s.manthira moorthi, indranil misra, r.ramakrishnan and p.k.srivastava {smmoorthi, indranil, rama,...

17
S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing Area, Space Applications Centre, ISRO Ahmedabad-380015 India Presented By: S.Manthira Moorthi Presented By: S.Manthira Moorthi SAC,ISRO,AHMEDABAD SAC,ISRO,AHMEDABAD osgeo conference, Hyderabad, january,2011 S.M.Moorthi

Upload: elvin-telling

Post on 01-Apr-2015

233 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava

{smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in

Signal and Image Processing Area, Space Applications Centre, ISRO

Ahmedabad-380015India

Presented By: S.Manthira MoorthiPresented By: S.Manthira Moorthi

SAC,ISRO,AHMEDABADSAC,ISRO,AHMEDABADosgeo conference, Hyderabad, january,2011S.M.Moorthi

Page 2: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

contentscontents

Introduction/objectivesTool set for Land and Crop StudiesFree and Open source Softwares (FOSS) as a situated software

and reuseStandard versus non standard requirementsGeospatial Processing requirementsFOSS ILWIS capabilitiesFOSS OSSIM capabilitiesCombining Geospatial ToolsCode Reuse and new developmentsRapid prototyping versus long term tool buildingConclusion

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 3: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Introduction/ObjectivesTask is to quickly build a geospatial tool box with rapid prototyping to satisfy an image analyst’s need to quantify per say land use and crop studies from the available open source projects falling in this category. There are many issues to be addressed in such attempts from handling data sizes, remote sensing data formats, data types, building process chains to taking advantage of batch processing capabilities and to put software processes in an integrated framework. This study brought many insights into existing open source geospatial tools, about their strengths and weakness.

S.M.Moorthiosgeo conference, Hyderabad, january,2011

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 4: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Tool set for Land and Crop StudiesTool set for Land and Crop StudiesDesktop Geospatial toolboxDesktop Geospatial toolbox A desktop geospatial processing toolbox needs to have an elaborate framework with workflows comprising data preparation, analysis and presentation tasks taking raster and vector data as inputs.

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

TasksData formats

GIS operationsGeneral Image Processing tasksGeoreferencing

Image Registration

Classification

Statistics aggregation

Page 5: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Popular FOSS

S.M.Moorthiosgeo conference, Hyderabad, january,2011

Open Source Software Image Map (OSSIM) is a high performance engine for remote sensing, image processing, geographical information systems and photogrammetry.

SAGA's first objective is to give (geo-)scientists an effective but easy learnable platform for the implementation of geoscientific methods.

QGIS•Multi Platform•Many Core Features•Plugin Architecture•GRASS integration•OTB integration

ILWIS is the Integrated Land and Water Information System.  As the title suggests, it is oriented to needs of the environmental science user community, so it is superlative in its user interface, analysis capabilities, and  image presentation features.

Page 6: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Foss as situated software Foss as situated software In contrast to standard software development, situated software usually isn’t constructed by a team of traditional software developers. Instead, it’s created by users who desire specific functionality that mainstream software doesn’t provide. Typically, it’s personalized, localized software that has evolved organically and has been created by the community that uses it. FOSS geospatial softwares can be compared to situated software, though there can be a good amount of overlap of functionalities with Commercially Available Off the Shelf (COTS) tools.

S.M.Moorthiosgeo conference, Hyderabad, january,2011

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 7: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Contd.Contd.situated software’s success lies in its ability to accommodate change and remain flexible and adaptive, as opposed to traditional software, which isn’t very forgiving and which operates in a well-defined boundary, is costly, takes significant effort to develop and maintain, and can be difficult to modify. Flexibility comes at the expense of quality attributes such as performance, security, and availability. As an example, most reuse efforts and techniques focus on reuse at design time.

S.M.Moorthiosgeo conference, Hyderabad, january,2011

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 8: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Geospatial processing requirementsGeospatial processing requirements

Task ILWIS OSSIM QGIS SAGA Open EV

General Image Processing tasks

Good Excellent okay Excellent Okay

GIS operations Good Very limited Excellent Good Good

Data formats GDAL library GDAL library GDAL library GDAL library GDAL library

Data Types Limited types Many Many Many Many

Large raster handling

Good Good not good Not good Excellent

Georeferencing Good Very advanced Very Limited Very Limited Very Limited

Image Registration Through geo referencing module

Very advanced NIL NIL NIL

Classification Maximum Likelihood

NIL NIL Maximum likelihood

ISODATA

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 9: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Standard versus non-standard requirementsStandard versus non-standard requirements

Task Standard Non standard

Format handlers General RasterGeotiff HDFNITF

Variants of Landsat Ground Support Operations Working Group (LGSOWG)

Committee on Earth Observing Systems (CEOS) format

Georeferencing GUI based and using reference images, maps and external GCPs

Reading header GCPs etc, sensor model parameters etc.

Supervised Classification : Training sample creation

Tools to create training samples by allowing to draw polygons

To accept a shape file as a input training sample set directly

Image registration

GUI based and using reference images, maps and external GCPs

Automatic robust image registration using georefencing information etc.

Image calculus: Aggregation

Pixel grid based Overlaying a vector layer for doing aggregation under this vector layer and attribute based calculations

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 10: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

FOSS: ILWISFOSS: ILWIS ILWIS provides a wide range of tools to perform operations on spatial data and uses a number of table window operations such as distance calculations, aggregations, etc. in the analysis of attribute data. Retrieval, reclassification and measurement operations mostly from the starting point in the data analysis process, and all these operations can be performed on vector or raster map, often in combination with attribute data, efficiently in ILWIS. The overlay operations form the core of any GIS project which is used to combine a number of maps together and to generate new information. ILWIS has a powerful tool to combine maps, called Map Calculation that can combine many maps at the same time using arithmetic, relational or conditional operators.

ILWIS is almost a complete tool for image processing and GIS operations but only for a data type of 8 bit unsigned char images only

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 11: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

FOSS OSSIM CapabilitiesFOSS OSSIM Capabilities Designed as a series of high

performance software libraries, it is written in C++ employing the latest techniques in object-oriented software design.

It has a powerful automatic satellite image registration plugin as an extension to OSSIM library.

Though vector and shape file support is present, its GIS capabilities are limited.

Other salient features include rigorous sensor modeling, universal sensor models, wide range of map projections and datums supported, non-destructive, parameter based image chains, native file access, precision terrain correction and ortho-rectification, advanced mosaicing, elevation support etc. osgeo conference, Hyderabad,

january,2011S.M.Moorthi

Its API documentation is exemplary. Many of the applications are available as command line tool and not integrated into the GUI tool known as imagelinker

Open Source Software Image Map (OSSIM) is a high performance engine for remote sensing, image processing, geographical information systems and photogrammetry. It has been actively developed since 1996.

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 12: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Combining geospatial tools Combining geospatial tools FOSS ILWIS is a good choice for all GIS tasks with

raster and vector data layers, visualization and presentation but

OSSIM library scores better on image processing tasks especially on remote sensing images with its rich plugin tools on image registration and mosaicing.

ILWIS does not have the capability to classify images other than 8 bit images.

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 13: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

New developments without FOSS: New developments without FOSS: A Supervised ClassifierA Supervised Classifier

osgeo conference, Hyderabad, january,2011

ILWIS’ supervised classification module was not satisfying the custom requirement in this study because there’s no functionality to accept the training sample data in the form of ESRI shape file before image classification. ILWIS does not have the capability to classify images other than 8 bit images.

S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 14: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Rapid prototyping versus long Rapid prototyping versus long term tool buildingterm tool building

In rapid prototyping, the development simply wants to amalgamate functionalities in a non cohesive manner. This means accepting tools instead of just functionalities and accommodates them in tools management architecture. Rapid prototyping approach can also be the approach taken by not so experienced software developer or by an effort not well resourced and there is a pressing need for prototyping of potential new functionality.

osgeo conference, Hyderabad, january,2011S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 15: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Conclusion

osgeo conference, Hyderabad, january,2011

If the developmental effort is goal oriented and requirements are sacred, then over and above the aggregated functionalities from FOSS you need to build additional functionalities using the relevant tool APIs or even fresh developments to satisfy the complete list of functionalities. It will be an ideal situation to expect any open source software to satisfy the custom requirements. An experienced software development team should be ready to develop functionalities those are not catered to by any combination of FOSS.

S.M.Moorthi

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 16: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

References1. Sriram Balasubramaniam, Grace A. Lewis, Soumya Simanta, and Dennis B.

Smith (2008) : Situated Software: Concepts, Motivation, Technology, and the Future, IEEE Software, November/December 2008

2. Steiniger S and Bocher E (2008) : In press An Overview on Current Free and Open Source Desktop GIS Developments, Int. J. of Geographical Information Science , doi:10.1080/1365881080802634956

3. http://www.gnu.org 4. http://www.ossim.org/ 5. http://shapelib.maptools.org/6. http://www.remotesensing.org/proj7. http://wiki.osgeo.org/wiki/Open_Source_GIS_History#Software_projects 8. Michel R. V. Chaudron Clemens Szyperski, Ralf Reussner (Eds.), 2008:

Component-Based Software Engineering, proceedings of 11th International Symposium, CBSE 2008 Karlsruhe, Germany, October 14-17, 2008.

9. http://www.itc.nl/Pub/research/Research_output/ILWIS__Remote_Sensing_and_GIS_software.html

S.M.Moorthiosgeo conference, Hyderabad, january,2011

Discovery of software systems for composition and their customizability in free and open source geospatial software environments – A Case Study

Page 17: S.Manthira Moorthi, Indranil Misra, R.Ramakrishnan and P.K.Srivastava {smmoorthi, indranil, rama, pradeep}@sac.isro.gov.in Signal and Image Processing

Thank you

S.M.Moorthiosgeo conference, Hyderabad, january,2011