Imagery Speaks on GISCafé
Manager of Technical Marketing at Exelis Visual Information Solutions
Local Geospatial Companies Assist Colorado Flood Responders
October 23rd, 2013 by Patick Collins
GIS and Image Analysis are Ideal for Natural Disaster Recovery and Response
The recent floods in Colorado exceeded previous assessments of the impact that such a large amount of rain would have along the front range of the Rocky Mountains. Many flood zone maps and other pre-disaster plans were rendered irrelevant as soon as rivers began jumping their banks and re-routing into neighborhoods and fields. The impact to the roadways into the mountains has been unimaginable, forcing road closures and detouring residents for many weeks, causing them to drive hours daily in order to get from their homes out of the mountains. In the aftermath of this disaster GIS analysts in and around the flooded areas, including the city and county of Boulder, the city of Longmont, , were charged with assessing the flood extent and impact that the floodwater had on their communities. Initial concerns included destroyed transportation routes, private and government owned property, the stability of soils near houses and roadways, and much more.
Geographic Information Systems (GIS) and Remote Sensing Analysis are ideal technologies for conducting disaster response and recovery. The ability to remotely capture imagery of the affected areas and extract features such as flood water extent, destroyed buildings, and elevation models allows responders to quickly assess disaster conditions and guide recovery efforts. Advanced analysis techniques, such as change detection, allow users to quantify changes on the ground and identify areas most in need. These techniques are proven to save time over manual identification and annotation of images and maps, which allows the analyst to be more efficient in their duties.
How can we help?
In response to this unprecedented disaster, Exelis Visual Information Solutions teamed up with DigitalGlobe™ Inc. of Longmont and the local Esri® office to conduct a series of workshops for local GIS analysts dealing with the flood response. Limited term licenses of ENVI® image analysis software were provided to the attendees, along with post-flood data donated by DigitalGlobe.The class itself provided overviews of remote sensing concepts and the value that geospatial data, like the imagery provided by DigitalGlobe, brings to disaster response. Examples were given outlining how ENVI is used in various disaster situations ranging from forest fires to tsunamis. The last half of the class was dedicated to hands-on training. During this time, the analysts were given individual instruction on how to conduct image analysis over flooded areas in the region. The portion of the workshop consisted of the following exercises:
Delineate Flood Extent over Longmont
This scenario utilized ENVI supervised classification to extract the extent of flood waters over downtown Longmont. This highlighted how image analysis can eliminate the need to manually digitize flood extents.
Detect Change over Longmont
This scenario took the previously classified image and ran change detection between it and a pre-event image classification to determine areas of change within the city. A major focus of this scenario was extracting the washed out South Sunset Street bridge.
Structure Inventory over Boulder
This scenario used LiDAR data and ENVI LiDAR to extract building features over an affected area. These features were then analyzed, corrected, and ported over to ArcGIS for further analysis. This scenario showcased how LiDAR can be used to quickly extract features such as buildings or trees, particularly in remote or inaccessible areas.
Terrain Analysis over Four Mile Canyon
This scenario introduced the elevation component to our analysis. Digital elevation models and imagery were combined to determine ideal routes for ingress and egress to affected areas, helicopter landing zones for evacuation, and preferred routes for moving supplies into damaged areas.
Flood Water and Affected Roads using Lines of Communication along the North Saint Vrain River
This scenario used the lines of communication workflow to identify flowing muddy water vs. rocks and sand along the North Saint Vrain river. These areas were then analyzed to see where roads were damaged along the river to allow responders to better assess damage and distribute resources more effectively.
According to the feedback received from the class participants this workshop contained information and concepts that were valuable and improved their ability to respond to the flood. Many analysts were unaware of the functionality that image analysis software such as ENVI can offer when responding to natural disasters. While the focus was on responding to the current situation, many attendees were quick to see how image analysis could assist them in other areas of their jobs. During each of the workshops, an open discussion was held that allowed responders to share their stories of struggle and success from the previous weeks. Regardless of the technical information learned in the session, this exchange was a valuable activity for everyone involved. Resources were shared, contacts were made, and relationships developed that will help improve communication among local analysts now and into the future.
Webinar and Training Series
Due to interest in the workshop from outside the Colorado region, Exelis will hold an additional webinar on October 29th, from 12-1pm mountain time that will cover the information we presented in the flood workshop. Please join us if you are interested in learning how ENVI image analysis can be used to address disaster situations, using the recent flooding as a model for disasters for which you might be preparing. We’ve also begun a series of webinars called the ENVI Rapid Learning Series, which consists of short, 30 minute webinars each week addresses relevant tips and topics to help any image analyst become more proficient in their remote sensing techniques.