Open source tools facilitate image georeferencing over internet

Open source tools facilitate image georeferencing over internet

Sampath Kumar P
National Remote Sensing Centre
— Select —
Vivek Kumar Gautam
IRS-Programme Management Office, ISRO Satellite Ce
— Select — –


The advent of GIS software has made spatial data interpretation easy and increasingly simple to understand. But, everyone does not have access to commercial software due to their high cost and steep learning curve. Internet is coming to the rescue in such situations. The ability to easily disseminate information has made the internet a viable distribution platform. This article focuses on development of open source tools for image processing in a web environment enabling use of spatial data even by non-experts, without the need for dedicated commercial image processing software


Recent advances in the areas of satellite, optics and communication technology have provided us with high resolution digital imagery across the globe. These advances in remote sensing and Geographic Information Systems (GIS) are immensely beneficial in diverse applications for planning, resource management, environment monitoring, and rural and urban development programmes etc. Progress in the areas of open source software and development of internet as the major distribution platform has brought these technologies within the reach of common man. Therefore an attempt has been made to develop open source tools for performing image processing over web environment that can be used even by a non-expert without the need for dedicated commercial image processing software on client’s computer. Few satellite image data processing tools were identified which are necessary for their analysis such as image restoration, georeferencing, shift resampling and edge detection. These tools were developed for desktop and used over web environment. For implementing client server architecture, an open source RDBMS and MapServer were also used for facilitating concurrent user access to these tools and visualisation of data. Subsequently the image data was automatically processed at server, which was uploaded using web browser at the client end. The processed image data, along with the statistical information, was presented over web browser for visualisation and downloading the same for further analysis. These tools have been successfully evaluated and tested for its utilisation. This article presents this online image processing software architecture based on open source technologies and its implementation methodology.


Advances in the areas of satellite, optics and communication technologies are producing high resolution digital imagery from various satellites and sensors across the globe. Also, free satellite data is provided by various national and international organizations worldwide. Simultaneously, accessibility of information over the internet has increased tremendously. This has provided a good platform for the growth of Web GIS technologies and its utilisation in India, such as Bhuvan, Bhoosampada, NRDB, NSDI, etc. But there have been a few attempts to process satellite images over web. Satellite image processing over the web can use open source software which in conjunction with greater availability of satellite data can increase the volume of processed satellite data. This is particularly relevant for developing countries for accelerating the utilisation of satellite data by resource managers and planners.

With the advent of GIS software, interpretation of spatial data has become easy and increasingly simple to understand. But, everyone does not have access to commercial software due to their high cost and steep learning curve. Many organisations are interested to distribute maps and processing tools without time and location restriction to users. Internet has made its way to many government organizations as well as students, researchers etc. The ability to easily disseminate information has made the internet a viable distribution platform.

For remotely sensed satellite data, image processing is involved at every stage. Digital image processing can improve image visual quality, selectively enhance and highlight particular image features and classify, identify and extract spectral and spatial patterns representing different phenomena from images. Geometric correction/registration of remote sensing data is a basic and essential step to correlate the data with the respective spatial extent on earth for proper utilisation of the data.

Hence, an attempt has been made with an objective of developing open source tool to perform geometric correction/georeferencing of remote sensing data by anyone having a network connection (LAN/WAN/Internet) with a server.

Tools and technologies

The tools and technologies used to achieve the objectives of the study are given in Table 1.

Table 1: Tools and Technology used in this study


In this study, we use Apache Web Server coupled with a UMN map server for visualisation of georeferenced data. Server side data processing is done by the tools that were developed using Matlab. PHP was used to tie all these applications to form a coherent and functional environment. The end product is in the form of JPEG/PNG. The architecture of a typical WebGIS is given in Figure 1.

Figure 1.Typical Web GIS model

Data used

This study has been carried out for three study sites in Kanpur, Lucknow and Bhopal. Locations of these three cities in India and the corresponding false colour composites (FCC) are shown in Figure 2. National Remote Sensing Centre (NRSC), Hyderabad the sole distributor of RS data, provides data at four different levels [3].The Level 1 data is corrected only for radiometric errors. It is not corrected geometrically, which means it has not undergone resampling. Therefore Level 1 data of IRS-1D, PAN sensor is being used as test data in the present project. Details of data corresponding to the study sites are provided in Table 2. A subscene of PAN sensor consists of about 4,600 lines 4,300 columns, which is difficult to process and analyse intensively. For this reason, small test areas of 512×512 pixels have been extracted from each subscene.

Figure 2. Locations and FCCs of three study sites. Yellow boxes represent the location of test sites used in present study. FCCs were prepared using Band2, Band3, Band4 images from IRS-1D, LISS-3 sensor

Table 2: Details of IRS-1D, PAN, Level 1 corrected data used in study

The steps followed in this study are as follows:
Development of algorithm: Digital image processing algorithms were developed for performing georeferencing on a digital image.
Development of image processing functions: The algorithm for image georeferencing was converted into a Matlab programme. This programme was then converted to independent executable using Matlab’s Deployment tool. The resulting executables can be run on any computer without having Matlab installed on it.

Processing the data: The input data that comes from the web environment i.e. user was processed by the standalone programmes.
Analysis/visualisation: The input and output images were analysed and their statistical properties calculated. The result of the processing was visualised on the browser.
Testing and evaluation: Developed applications were deployed and tested using sample data. Required corrections were made in the applications.
Application deployment: After the applications were tested and ready for use, they were deployed on the web server.
The general flow of steps for the study is given in Figure 3.

Figure 3 Flowchart for preparation of web tools

An interactive, intuitive and easy to use interface was developed in HTML and JavaScript for user input. User inputs i.e. input image and processing parameters were then uploaded to their respective databases created in PostgreSQL where the server performs the requested image processing. While the processing is performed, user is redirected to a ‘wait’ page. After the image is processed, the processed image and other statistics are uploaded to their respective databases. The result was presented to the user along with the statistical properties for the input and output image. The communication between client and server was handled by PHP which is a server side scripting language. Visualisation of georeferenced image was done using OpenLayers API. The processed image can then be downloaded from the results page.

Results and discussion
Tools for performing geometric correction over web environment have been developed and successfully deployed over network. This tool allows a user with internet access to process any image without requiring high processing power on the client-side. The architecture is such that it allows simultaneous connections from multiple users.

It allows the user to input any uncorrected image and produce a geo-rectified image. The user can select an input image using the browse button, a space delimited file for GCPs, and other inputs like coordinate system of GCPs, coordinate systems of output image, resampling method, distance method etc. Figure 4 shows the web interface for performing georeferencing.

Figure 4 Web Interface for Geometric Correction

Figure 5 Example of a space delimited file for GCPs

The user input is validated and if valid, it is submitted to the server for processing. After the image is processed, it is presented to the user in the results page using the OpenLayers API which provides functionalities like zoom, pan etc. The geo-referenced image can also be downloaded using the download button. Figure 5 gives a sample space delimited file for GCPs. A sample output georeferenced image is shown in Figure 6.

Figure 6 Results of Georeferencing operation visualised in Web Browser

In the present study, a raster-based georeferencing system over web has been successfully developed using client-server architecture. This article has attempted to develop a system which can be used by a non-expert without having dedicated image processing software on the client side. Standalone applications are used for server side processing. The storage and management of input data collected from web is handled by PostgreSQL database which is a robust solution for getting good performance and quick response. The performance of the web image processing system is important when it is implemented for concurrent use of multiple clients across the network.
Security of data with an authorization mechanism is also implemented in PostgreSQL which provides additional security for the system. Pre-processing is essential before using acquired RS imagery for analysis and use in an application. The system developed in this study has the capability to generate georeferenced images according to user requirements in a web based environment. Hence, a web browser is being used as a simple client. The resultant georeferenced output images can also be visualised using GIS navigation tools on the browser itself as the web interface has been customised for such tasks. The processed images can also be downloaded to the user’s computer for further usage.

Future directions
As the penetration of web is increasing to encompass more and more devices, it is becoming the dominant platform for information distribution. Therefore a system for online image processing will be essential in the near future.
• More frequently used image processing tools can be added to the system.
• Interactive online GCP collection can replace the file uploading the GCP data for geometric correction process.
• With the increase of internet connection speed all across the country, the limit on input image size can be lifted.
• A database of freely available remote sensing images can be made so that the user can select an area of interest and get the relevant image on the server itself.
• The performance of the system can be increased by using better architecture for processing. Distributed computing and GRID computing are possible alternatives.
• The safety of the system can be strengthened by taking more security precautions. In addition to client side validation of input data, server-side validation can be deployed.
• It can be extended to mobile and other personal computing devices i.e. the web interface can be optimised for display on small screens.
• With the advent of HTML5, lightweight client side JavaScript processing of raster data can be realised.

[1] Alesheikh, A. A., Helali, H., Behroz, H. A., “Web GIS: Technologies and its applications,” Symposium on Geospatial Theory, Processing and Applications. CIG: Canadian Institute of Geomatics, July 2002.
[2] Lawrence, S. and Giles, C.L., “Searching the Web: general and scientific information access,” Proceedings of the First IEEE/Popov Workshop on Internet Technologies and Services, 1999, pp. 18-31.
[3] NRSA, 1997, IRS-1D Data users’ handbook. National Remote Sensing Agency, Hyderabad.
[4] Vatsavai, R. R., Burk, T.E., Wilson, B. T., and Shekhar, S., “A Web-based browsing and spatial analysis system for regional natural resource analysis and mapping,” Proceedings of the 8th ACM international symposium on Advances in geographic information systems, 2000, pp. 95-101.
[5] Science organization that provides information on the health of ecosystems and environment, natural hazards, natural resources, impacts of climate and land-use change, and the core science systems and provides satellite imagery, url: