=Paper=
{{Paper
|id=Vol-2534/10_short_paper
|storemode=property
|title=An Interactive Enironment for Modeling the Processes of ERS Data Processing and Analysis
|pdfUrl=https://ceur-ws.org/Vol-2534/10_short_paper.pdf
|volume=Vol-2534
|authors=Ruslan V.Brezhnev,Yury .A.Maglinets,Ksenia V.Raevich
}}
==An Interactive Enironment for Modeling the Processes of ERS Data Processing and Analysis==
https://ceur-ws.org/Vol-2534/10_short_paper.pdf
An Interactive Environment for Modeling the Processes of ERS Data
Processing and Analysis
R.V.Brezhnev(1), Yu.A.Maglinets(1), K.V.Raevich(1)
(1)
The Institute of Space and Information Technology SFU, Krasnoyarsk
Abstract. Article presents the concept of interactive technology modeling processes and
the data analysis of the Earth remote sensing developed at Institute of space and the infor-
mation technology SFU. The technology allows to interpret models of processes in real al-
gorithmic designs data processing in the local and distributed conditions, and also to use
them for automatization permission queries of the end users in the remote monitoring mul-
ti-purpose system of ISIT SFU (MSRM).
Keywords. Remote sensing, information systems, satellite data processing, processes mod-
eling, interactive modeling environment, algorithmic sequence, information query, end us-
er.
1 Introduction
The development of the global ERS satellites’ constellation will significantly broaden the range of issues to be solved
by the RS data. This will increase the demand for geospatial data and services in government control, commercial activ-
ities and among the public. The concept of the remote monitoring software is changing, they become multi-purpose.
The multi-purpose remote monitoring systems (MPRMS) are characteristic of the end-user orientation which assumes
availability of comprehensive facilities for spatial monitoring definition and solution and reduces the role of experts in
data processing and interpretation as the mandatory problem solution participants. Such possibility is identified by the
following factors: “capacity” of the spatial data processing system which is defined by the number and variety of the
software models to process, analyze and interpret data, such number and variety being enough to solve the common
jobs’ pool in one subject area or another, and by degree of such system reasonableness marked by the expert knowledge
transfer to the system knowledge data base.
The number and variability of the processing algorithms for raster and vector spatial data as developed nowadays by
the world society and realized in the respective software packages such as ENVI, Erdas Imagine, ScanEx Image Proces-
sor, QGIS and others are enough to solve the set of common jobs. In addition, the algorithms and software are being
continuously improved in response to development of the new tools to acquire the spatial information and the computa-
tional accuracy is increased. However, most of the existing software packages are meant to be used by the image pro-
cessing experts.
Development of the web-service architectures which make it possible to solve the common jobs shall be noted
among the end-user solutions. Nevertheless, the end users often have the individual demands which means a specific
procedure of the data input, processing and presentation of results. As a rule, the particularity of the solved topic-related
issues makes it necessary to build the algorithmic sequences of the various range of complexity based on the combina-
tion of the known data processing and analysis algorithms. Thereat solving of new jobs requires not only the data pro-
cessing expert being engaged to build the processing logics which is quite natural, but also the programmer to program
the respective sequences. Such approach is inflexible as the system functioning logic is securely fixed as regards its
internal implementation and enhancement of functionalities requires, as a matter of fact, a new cycle of development.
This fact makes actual the researches aimed at development of unified technological platforms providing the data
processing expert with the interactive tools for the adaptable arrangement and debugging of the chain of operators
which present the algorithmic structures solving some job [1, 2, 3] and, thereat, allowing to exclude the source-weaving
of data processing operators which is inappropriate as regards the technology buildup process. The example of such
approach is the Sentinel Application Platform (SNAP) software package of European Satellite Agency (ESA) or the
Russian Image Media Center (IMC) of the Innovative Technologies Center.
2 Structure of the interactive process modelling environment
The major concepts of the spatial systems development are more specifically discussed in [4-8]. The systems con-
sidered in the mentioned works are of various unifications and, in most cases, require the systematic approach and pro-
gramming experience to adapt for the new ranges of the monitoring jobs.
Copyright © 2019 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 Interna-
tional (CC BY 4.0).
As an alternative this article presents the architecture of the software complex providing the interactive environment
for ERS processing and analysis modelling (Fig. 1, а). It includes the below components presented on the conceptual
level.
a b
Fig.1. a - Structure of the process modelling software package; b - algorithmic structures’ View Model.
2.1 The Guide of Software Operators
The Guide of Software Operators (Table «Operators», Fig. 1, b) keeps the list of operators available for experts dur-
ing the algorithmic design. Operator is a complied library or a software unit which can be realized in any interpreted
language (Python, Perl, PHP and others). The software units shall have the below properties to function as part of the
algorithmic structures:
Autonomy is one of the key properties of the operator functioning which characterizes its ability to execute the
given set of functions without other system operators and allows considering it as an independent element of
the control chain. Thereat the operator is considered as the programming interface for cooperation between a
user being a customizer and an objective system model represented by the collection of libraries. Thus, the au-
tonomy does not result in duplication of functions by different operators. The operator programming interface
may be realized as a command line or a graphic interface to uptake the user setups (Fig. 3).
Controllability. This property means the restricted number of input 𝐼𝐴 and output 𝑂𝐴 parameters which are par-
tially designated by an expert during algorithmic design and partially automatically determined when obtaining
the interim processing results. Input parameters dictate the scenario of an autonomous operator behavior. This
means that the operator itself shall build-up its functioning logics depending on the parameters’ contents and
their possible values.
Parameters are determined by the “qualifier-value” principle where, for instance SQL, SPARQL-structures
may be used as the values instead of common numerical or line data in order to communicate the interim oper-
ation results from the preceding i-d operator to the successive i+1 or i+n (𝑖 = 1,´ 𝑛) operator when the results is
not apriori known. Also, the files’ search, files’ content search, regular expressions and etc may be considered
in parameters.
Stability. The property which gives an indication of an operator capability to function properly with the wrong
set content or sequence of parameters. This means that the parameters should not be listed in a strict order and
the surplus or non-existent parameters coming out in the list shall not lead to the operator failure. If the re-
quired parameter has been missed, the operator shall inform the user thereof in a dialog mode.
Security. The property which gives an indication of the operator capability to resist the willful or unintentional
destructive actions of the user.
Understandability is indicative of a well-documented and structured operator description, in particular, of its
functionability, content and assignment of input parameters, combination of parameters and behaviors.
2.2 Algorithmic Data Base
Each algorithmic structure is characterized by its own specific set of parameters which are generalized in the below
expression:
𝐴 =< 𝐼𝐴 , 𝑂𝐴 , 𝑅𝐴 , 𝐹𝐴 >, (1)
where 𝐼𝐴 is a set of input parameters of algorithm A; 𝑂𝐴 is a set of input parameters; 𝐹𝐴 is a set of algorithmic func-
tions; 𝑅𝐴 is a set of relations with the other algorithms which shall be taken into account as 𝐹𝐴𝑖 ∩ 𝐹𝐴𝑗 and at the physical
level this property is indicative of the modular approach to algorithm design which, in its turn, implements the principle
of the program code re-use or the principles of the object-oriented approach.
Relational approach is the most correct and flexible as regards presentation of the system knowledge of the structure
and content of algorithmic structures, such approach allowing the operators to be bound in the logical sequences, defin-
ing the set members 𝐼𝐴 and storing the unlimited sets of such sequences. Fig. 1, b, shows the part of the data base meant
for algorithmic structures. Model includes the following tables:
Algorithms is a reference guide of generated algorithmic structures.
Tables Operators_Settings and Operators_Order are binders to establish the conformities between algorithms,
their integral operators and parameters of the operators. They identify the set of operators’ sequences, their pa-
rameters and their order in the structure.
2.3 MpRMS Operator Interface
Provides the customizer with interactive tools for the process model visualization and the process initialization and
startup (Fig. 2). Interface shall also account for the possibility to manage the Operators’ Reference Guide, algorithmic
structures and task performance schedule, namely, to add, edit and delete lines, as well as the possibility to establish
binds between information queries of users and algorithmic structures designed to settle the queries.
2.4 Data base of user information queries
Data base of user information queries is a part of the Data Base (Table Processes) meant to record the incoming data
processing queries of end users. Therewith there may be two types of queries: single and cyclic. Single queries mean the
queries of the one-time calculation of some set parameters, for instance, calculation of area. Cyclic queries provide for
the multiple calculation of the set parameters for the set objects at the certain time interval which corresponds with the
monitoring jobs. This way the model of the user information query allows presenting the query as the monitoring plan
be described by certain status and binding the job definer, group of objects and the respective algorithmic structure (Fig.
1, b).
2.5 Job manager
Job manager is a programming module meant to be activated to perform the algorithmic structures in accordance
with the performance time schedule. The activated algorithmic structure is a system process of various status as regards
the operational system. The Manager allows interpreting the system status and changing it. As the multi-job demand
means that the operators of various algorithmic structures may be performed in parallel, then the job manager shall con-
trol the system resources which are a priori limited. Therefore, on the one hand, the thread (Thread) queue (Queue) pro-
cedure is foreseen for the structure operators to be performed in strict compliance with the expert-set order. On the other
hand, the control subsystem shall control the system memory and prevent the simultaneous performance of more than 5
structures.
2.6 Planner
The system planner shall activate the job manager from time to time, namely once in a minute. The alternative solu-
tion of the automated no-planner operators’ start job is to create the services that continuously supervise the jobs’ status
[9].
3 The Experimental Testing of Technology
The status of spring sown cereal during the growing season was chosen as a monitoring job for the experimental
testing. The status is assessed based on the periodical calculations of the vegetation index and status charting to be pub-
lished in web-interface of ISIT MSRM [10, 11].
The system has an algorithmic structure to deal with this query, such structure consisting of the following operators
presented in the relevant order (Fig. 2):
Archive Retrieval Module. USGS open-source data are used where image channels are distributed in ZIP or
TAR archives. That is why the first action is to expand the archive to the specified temporary directory. The
module shall self-recognize the type of the archive. The batch mode is used whereat all the archives of one date
are successively extracted to the set directory. The preloaded data for the set date and area are the pre-
requisites for the module operation.
Files’ Moving Module is used to move the initial archives to the data storage.
Mosaic Gathering module is meant for the channel-by-channel gathering of mosaics from the individual imag-
es. The module shall determine the number of images of the set spectral channel. When there is more than one
image of one channel after the archives’ extraction, the mosaic shall be gathered, otherwise it should not. Thus,
two channels are required to calculate NDVI: RED and NIR, therefore the module in the structure shall be used
twice.
Multiplexing Module is meant to combine the individual spectral channels into one multi-channel file to opti-
mize the processing.
NDVI Calculation Module calculates the spectral NDVI index and creates a channel with NDVI values.
Regional Statistics Module shall function for calculation of statistical characteristics and their storage in one of
the vector formats. Used for calculation of the NDVI average (𝑁´) for each object. The calculation results are
stored in GeoJSON file with reference to which the vegetation status map is being generated (Fig. 2).
Format Conversion Module is meant for conversion of one bit map format into another. Used for conversion of
the Sentinel satellite channels from JPEG2000 to the conventional GeoTIFF format to be further treated under
the other processes.
Files’ Delete Module is meant for recursive cleaning of directories. Used to delete the temporary directory with
the interim findings of the algorithmic structure function.
Fig 2. Example of the algorithmic model and its findings.
The described algorithm allows mapping the status of the cultures under review as per Sentinel-2A, B data.
Process modelling provides for just several major steps: selection and addition of a listed program operator to the
workspace and setting a subset of parameters for it (Fig. 3); the operators’ sequence is set by arrows (Fig. 4).
Fig 3. Steps 1 and 2 of the process modelling dialogue
Fig 4. Steps 3 and 4 of the process modelling dialogue
4 Conclusion
The interactive environment for modelling of ERS data processing and analysis as presented herein is realized as the
software and hardware complex being part of the multi-purpose ISIT remote sensing system and was successfully test-
ed. The algorithm classifier is being created based on the algorithmic structures for this environment, such classifier
being used in the designed job setting interface [12] in the form of the operations with the set spatial objects available
for an end user. Besides that, the development also allows arrangement of computational procedures in a distributed
environment when program operators indexed in the unified guide are physically located at different servers.
The research was financially supported by the Russian Fund of Fundamental Research (Project No18-47-242002
r_mk), Government of the Krasnoyarsk Territory and the Krasnoyarsk Territory Science Fund as part of the Project
named Development of the technology to build up the intellectual information systems for the object-oriented monitor-
ing of the areas as per the RS data.
LITERATURE
[1] V.N.Lobzenyov Complete cycle of ERS data processing in IMC software package // Publications of the Interna-
tional Conference “From Image to Map: Digital Photogrammetric Technologies”. Hainan, China, 2014. P. 13-19.
[2] Michael Foumelis, Jose Manuel Delgado Blasco, Yves-Louis Desnos and other ESA SNAP – StaMPS Integrated
Processing for Sentinel-1 Persistent Scatterer Interferometry // IGARSS. 2018. P.1364–1367.
[3] R.V.Brezhnev, A.A.Perevalova Technology of the graphic development of the ERS data processing and analysis
procedures// Publications of the International Conference Regional issues of the Earth Remote Sensing. Krasno-
yarsk : Siberian Federal University, 2018. P. 40-43.
[4] A.A.Matveev, A.S. Mamaev, A.A.Proshin, E.V.Flitman. Arrangement of control over operation of the distributed
remote sensing information system of RosLesKhoz//Modern issues of the Earth Remote Sensing from space. Mos-
cow, 2009. V. 6, No 2. P. 535-541.
[5] Yu.A.Maglinets, R.V.Brezhnev Development of the automation tools to receive and process the satellite data of
the regional ERS system of SFU//Modern issues of the Earth Remote Sensing from space. Moscow, 2011. V. 8,
No 3. P. 120-128.
[6] I.V.Balashov, V.Yu.Efremov, A.A.Mazurov Junior, A.S.Mamaev, A.M.Matveev, A.A.Proshin Peculiarities of
arrangement of monitoring and control of the distributed RS systems//Modern issues of the Earth Remote Sensing
from space. Moscow, 2011. V. 8, No 3. P. 161-166.
[7] P.A.Loshkarev, O.O.Tokhiyan, A.M.Kurlykov, K.V.Koshkin, A.P.Gladkov Development of ERS UGDIS with
clouds//GEOMATIKA. Moscow, 2013. No 4. P. 22-26.
[8] Brezhnev R.V., Maltsev E.A. An Ontological Spatial Monitoring System for Agricultural Land Monitoring // Pat-
tern Recognition and Image Analysis. 2015. Vol. 25, No. 2. P.201–208.
[9] A.V.Kashnitski, I.V.Balashov, E.A.Lupyan, V.A.Tolpin, I.A.Uvarov Creation of tools for remote processing of the
satellite data in the modern information systems//Modern issues of the Earth Remote Sensing from space. Mos-
cow, 2015. V. 12. No 1. P. 156-170.
[10] Brezhnev R.V., Maglinets Yu.A., Raevich K.V., Tsibulski G.M. Modeling of Agricultural Spatial Objects with
Heterogeneous Dynamically Changing Spatial Structure // CEUR Workshop Proceedings. Samara, Russia, 2018.
Vol. 2210 P.316–322.
[11] Brezhnev R.V., Maglinets Yu.A. The dynamic model of agricultural land structure on the space images in the pre-
cision agriculture tasks // E3S Web of Conferences. 2019 Vol. 75, 01001.
[12] Brezhnev R.V., Maglinets Yu.A. Information Support Technique for Solving Agricultural Land Monitoring Tasks
Based on Earth Remote Sensing Data // Journal of Siberian Federal University. Engineering & Technologies.
Krasnoyarsk, 2017. Vol.10. NO 6. P. 819-827