XXXIX-B4, 2012 
Human Face Detection 
f Shanghai Jiaotong 
e) 
orithm to gather colors 
oftware, (8), pp. 466- 
Space Conversions". 
pdf (16 April. 2012) 
lection Method Based 
n Images. Journal of 
38(5), pp. 563-567 (in 
Map Vectorization 
China University of 
81 (in Chinese) 
ang, 2005. Interactive 
Computer Application, 
1e Application of the 
n of Maps. Journal of 
Chinese) 
quan, 1999, A Colour 
n Maps with Its 
puter-aided Design 
(in Chinese) 
2001. Constrained K- 
d Knowledge. In: 
ional Conference on 
pp. 577-584 
Method of Interactive 
Micro Systems, 19(5), 
Hangen, 2004. A 
ion of Contour Line. 
40(3), pp. 56-57+62 
inyu, 2003. Research 
Color Segmentation 
ial of Computer-aided 
29-33 (in Chinese) 
yber Group Co., LTD 
iments. 
HIGH PERFORMANCE PHOTOGRAMMETRIC 
PROCESSING ON COMPUTER CLUSTERS 
V. N. Adrov, M. A. Drakin, A. Yu. Sechin 
JCSC Racurs, Moscow, Russia - adrov@racurs.ru, mike@racurs.ru, sechin@racurs.ru 
Commission IV/3: Mapping from High Resolution Data 
KEY WORDS: Satellites, DEM, Digital, Model, Mosaic, Networks, Orthoimage, Orthorectification, Photogrammetry 
ABSTRACT: 
Most cpu consuming tasks in photogrammetric processing can be done in parallel. The algorithms take independent bits as input and 
produce independent bits as output. The independence of bits comes from the nature of such algorithms since images, stereopairs or 
small image blocks parts can be processed independently. Many photogrammetric algorithms are fully automatic and do not require 
human interference. Photogrammetric workstations can perform tie points measurements, DTM calculations, orthophoto 
construction, mosaicing and many other service operations in parallel using distributed calculations. Distributed calculations save 
time reducing several days calculations to several hours calculations. Modern trends in computer technology show the increase of 
cpu cores in workstations, speed increase in local networks, and as a result dropping the price of the supercomputers or computer 
clusters that can contain hundreds or even thousands of computing nodes. Common distributed processing in DPW is usually 
targeted for interactive work with a limited number of cpu cores and is not optimized for centralized administration. The bottleneck 
of common distributed computing in photogrammetry can be in the limited lan throughput and storage performance, since the 
processing of huge amounts of large raster images is needed. 
1. INTRODUCTION 
Many operations in processing remote sensing data can be split 
into numerous tasks that can be easily done in parallel. 
   
Figure 1. Building orthomosaic from a set of images 
Figure 2. Conventional approach: many workstations and one 
The original images are similar in nature and can be processed server 
independently to build ortho mosaics. This is due to the fact that 
cutlines calculation, orthorectification, image format 
conversion, pansharpening, many service opertations performed are used. 
on one image or on few images and are independent on other 
images. Parallel processing of data can reduce the computer 
This approach is not efficient when modern computer clusters 
2. DISTRIBUTED PROCESSING ON 
time needed to get the final result considerably. The current COMUTER CLUSTERS 
trends in computer hardware are the increasing number of cores Special adjustments of software algorithms are needed for 
(identical computer units) on the off-the-shelf CPUs, larger and computer clusters. Such algorithms predict data workflow and 
faster computer networks and more and more affordable split the computing tasks into several independent processes. 
supercomputers as clusters of thousands of computer nodes. The approach is different from the conventional one: the main 
Conventional digital photogrammetric workstation (DPWs) use ideology is: one workstation and multiple servers 
distributed processing that is based on interactive work and uses 
the ideology — many workstations and one server. 
105