|
|
|
Georeferencing is a critically important step in remote sensing data processing. Geo-corrected remote sensing data ensures suitability of long-term data series for change detections such as climate change and environment changes. The earth location error of AVHRR level 1b dataset at the satellite sub point region remains within 2-4 km for recent datasets but errors near the orbit’s edge are often as large as 10 km (http://www.osdpd.noaa.gov/PSB/PPP/NAVIGATION/overview.html). Large quantity AVHRR satellite datasets were produced daily by NOAA series satellites. Georeferencing of these data is time consuming and computationally intensive. To improve accuracy and efficiency of processing, an automatic processing system was required. MRSIP_AVHRR was developed to solve this problem.
|
|

|

|
|
Ungeoreferenced orbit from 1981
|
Ungeoreferenced orbit from 2006
|
|

|

|
|
Sample of georeferenced composite using popular algorithm
|
Georeferenced composite produced by MRSIP_AVHRR
|
|
|
|
The MRSIP_AVHRR system component is specific to data captured by NOAA's AVHRR sensor. It was successfully developed by Earth-To-Map GIS Inc. It automatically processes archived or near real time NOAA satellite data at 1km spatial resolution. High georeferencing accuracy and radiometric consistency are achieved for all AVHRR datasets from NOAA-6 to NOAA-18. In addition, it automatically generates daily and multi-day clear sky composite over a very large area. This includes the whole of Canada, Greenland, Iceland, most of the contiguous US, most of Alaska, parts of the UK, Russia and surrounding oceanic regions.
|
|

|
- Automatically georeferences orbits
- Automatically generates 1 day clear sky composite
- Automatically generates multi-day (user specified) clear sky composite
- Automatically updating satellite navigation parameters and radiometric calibration info
- Processes archived AVHRR datasets dating back to 1981
- Processes real time AVHRR datasets in minutes
- High speed performance and efficient file storage design
- Combines orbit processing and one day composite generation into a single operation
- Composites using all available orbits is ensured by orbit quality ranking design
- Users can subset the output coverage inside the range of 7000x6000 pixels
- Users have 3 options to select composite criteria
- Users can specify output file format: HDF or Binary
- UML design and OOD
- WIN32 GUI application implemented in VC++ under Visual Studio 2005
Our team implemented the system in VC++ under Visual Studio 2005 using the latest algorithms and technologies. The system automatically georeferences the orbits and composites based on a Convex Hull Buffer approach which was first proposed by Ji Chen and achieves a significant improvement in georeferencing accuracy. In addition, additional ground control points (GCP) in the US, Russia, Greenland, Iceland and UK improve the georeferencing accuracy. Furthermore, new orbit quality ranking design ensures that all available orbits are used by composites which increases orbit usage by 30 percent. The data coverage of previous outputs is also increased by 50% (7000 pixels x 6000 lines). This extends the coverage significantly for ocean studies and monitoring.
|
|

|
|
Data coverage of 7000x6000, white rectangle represents data coverage of 5700x4800
|
|
|
| |
Unique Advantages
|
Benefits
|
|
Output Quality
|
High Geometric Accuracy and Radiometric consistency ensured by using new and refined algorithms. To date, it provides the highest success rate and overall accuracy in georeferencing amongst all existing AVHRR automatic processing systems.
|
Ensures suitability of long-term data series for change detections such as climate change, environment changes, etc. MRSIP provides a reliable and fundamental base for generating higher level products.
|
|
Near Real Time Processing
|
Processes level 1b orbits and generates daily composite over a very large area (7000x6000) in less than 2 hours.
|
Makes it possible to provide accurate, timely, and high quality EO data for those kinds of near real time monitoring services and programs.
|
|
Archive Processing
|
Ability to process AVHRR NOAA 6 to 18 archived data dating back to 1981.
|
Makes it possible to build a 25+ year EO historic database for various services, studies and researches.
|
|
Near Real Time Processing
AND
Archive Processing
|
MRSIP can process archived data and Near Real Time Data.
|
Single system improves productivity, simplicity and reduces cost.
|
|
System Architecture Design and Performance
|
Implements a high performance and efficient file storage design based on a fully automatic approach.
|
Saves at least 60% data storage and improves performance by at least three times.
|
|
Automated Processing
|
Highly automated one step processing combines orbit processing and one day composite generation into a single operation. Automatically updates TBUS info used for orbit navigation calculations. Automatically updates calibration info used for orbit radiometric corrections.
|
Manual processing is eliminated. Performance is increased.
|
|
User Interface
|
Simple and user friendly graphical user interface implemented in VC++ under Visual Studio 2005.
|
Ease of use and minimal training
|
|
Customization
|
Users can adjust the output within the 7000x6000 data frame. Output data frame is easily customized to meet specific needs and regions using the same methodologies.
|
Provides flexibility to users to create products over their area(s) of interest.
|
|
Should you be interested in a trial version of MRSIP_AVHRR for 30 days, please contact us.
Latifovic, R., Trishchenko, P. A., Chen, J., Park, M. W., Khlopenkov, V. K., Fernandes, A. R., Pouliot, D., Ungureanu, C., Luo, Y., Wang, S., Davidson, A., and Cihlar, J. 2005 Generating historical AVHRR 1-km baseline satellite data records over Canada suitable for climate change studies. Canadian Journal of Remote Sensing vol. 31, N5, pp 324-346.
|
|
|
|
|
|