w w w . p c i g e o m a t i c s . c o m

Challenge

Natural Resources Canada (NRCan) is responsible for the acquisition, management and dissemination of topographic information for the Canadian landmass to all federal government departments and the general public. As part of their mandate to provide quality geospatial information to Canadians, NRCan awarded TELUS Corporation, Iunctus Geomatics and PCI Geomatics, a $2.4 million contract to process recently acquired high-resolution imagery of Canada, from SPOT 4 and 5 satellite data.

The goal to implement a national imagery database (11.5 million square kilometers) at 10m panchromatic and 20m multispectral resolution stemmed from the need to update the out-dated Landsat 7 coverage. Natural Resources Canada will freely distribute this new high-resolution imagery to various users through GeoBase, a Web portal governed by the Canadian Council on Geomatics (CCOG).

To meet the project timelines and high quality expectations, Iunctus Geomatics chose PCI Geomatics’ image-centric technology to develop an enterprise production system. Such technology will be used for the generation of accurate satellite images covering all of Canada, south of 80° North latitude. The 5-year contract from 2006 to 2011 involves both the delivery of software, as well as production services for the final orthorectified products and associated metadata.

PCI Geomatics opted for an advanced automated approach to address the challenge of generating and delivering high quality orthorectified SPOT imagery in a timely and cost effective manner. At the core of the automated production solution, a workflow was designed to ingest, orthorectify, and manage the SPOT data and its derived orthorectified products.

Through automation, volumes of imagery will be corrected without the time-consuming task of orthorectification. A turn-key enterprise approach is ideal for delivering corrected imagery to Natural Resources Canada. This approach will manage the quarter million individual files that will be produced over the course of the project and will also allow for the delivery of the corrected imagery on time and on budget.

Key Points

Providing Canadians with high resolution, orthorectified SPOT imagery of Canada.

With a national imagery database being implemented, new imagery of Canada will be freely distributed via a web portal.

PCI Geomatics image-centric technology is used to develop an enterprise production system.

A timely and cost effective automated approach for delivering orthorectified satellite imagery to Canadians.

Developing an automated workflow to produce timely and accurate orthorectified SPOT imagery over the Canadian landmass

C a s e S t u d y

National Imagery Project

National Imagery Project

w w w . p c i g e o m a t i c s . c o m

Project Objectives

Two main objectives were set to support the final goals of the project:

to orthorectify all satellite imagery at low production coststo guarantee the quality of the final products

In this context, the automated production system was designed to:

produce quality above-standard ortho-images which can be delivered to the end-user in conjunction with DEM data and the associated metadatautilize automated processes to generate the final products, product reports and statistics, and archive the data and the derived output

Solution

The objective of the automated production system is to alleviate bott lenecks of t radi t ional orthorectification. With automation at the core of the system, the orthorectification process will produce orthoimages, corresponding metadata and process reports, all within a standard approach. Such standardization guarantees quality geospatial imagery is achieved with high levels of throughput.

Furthermore, with the automated workflow, an operator’s skills would be only required for the overview of data management and quality assurance tasks, resulting in a cost effective method for enterprise data production.

An automated orthorectification production workflow The management of the geospatial data and products, process reports and metatdata files are vital in ensuring the successful delivery of 5000 SPOT scenes over the duration of the 5-year project. The production system leverages a scalable data management framework designed to assist image management and product delivery. The automated orthorectification workflow for the raw SPOT 4 and 5 is the core of the production system. To orthorectify a single SPOT scene, the system reads the raw data from the SPOT sensor format and assigns the image an estimated geographic location obtained from its

corresponding satellite ephemeris data. With the approximate georeference position of the image known, the control geospatial data is compiled.

The required Digital Elevation Model (DEM) used to correct the relief distortion in the image was assembled from elevation data tiles. Three Ground Control Points (GCP) sources: the national road network vector layer, a chip database, and orthorectified Landsat 7 imagery were assembled as reference data to be used in the image orthorectifcation process.

The next step within the automated process is to collect the GCPs between the uncorrected image and the various geospatial control data. A GCP distribution algorithm is then used to prioritize, merge, and select the GCPs from the various control sources for use in the orthorectification stage. The accepted GCPs (those that satisfy the quality threshold) are then used to compute a rigorous satellite math model used for image orthorectification. When the image orthorectification is completed the output files with associated metadata, DEM, and product reports are then packaged for delivery.

C a s e S t u d y

w w w . p c i g e o m a t i c s . c o m

The complete automated processing of scenes is tracked by the database application, providing a centralized location for processing new scenes, performing quality assurance, and product delivery. Once a final automated check of GCP distribution is completed, a quality control process is executed to validate the accuracy of the satellite model.

GeoSpatial products that meet a high standard of quality The automated production workflow used for the correction of the SPOT imagery alleviated a common time consuming task within orthorectification – Ground Control Point (GCP) collection. Collecting sufficient quality GCPs for large data volumes has traditionally required a significant amount of human interaction, making large scale ortho-imagery production an overwhelming and costly endeavour. The ability to collect good quality GCPs directly impacts the accuracy of the orthorectified results. To ensure the generation of quality ortho-images, three automated GCP collection methods were used in the production system.

The first module was an image-to-image matching technique used with the uncorrected imagery and Landsat 7 orthorectified referenced data. The second module collects GCPs between the uncorrected image and chip database (a database containing georeferenced thumbnail images based on GCP points).

The third module is a vector-to-image matching algorithm that was developed, tested and integrated into the system to support GCP collection from the road network vector data. The vector-to-image and chip database techniques were the two sources of

ground control with a higher accuracy rate, which were used to supplement the Landsat 7 control data.

To further improve the efficiency of the system and guarantee the accuracy of the output, an additional quality assurance procedure - a GCP distribution algorithm - was added to the system before the heavy computational step of image orthorectification. The GCP distribution algorithm ensured the ground control

was distributed equitably across the spatial bounds of an input image, as well as distributed across varying elevations. All accepted GCPs, evaluated against standard thresholds, are then used in the subsequent process of the automated workflow. As a last step, the automated production system also produces metadata and accuracy reports for every orthorectified image to ensure the final product has attained appropriate quality standards.

w w w . p c i g e o m a t i c s . c o m © 2007 PCI Geomatics Inc. All rights reserved.

Benefits

The PCI Geomatics automated orthorectification enterprise system supports rapid and cost-effective production of more than 5000 10m and 20m images from SPOT satellites.The automated production system generates ortho images around the clock, with little human intervention, resulting in reduced error and increased company productivity.

The operating costs for the project are largely decreased by reallocating analysts to product management and quality assurance activities. The automated processes save a significant amount of labour by removing the repetitive time-consuming operations from an analysts workload.

With the ability to automate the GCP collection process, traditional operational congestion is eliminated. Additionally, the advanced GCP algorithms (image-to-image, vector-to-image, chip database) and the rigorous math modeling for the orthorectification, ensure that accurate and quality geospatial products are produced and delivered to NRCAN.

The orthorectification automated system was built on an open and scalable architecture, which allows the system to be modified to meet new project requirements without rescheduling final product delivery dates. If required, the system can be extended by integrating new image processing algorithms and/or adding support for different data resolution.

An additional benefit of the production system lies with the data management component which supports the spatial as well as the non-spatial data organization; this sub-system maintains information about product processing status, accuracy reports, as well as the raw and the processed results.

The tracking capabilities inherent to the data management component improve internal operations and further assist in the distribution of the orthorectified imagery to the end user, allowing them to meet their internal and external demands in a timely manner.

About PCI Geomatics:Founded in 1982, PCI Geomatics is a world-leading, innovative developer of image-centric software for the geospatial industry. Specializing in remote sensing, digital photogrammetry, spatial analysis, map production and automated production systems, PCI Geomatics’ technology turns images into information. Our geospatial solutions enable our customers to find answers to real-world questions using satellite, aerial and other geospatial data.

From the desktop to the enterprise, our technologies are applied to a wide array of challenges, including natural resource monitoring, homeland security and defense, insurance risk assessment, map production, image management, site selection, feature extraction, and education. With our trusted Geomatica® brand, PCI Geomatics provides the image-centric solutions demanded by a large and expanding industry. With over 21,000 Geomatica licenses in 135 countries, PCI Geomatics is driven to develop leading-edge applications that our customers require to remain productive and competitive.

Email: info@pcigeomatics.com

C a s e S t u d y

Gatineau, Quebec Phone: (819) 770-0022 Fax: (819) 770-0098

Richmond Hill, Ontario, Phone: (905) 764-0614Fax: (905) 764-9604

National Imagery Project