International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B3, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Now, all 108 RMK DX images and 71 known high precision
ground points are used for quality validation. The efficiency of
the new method is also relatively compared to the commercial ?
software LPS/ERDAS Imagine 2010. The latter provides PEL
automatic tie point measurement based on least-squared image | LPS
matching and a default density of 5 x 5 standard tie points per
image. Table 5 shows the number of N-fold tie points
determined by the new method of SIFT-supported dense point
cloud matching and by the LPS/ERDAS Imagine 2010.
Apparently, the new method can provide denser cloud of tie ?
points than the default density of LPS/ERDAS Imagine 2010. | mi
LP:
N-fold point 2 3 4 5 6
New method | 60198 | 15839 | 6712 | 3415 | 1211 es
LPS 2010 238 259-1: 230.| 231 | 238
N-fold point 7 8 9 | 10111:] 12 | 13 ] 14 ]
New method | 645 | 444 | 294 | 164 | 98 | 80 | 63 | 24 | m
LPS 2010 157 | 139 | 123]. 66137122 .| 12] .8 LP:
(c
r N-fold point | 15 | 16 | 17 | Total number of tie points iin
p xd "nC UU New method | 19| 2| 2 89210
Figure 11. Location of key points without QF (top) and with QF LPS 2010 2 1 i 1764 1
(bottom) for the image 10123 (left) and 60185 (right) :
Table 5. Number of N-fold tie points —
Number | Matching | Number | Skip Lr
of Time of Rate Number'or Number of Skip Rate
Matched (Unit: Skipped | (Unit: Points Skipped (Unit: %) T
Points Seconds Points % Points a
2716 5915 19 0.70 New method 235002 2548 1.08
37505 11621 512 137 LPS 2010 9736 174 1.79
5767 5920 60 1.04 6. (um) | Test Value TO Ald
: X, Redundancy cu
77426 11623 155 New method 1.5 1.05 193965 D
LPS 2010 2.1 1.00 14371 det:
10026 5933 0.94 ima
129161 11626 1.78 Table 6. Results of free network adjustments done by the new was
method and LPS2010 proj
and
Jag Joes TOS These tie points are also first checked by means of free network feat
187295 11632 2.06 adjustment, and the results are shown in Table 6, where the test adv
value is the ratio of the a priori accuracy divided by posteriori thos
accuracy. It shows that the new method provides the denser tie all
Table 4. Computational efficiency of AFTP and QF points with the photo coordinate accuracy 6, = +1 5pm = are:
Average matching speed (paite seconda Skip rate (tai +0.21pixel. The skip rate is 1.08%. =
2 34
M Then the bundle block adjustment with control data is gut
v performed, where 6 full control points and 65 independent ed n
2% A check points shown in Figure 6 are used. The statistic figures of pe
6 both bundle block adjustments using the tie points measured by p
ài (e the new method proposed in this paper and by the commercial thre
software LPS 2010 are listed in Table 7. The root mean square AS
23 es value of ground coordinate differences on all 65 check points
v * on shows that the points determined by the new method have the Tes
e: ame tal ea horizontal accuracy +3.4cm and the vertical accuracy +1 1.9cm. ope
Figure 12. Average matching speed and skip rate for the cases I, effi
II and III denoted by blue, yellowgreen and orange colour; solid Moreover, they show that the new method is available to aerial whe
and dashed lines denote average matching speed and skip rate, triangulation with high precision and good efficiency. gen
respectively Especially, the new method don’t need any information on per
image overlap. In other word, the new method don’t need the robi
3.2 Bundle Block Adjustment well-known input data of block parameters and strip parameters ope
Which are often adopted by general commercial AT softwares. dete
72
