The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
389 
controls the point position in the normal direction to the 
enclosing feature. 
(3) The transformation parameters for the individual strips as 
well as the adjusted strip coordinates of the observed 
features are determined through a LSA procedure. During 
such an adjustment procedure, the individual strips are 
independently rotated and shifted until they fit together as 
well as possible. The datum for the strip adjustment 
procedure can be established by using one of the strips as a 
reference strip. In other words, overlapping strips will be 
rotated and shifted until they are compatible with each 
other and fit with the reference strip as well as possible. 
(4) The estimated transformation parameters in the previous 
step are applied to the respective strips leading to 
improved compatibility among these strips. 
4. EXPERIMENTAL RESULTS 
well as the adjusted strip coordinates. The estimated 
transformation parameters are listed in Table 2. A closer look at 
this table reveals the following: 
1. Significant differences from the expected zero shifts and 
rotations indicate the presence of biases in the system 
bore-sighting parameters. 
2. The most significant deviation is observed in the X- 
direction (refer to the X T values in Table 2). 
3. The least deviation from the optimum values is observed in 
the Z-direction (refer to the Z T values in Table 2). 
Strip 
Lines 
Measured 
X T 
(m) 
Y T 
(m) 
Za 
(m) 
(Û 
O 
<P 
0 
K 
<°> 
1 
20 
0.39 
0.07 
-0.02 
-0.017 
-0.008 
0.003 
2* 
26 
0 
0 
0 
0 
0 
0 
3 
17 
0.66 
-0.18 
0.11 
-0.037 
-0.012 
-0.009 
‘Reference strip 
The experimental results section aims at testing the validity of 
the presented procedure using a real dataset as well as exploring 
the impact of such an adjustment procedure. It will start by 
describing the dataset used followed by the strip adjustment 
results. 
4.1 Dataset Description 
The dataset used in the experiments covers and urban area and 
consists of three strips as shown in Figure 5. The specifications 
of this dataset are shown in Table 1. 
Table 2. Estimated transformation parameters using conjugate 
linear features in overlapping strips 
Figure 6. Location of the linear features used in the strip 
adjustment 
Figure 5. Dataset used in the experiments 
Sensor Model 
Optech 2050 
Flying Height 
-1000 m 
Ground Point Spacing 
-0.75 m 
1 survey day 
3 strips 
Horizontal accuracy 
50 cm 
Vertical accuracy 
15 cm 
Table 1. Dataset specifications 
4.2 Strip Adjustment Results 
The proposed semi-automated procedure for the extraction and 
matching of corresponding linear features has been applied 
leading to the identification of twenty-eight lines in the three 
strips. These features are identified in the strips they appear 
(Figure 6). The points representing these features, after the 
variance expansion, are then used in a LSA procedure to 
estimate the transformation parameters for the involved strips as 
The improvement in the strips’ compatibility after the strip 
adjustment procedure can be observed in the profile shown in 
Figure 7. The surface shown in this profile is tilted with the 
aspect almost parallel to the X-axis, which is the direction 
where the most significant discrepancy takes place (refer to the 
X T values in Table 2). 
Figure 7. Profile along the X-axis crossing a tilted surface 
before (a) and after (b) the strip adjustment