Skip Maine state header navigation
Tags
Maine, USA, 2001, 2007
This is a representation of the changed impervious surfaces meeting specific criteria within the state of Maine, between the time of the early-date imagery (circa 2001 - 2003) and the late-date imagery (2007), depending on the location within the state. It was created as part of a project designed to quantify change in development between T1, or early date, and T2, or late date.
This project resulted in the creation of raster data layers that document changes in the spatial pattern of commercial and residential development between two time periods in organized towns in Maine. This analysis was based primarily off leaf-off imagery from 2001-04 (T1) and leaf-on imagery collected in 2007 (T2) through the NAIP program (National Agriculture Imagery Program). Sanborn produced a yes/no “baseline” development raster from the 2001-04 imagery (T1), a yes/no “current” development raster from the 2007 imagery (T2), and a yes/no “new development between T1 and T2” raster. Sanborn classified the imagery for the 2001-04 baseline raster, the 2007 current raster, and the 2004-07 change in development raster and carried out a statistically valid accuracy assessment for the 2001-04 raster and the change analysis.
State of Maine, Sanborn Map Company
There are no access and use limitations for this item.
Extent
There is no extent for this item.
State of Maine, Sanborn Map Company
See access and use constraints information.
Internal feature number.
ESRI
This project resulted in the creation of raster data layers that document changes in the spatial pattern of commercial and residential development between two time periods in organized towns in Maine. This analysis was based primarily off leaf-off imagery from 2001-04 (T1) and leaf-on imagery collected in 2007 (T2) through the NAIP program (National Agriculture Imagery Program). Sanborn produced a yes/no “baseline” development raster from the 2001-04 imagery (T1), a yes/no “current” development raster from the 2007 imagery (T2), and a yes/no “new development between T1 and T2” raster. Sanborn classified the imagery for the 2001-04 baseline raster, the 2007 current raster, and the 2004-07 change in development raster and carried out a statistically valid accuracy assessment for the 2001-04 raster and the change analysis.
This is a representation of the changed impervious surfaces meeting specific criteria within the state of Maine, between the time of the early-date imagery (circa 2001 - 2003) and the late-date imagery (2007), depending on the location within the state. It was created as part of a project designed to quantify change in development between T1, or early date, and T2, or late date.
State of Maine, Sanborn Map Company
T1 Impervious automated classification
1.) All imagery not in UTM Zone 19N NAD 1983. Any imagery that was not in that projection system was reprojected to that system.
2.) All T1 Imagery was resampled to 1 meter.
3.) Imagery was divided up into 475 production areas. The production areas were created based on imagery type. There were 5 different imagery types for T1. The imagery types were 2003 2ft, 2003 1ft, 2004 2ft, 2003 1ft, 2001 6inch.
4.) The imagery was mosaicked in to their respective production areas.
5.) Training models were created for each set of different imagery in Feature Analyst (Overwatch Geospatial). Once a model passed Quality Control checks, it was run on all imagery of the same vintage.
6.) Each production area’s Feature Analyst output was reviewed and a new model was created as necessary to further improve the output.
7.) Once a production area’s outputs were accepted, they were combined to make one file.
8.) Minimum mapping unit for structures captured was 900 sq. ft.
9.) Road types captured were paved roads, unpaved roads, driveways and logging roads.
10.) Other features captured landfills, road in gravel pits and cemetery roads.
T2 Impervious classification
1.) All Imagery that was not in UTM Zone 19N NAD 1983 was reprojected to that system.
2.) All T2 Imagery was resampled to 1 meter.
3.) Imagery was mosaicked by type into the production areas created for the T1 imagery. There were 3 imagery types for T2, these were NAIP, 6inch 2006 York imagery and 6inch 2006 Columbine imagery. For editing purposes the 6inch imagery was used when available, otherwise NAIP was used.
4.) The T2 was edited using manual heads-up methods.
5.) Minimum mapping unit for structures captured was 900 sq. ft.
6.) Road types captured were paved roads, unpaved roads, driveways and logging roads.
7.) Other features captured landfills, road in gravel pits and cemetery roads.
Editing Impervious datasets
1.) An ArcMap document which was created by the state of Maine was used for all editing. The document contained a vector editing tool created by Maine to help edit missed buildings.
2.) A raster version of the automated T1 impervious was contained in the document. A blank raster to add T2 change such as roads, driveway or features that could not be captured by the building tool was added in. Two vector files were added from a geodatabase, one for T1 edits and another for T2 edits.
3.) The raster files were edited using the ArcScan Raster Paint tool. The vector files were edited using the tool that Maine created for editing buildings.
4.) Once editing was complete data were reviewed to make sure that all T1 and T2 impervious was captured.
Creation of Final Dataset
1.) Once all impervious tiles were accepted, all T1 vector files were merged into a single vector file. The same was done with the T2 Change vector data.
2.) A Field was then calculated in the T1 one and T2 change vector data with a value of 1.
3.) Each file was converted to raster using Spatial Analyst.
4.) The Edited T1 Raster files were mosaicked into one file using the mosaic tool in Erdas Imagine. The same was done for T2 raster data.
5.) The Erdas mosaic tool was used to combine T1 raster data with the rasterized T1 vector data. This same process was used to combine the T2 raster data and the rasterized T2 vector data.
6.) Once that was complete the output T1 combined dataset was combined with the T2 datasets using the Erdas mosaic tool. This created the 2007 Impervious dataset.
7.) A model was written in ERDAS to subtract the T1 impervious from the 2007 T2 impervious dataset to create the change raster dataset.
8.) Each dataset was subset by township using the shapefile towns_dissolved.shp as a template. This was done using the Erdas mosaic tool.
See access and use constraints information.