TCMA 1-Meter Land Cover Classification metadata

Metadata: TCMA 1-Meter Land Cover Classification

TCMA 1-Meter Land Cover Classification
This page last updated: 06/30/2016Metadata created using Minnesota Geographic Metadata Guidelines

Go to Section:
1. Overview
2. Data Quality
3. Data Organization
4. Coordinate System
5. Attributes
6. Distribution - Get Data
7. Metadata Reference

Section 1 Overview
Originator Remote Sensing and Geospatial Analysis Laboratory, University of Minnesota - Version 1
Title TCMA 1-Meter Land Cover Classification
Abstract A high-resolution (1-meter) land cover classification raster dataset was completed for three different geographic areas in Minnesota: Duluth, Rochester, and the seven-county Twin Cities Metropolitan area. This classification was created using high-resolution multispectral National Agriculture Imagery Program (NAIP) leaf-on imagery (2015), spring leaf-off imagery (2011- 2014), Multispectral derived indices, LiDAR data, LiDAR derived products, and other thematic ancillary data including the updated National Wetlands Inventory, LiDAR building footprints, airport, OpenStreetMap roads and railroads centerlines. These data sets were integrated using an Object-Based Image Analysis (OBIA) approach to classify 12 land cover classes: Deciduous Tree Canopy, Coniferous Tree Canopy, Buildings, Bare Soil, other Paved surface, Extraction, Row Crop, Grass/Shrub, Lakes, Rivers, Emergent Wetland, Forest and Shrub Wetland.

We mapped the 12 classes by using an OBIA approach through the creation of customized rule sets for each area. We used the Cognition Network Language (CNL) within the software eCognition Developer to develop the customized rule sets. The eCognition Server was used to execute a batch and parallel processing which greatly reduced the amount of time to produce the classification. The classification results were evaluated for each area using independent stratified randomly generated points. Accuracy assessment estimators included overall accuracies, producers accuracy, users accuracy, and kappa coefficient. The combination of spectral data and LiDAR through an OBIA method helped to improve the overall accuracy results providing more aesthetically pleasing maps of land cover classes with highly accurate results.
Purpose Land cover information offers important inputs to local, regional, and state land use planning and natural resource monitoring.
Time Period of Content Date 2015
Currentness Reference A multitemporal composite of aerial imagery from the summer of 2015, fall of 2009-2011, lidar data of 2011 and 2012 were classified.
Progress Complete
Maintenance and Update Frequency As needed
Spatial Extent of Data Twin Cities Metro Area
Bounding Coordinates -94.03
Place Keywords Twin Cities Metropolitan area (TCMA)
Theme Keywords NAIP, Lidar,object based image classification, land cover
Theme Keyword Thesaurus NAIP: National Agriculture Imagery Program, 4 band, 1-meter resolution aerial imagery collected periodically. Lidar: Light detection and ranging. Lidar data provide information on height and elevation.
Access Constraints The Remote Sensing and Geospatial and Analysis Laboratory, University of Minnesota, has attempted to produce accurate maps, statistics and information of land cover and impervious surface area. However, it makes no representation or warranties, either expressed or implied, for the data accuracy, currency, suitability or reliability for any particular purpose. Although every effort has been made to ensure the accuracy of information, errors and conditions originating from the source data and processing may be present in the data supplied. Users are reminded that all geospatial maps and data are subject to errors in positional and thematic accuracy. The user accepts the data -- as is -- and assumes all risks associated with its use. The University of Minnesota assumes no responsibility for actual or consequential damage incurred as a result of any user's reliance on the data. The data are the intellectual property of the University of Minnesota.
Use Constraints This data may be used for educational and non-commercial purposes, provided proper attribution is given. Secondary distribution of the data is permitted, but not supported by the University of Minnesota. By accepting the data, the user agrees not to transmit this data or provide access to it or any part of it to another party unless the user includes with the data a copy of this disclaimer.
Contact Person Information Joe Knight, Associate Professor
Remote Sensing and Geospatial Analysis Lab, University of Minnesota
1530 Cleveland Avenue North
St. Paul, MN  55108
Phone: 612-625-5354
Fax: 612-625-5212
Browse Graphic None available
Associated Data Sets
Section 2 Data Quality
Attribute Accuracy The TCMA dataset has 88% overall level 1 classification accuracy and 86% for level 2 Deciduous and Coniferous Tree Canopy.
Logical Consistency
Completeness Data provides complete coverage of the Twin Cities Metropolitan Area (TCMA).
Horizontal Positional Accuracy NAIP Summer 2015: FSA Digital Orthophoto Specs. The quarter-quad files are 1 meter ground sample distance (GSD) ortho imagery rectified to a horizontal accuracy of within 6 meters of absolute ground control ("true ground") rather than reference imagery.

Spring 2010 aerial imagery TCMA: An overall accuracy for this product was targeted for 2.4 meters, CE95. Surdex collected Airborne GPS data at the time of image acquisition and supplemented with several control points surveyed by ground crews. Surdex used these ground control points and additionally some QC points dropped out of the Triangulation in order to validate the accuracy of the final digital orthophoto quarter quads (DOQQ) product. Of all the points collected, only 20 were easily readable in the final product and were used to calculate the stated residual. 0.85, value in meters, CE95.
Vertical Positional Accuracy The vertical accuracy of the lidar data meets or exceeds 12.5 cm RMSE.
Lineage Mosaic of the National Agriculture Imagery Program (NAIP) imagery: Leaf on NAIP imagery was collected in the Summer of 2015 at 1-meter resolution. Further information about NAIP 2015 is available at

Leaf off NAIP imagery was collected in the Spring of 2010 at 0.3-meter resolution. Further information about this Spring imagery is available at
Both imagery sets include Near-infrared, Red, Green, and Blue bands. The Normalized Difference Vegetation Index (NDVI) was derived for the Leaf on and Leaf off imagery.

Lidar Data collected for Minnesota between 2006-2012 and made publicly available by the Minnesota Department of Natural Resources (DNR). For more information, the 2011 lidar metadata are available at:
LAStools was used to create 1-meter bare earth Digital Elevation Models (DEM), Digital Surface Models (DSM), and LiDAR Intensity. Additional lidar-derivative layers, such as slope and normalized Digital Surface Model (nDSM) were created as well.

Vector layers: roads, railroads centerlines, and airports were obtained from Open Street Map and the Minnesota Department of Transportation (MnDOT). The OpenStreetMap data is available under the Open Database License see:
The National Wetland Inventory (NWI) was used for water and wetland classes. For more information about the updated NWI data, see:

A set of independent stratified randomly generated points were used as reference data for accuracy assessment. These independent randomly points were checked using photo-interpretation techniques and high-resolution aerial photos available through the MnGeo Geospatial Image Service (, 3-band 1-meter, 4 bands meter natural color and color-infrared USDA National Agricultural Program Imagery (NAIP) and Spring Aerial Imagery 4-band leaf-off (2009-2015). For classes, such as extraction and agriculture, additional ancillary datasets were used.

Image Classification: We mapped 12 land cover classes including Deciduous Tree Canopy, Coniferous Tree Canopy, Grass/Shrub, Agriculture, Extraction, Buildings, Roads/Paved Surfaces, Bare Soil, Emergent Wetland, Forested/Shrub Wetland, Lakes (including ponds and lakes), and Rivers.

We utilized an Object Based Image Analysis (OBIA) approach by creating rule sets for the selected areas within the State of Minnesota. We used the Cognition Network Language (CNL) within the software package Definiens eCognition ® Developer version 9.2.0 was used to develop the rule sets. A previous high resolution land cover ruleset was adapted to create the land cover classification maps. We started by classifying the most easily separable classes and worked toward the more complex classes that were more difficult to separate. New rules were developed to classify agriculture, extraction, and to separate tree canopy type into deciduous or coniferous. The ruleset functions as a decision tree. For example, the nDSM layer was used to identify all pixels greater than 3 meters as “_Tall” (temporary class). Subsequent classes were classified from only the remaining unclassified pixels effectively masking out all pixels/objects already assigned to a class. This means that the order of the rules for each class is critical to avoid errors. The classification structure used in our ruleset follows the following order:

1. Lakes, Rivers, Emergent Wetland, Forested/Shrub Wetland – imported from the updated NWI shapefile .Given the availability of the recently available new high resolution and accurate NWI 2010-2015 data for the Northeast and Southern part of the state; we masked all the water bodies including wetland types, lakes, rivers and ponds. We used the Chessboard segmentation algorithm in eCognition for masking out these features.

2. Buildings: The LiDAR building footprints was used as the initial vector layer to classify buildings. Additional rules were created using spectral and contextual information to classify buildings that were not identified by the LiDAR building footprints layer.

3. Tree Canopy: Coniferous vs Deciduous
We classified preliminary objects (unclassified versus tall) into temporary classes. The nDSM layer was used to separate unclassified objects from tall objects, using a 3 meters threshold for separation. We executed the multi-threshold segmentation algorithm with the following parameters: a threshold value of equal or greater than 3 for tall object and anything less than 3 for unclassified objects. Previous parameters were determined after several trial-and-error experiments and a detailed visual assessment for separating tall classes. A series of rules then separated the _Tall class into the Tree Canopy class or the Building class using image segmentation to create image objects which were then classified by optical image characteristics (NDVI, brightness, etc.) and LiDAR point cloud properties (number of returns, intensity, etc.). Coniferous and deciduous were classified from the tree canopy class using optical properties, NDVI, and the difference in NDVI between leaf-on and leaf-off imagery.

4. Extraction: a mining, gravel pits, and quarries location point vector layer was used as the starting base information to identify potential extraction locations. Further refinement of this class was completed using customized rules along with spectral and contextual information.

5. Impervious (roads/other paved): OpenStreetMaps roads vector layer was used as the initial identification of roads which were then verified using optical properties of the imagery. Other paved surfaces were classified by object features distance to roads and building classes as well as object size, shape, and optical characteristics.

6. Agriculture: we used a customized polygon vector layer to identify initial potential classes for agriculture areas. This vector layer contained row crops, hay and pastures classes that were derived from Landsat 8 imagery using the Random Forest (RF) algorithm within eCognition. Further enhancement of this class was finalized using customized rules that made use of spectral and contextual information (association, size, shape, and pattern).

7. Bare Soil: A polygon vector layer from OpenStreetMap’s was used to identify potential areas of bare soil based on land use context which was then verified by optical characteristics from the NAIP imagery.

8. Grass/Shrub: Grass/shrub was the most widespread and spectrally diverse class in the landscape, so a process of elimination was used for this final class. After all other land cover classes have been identified, the remaining unclassified objects were classified as grass/shrub.

Accuracy Assessment was evaluated by comparing the classification map to an independent stratified random reference points for each area. There were 1,097 points in the Twin Cities metro area, 296 points in the city of Duluth, and 155 points in the city of Rochester. The classes from the National Wetland Inventory (Lake, River, Emergent Wetland, and Forested/Shrub wetland) were not assessed since an independent accuracy assessment was performed for these classes (details here:

Section 3 Spatial Data Organization (not used in this metadata)
Section 4 Coordinate System
Horizontal Coordinate Scheme Universal Transverse Mercator
UTM Zone Number 15
Horizontal Datum NAD83
Horizontal Units meters
Cell Width 1
Cell Height 1
Section 5 Attributes
Overview See table below for the Level 1 and Level 2 Land Cover class values and descriptions.

Descriptions of the lakes, streams and wetlands classes can be found at:
Detailed Citation
Table Detail:
Classification Scheme Land Cover
Field Name Valid Values Definition Definition Source
Grass/Shrub enumerated
1 Grass/Shrub Golf courses, Parks, Natural grass and Herbaceous vegetation
Urban/Developed enumerated
2 Bare Soil Baseball fields, golf course sand traps
3 Buildings Commerical, residential, and all structures over 3 m tall
4 Roads/Paved Surfaces All roads, parking lots, sidewalks, and paved surfaces
Lakes/Ponds enumerated
5 Lakes/Ponds Lacustrine L1, L2, Palustrine PAB, PUB, PUS
Tree Canopy enumerated
6 Deciduous Tree Canopy Oak, Red Maple, Black Maple, Paper Birch, Black Ash, Aspen, Silver Maple
7 Coniferous Tree Canopy White Pine, Red Pine, Balsam Fir, Jack Pine, White & Black Spruce, White Cedar
Agriculture enumerated
8 Agriculture Annual and perennial crops
Wetland enumerated
9 Emergent Wetland Palustrine PEM
10 Forested/Shrub Wetland Palustrine PFO and PSS
11 River Riverine
Extraction enumerated
12 Extraction Gravel pits, quarries, and mines

Section 6 Distribution
Publication Date 06/30/2016
Contact Person Information Joe Knight, Associate Professor
Remote Sensing and Geospatial Analysis Lab, University of Minnesota
1530 Cleveland Avenue North
St. Paul, MN  55108
Phone: 612-625-5354
Fax: 612-625-5212
Distributor's Data Set Identifier TCMA 1-Meter Land Cover Classification
Distribution Liability The data are provided as is and without any warranty to their performance, merchantability, or fitness for any particular purpose. The University of Minnesota does not represent or warrant that the data or documentation are error-free, complete, or current. The user should consult the data documentation to determine its limitations and accuracy. The user is responsible for any consequences resulting from use of the data or reliance on it. The entire risk of the results of the use of the data are assumed by the user.
Ordering Instructions Contact Joe Knight at the University of Minnesota
Online Linkage None available
Section 7 Metadata Reference
Metadata Date 06/30/2016
Contact Person Information Joe Knight, Associate Professor
Remote Sensing and Geospatial Analysis Lab, University of Minnesota
1530 Cleveland Avenue North
St. Paul, MN  55108
Phone: 612-625-5354
Fax: 612-625-5212
Metadata Standard Name Minnesota Geographic Metadata Guidelines
Metadata Standard Version 1.2
Metadata Standard Online Linkage

This page last updated: 06/30/2016
Go back to top

Main menu