um 
djustment as well as all routines for managing the data, i.e. the 
ui 
(^f 
© 
t0 O 
^ 
2. BASIC CONCEPT OF THE SYSTEM 
The triangulation system consists of two independent parts, the measuring 
system and the bundle adjustment which share a common data base (Fig. 1). 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
7 : Sub-bl i i 
| PSR 1 Measuring A. hrs Dek ing Vis Bundle 
| interactions AS ci 
| System jus 
| ment 
\ Baarda!s test 
b criteria 
7 7 
  
  
  
  
  
  
  
  
  
  
Fig. 1. The triangulation system implemented in the DSR 1. 
| The communication between the two programs happens indirectly via the data 
base and directly with the help of two small routines. The block adjustment 
can be started up whenever wanted by choosing the data for it interactively 
(sub-blocking). From the bundle block adjustment a feedback is received in 
1 the form of weight coefficients of residuals (qyy values). They can be util- 
1 ized in an interactive routine which computes the test criteria (t-test 
values), thus making it possible e.g. to evaluate the necessity of remeasure- 
ments. 
Apart from some simple checkings Baarda's data snooping is the primary means 
Of checking gross errors. The data snooping is adapted In the following least 
squares computations: inner orientation, relative orientation, triplet for- 
mation, all similarity transformations, polynomial strip adjustment and 
bundle block adjustment. Interaction is always used when making final decisions 
concerning the detected gross errors. 
  
i 3. DATA BASE 
  
  
| The data base which is totally updated after each measurement consists of the 
| following items (capacity in the PDP-11/23 given in parentheses): 
1 - points to be adjusted (4 500) 
b - photos (500) 
To TOStrips (50) 
| - blocks (3) 
b - cameras (200) 
io - model set up (500 points) 
I - points to be intersected after the | 
| final bundle block adjustment (32 767; 2000 per model) 
I - point groups (30) 
ored on point records thus makin nd] loc!