negative for both techniques. Also
from an analysis of the number of points with small
residuals, in general, the results from the ImageStation
are slightly better than the ERDAS.
Min Template 7 7 11 9 11
(p)
Max Template 9 9 13 15 15
(p)
Max Parallax 7 3 3 3 3
(p)
Range (m) eo6 | eo7 | eo8 | eo9 | eo10
Under -0.20 7.5 75 6.9 6.5 6.8
-0.20>-0.15 3.2 3.4 2.9 3.4 3.1
-0.15>-0.10 4.0 4.2 4.4 4.5 4.2
-0.10>-0.05 51 5.6 5.5 5.6 5.6
-0.05> 0.00 7.3 7.5 7.5 Tal, 7 4
0.00<+0.05 8.9 8.8 9.2 9.5 9.4
0.05<+0.10 9.7 9.6.]. 10.2.1..10.3 |...10.5
0.10<+0.15 10.2 | 11.1 | 10.6 | 10.9 | 10.9
0.15<+0.20 9.6 9.6 | 10.2 9.7 110.1
Over +0.20 34.5 | 32.6 | 32.6 | 32.1 | 31.9
+0.05 16:24 163-4167: A720 17.1
+0.10 31.0 | 315} 324 |} 331 | 33.2
+0.15 45.2 | 46.8 | 47.4 | 48.5 | 48.3
+0.20 58.0 | 59.9 | 60.5 | 61.4! 61.3
Under -0.20 75 75 6.9 6.5 6.8
Over +0.20 34.5} 326 | 32.6- 32-1F 319
Table 3 OrthoMAX DEM Comparison Statistics
(% of points, p = pixels)
Min Template 7 9 11 7 9 4.3 Analysis of Different Coastal Zones
(p) The DEM comparisons can be studied further by
Max Template 9 11 13 15 15 analysing the individual coastal zones. Tables 4 to 9
(p) give a summary of results, in percentages, from the DEM
Max Parallax 5 5 5 5 5 generation over the beach surface only and the cliff face
(p) only. i.e. the beach points (1734 in total) and the cliff
Range (m) eol | eo2 | eo3 | eo4 | eo5 face points (647 in total) have been extracted from the
Under -0.20 7.9 7.0 6.4 7.4 6.6 DEM which was created over the whole model (9084
-0.20>-0.15 31| 31| 28| 33| 3.0| Points in total).
-0.15>-0.10 4.1 4.5 4.6 4.6 4.4
-0.05> 0.00 7.4 7.6 7.6 7.2 7.1 +0.05 21.1 | 20.6 | 18.2 | 20.8 | 18.9 | 17.8
0.00<+0.05 9.0 9.4 9.0 8.8 9.2 +0.10 41.2 | 39.9 | 35.3 | 38.8 | 33.9 | 33.1
0.05<+0.10 99. 10.1] 10.7] | i0.1] 10.6 10.15 57.8 | 55.0 | 50.2 | 54.1 | 49.1 | 45.7
0.10<+0.15 10.7 | 11.0] 105] 102] 11.3 £0.20 69.7 | 69.9 | 64.0 | 68.1 | 62.0 | 57.2
0.15<+0.20 95] 1001 103 | 10.3 9.9 Under -0.20 oS S. 11.1 9.6 | 11.4 | 13.6
Over 40.20 324| 3171| 325| 326 | 324 Over +0.20 | 21.0 | 20.8 | 24.9 | 22.3 | 26.7 | 29.2
+0.05 164.1 17.0 166 | 16.0. 183 Table 4 ImageStation DEM Comparison Statistics,
+0.10 322) 327) 328) 315| 323 Beach Only (% of beach points)
+0.15 47.0 | 48.2 | 47.9 | 46.3 | 48.0
+0.20 590.7 | 61.31 61.1 [ 60.0} 61.0 Range (m) eo1 eo2 eo3 eo4 | eo5
Under -0.20 7.9 7.0 6.4 7.4 6.6 +0.05 16.8 14.8 17.2 15.5 | 161
Over 40.20 32. 41.-31.7-1- 32.5 }--32.61-32,.4 +0.10 32.1 31.4 325; 34.01" 316
+0.15 45.0 | 465 | 46.2 | 45.5] 47.0
Table 2 OrthoMAX DEM Comparison Statistics +0.20 56.74 59.31 598 874.587
(% of points, p = pixels) Under -0.20 13.3 12.1 111 12.0 | 11,1
Over +0.20 30.0 | 28.6 | 29.1 30.7 | 29.9
Table 5 OrthoMAX DEM Comparison Statistics,
Beach Only (96 of beach points)
922
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996