Full text: Proceedings, XXth congress (Part 4)

  
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
After the process of partitioning sections along the river, make 
observation from the output meshes shown in fig. 3 and fig.4, 
  
It is found out that original assumption for numerical expression 
of the flow direction really works fairly well in those relative 
smooth areas of the river (fig.3); however, it is not so self- 
adapted when those abrupt variations come into being at the 
corner, “overlapping” phenomena will appear in the output 
meshes then (fig.4). It indicates that direction transition 
between sections is not smooth enough as is required to adapt 
the variation of the river flow all the way; some smoothening 
operations are in need to make up the limitation of the original 
assumption, now. 
We have tried to average the slope values of adjacent certain 
number of sections along the flow direction, after an attempt of 
this kind of smoothening processing, 
  
Fig.6 after slope-averaged process at the corner part 
  
All the output meshes in this paper are generated and 
displayed in software TecPlot9.0 (Tecplot, inc. 2004) 
directly from the result file of this algorithm, each mesh is 
an irregular quadrangle made up of four sampling points 
which would be introduced in detail in $3.4. 
^ 
it is found out that *overlapping" phenomena at the corner can 
be eliminated effectively as shown in fig.6, however, in those 
horizontal parts as in fig.5,as the slope value used to express the 
direction would be large in abstract values and would even be + 
co or -, there would be great discontinuity in the directional 
transitions between adjacent sections and “overlapping” 
phenomena will appear now. 
At last, we come to realize that it is not the smoothening 
methods but the measure taken to represent the direction of the 
sections are unfit for handling above “overlapping” phenomena. 
Angle is proved to be good at expressing directions of all the 
sections continuously, this could be seen from output meshes 
shown in fig.17 (a) ~ (d), in which sections along the river have 
been processed with a smoothening method based upon angle- 
value averaging; thus achieves good performance in partitioning 
sections self-adapted to variable flow direction. 
Further considerations on how to prevent “overlapping” 
phenomena taking place within the river way have been taken 
into account, also, we will not cover much on them as it is not 
difficult to achieve the goal after above transactions, now. 
3.4 Elevation interpolation of the sample points 
In fact, partitioning sections along the river is one way used to 
make terrain of the riverway become discrete in lengthways, in 
order to make it also discrete in widthways to complete the 
dispersion of the entire terrain, following solutions have been 
taken to do like this: 
|. Confirming both ends of each section in coordinate values 
on the river boundary and telling the starting from the 
ending, which also making certain the sampling direction 
in each section; 
2. Sampling out certain amounts of discrete points along 
each section; 
3 
3. Calculating clevation values of each sample points along 
every section by the use of interpolation algorithm. 
As all the DEM grids within the river way have their valid 
elevation values but those outside don’t and direction of each 
section-line together with one Secr-Point are known to us, it is 
not too difficult to scan from the Secr-Point to confirm both 
ends of each section-line, just as illustrated in fig.7. 
    
   
"s RS E n x 
Fée % 
FEFTE B 
UE I SE E 
IT TT EE 8 
ES 8 
171-13 f 
ERI EDS E | 
FTIIT 
EE BEES Ern 
*u BE ES NÉ 
TI ELIT 
ITLITI 
BEER RUE P 
BENE kt 
Fig.7 confirming both end points of the section-lines 
Successive computations in hydrodynamic models require that 
sequence of sample points in one section maintain some kind of 
parallel relationship with those of the next, as shown in the 
following two figures. , which are fig.8 and fig.9, 
  
"* S; (i21,2...n) represent the sequence of sample points in one 
section, T; (i-1,2...n) represent those in the adjacent next 
section. Incremental direction of / represents the sampling 
direction. 
454 
  
Internat. 
——— 
Fig.8 rig 
fig.9 sl 
samplin 
guarante 
sure to c 
Another 
along e: 
flow di 
(thumb 
corresp« 
from th 
samplin 
First, fo 
us analy 
and the 
Accordi 
describe
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.