GST 101 Introduction to Geospatial Technology��Unit 2 – Understanding Spatial Data �Module 2.1 Translating Reality Into a Digital World ��� � �

Author: Ann Johnson

Title: Associate Director

Email: ann@baremt.com

​

Based upon work supported by the National Science Foundation under Grants DUE 1304591, DUE 164409, DUE 1700496, DUE 1937177, Due 1938717 DUE 1937237, 2030206 and 2015927. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

From Real World to a Digital World

Geospatial software applications require real world objects be input into a geospatial “Data Model” format, using coding that is not generally accessible to the user to:

• Visualize geographic features spatially

as a map on a computer monitor or web page

• Analyze relationships between

features based on its location and

its attributes in a database

​

Bolstad, Fig. 2-2, pg. 28, 2019

Vector Data Model

Common Data Models For GIS

• GIS software traditionally used two Data Models for features (entities or objects in the real world):
• Vector Data Model – layers of points, lines or polygons
• Raster Data Model – a grid made up of cells or “pixels”
• Other Data Models are used and include:
• Triangular Irregular Networks – TINs
• Point Clouds – LIDAR (Light Detection and Ranging) LAS (3D) data
• Object Classification – method to create “objects” from grided digital data
• 3D Voxels – volumetric and
• 4D adding time and other attributes

​

A Closer Look at Data Models From Unit 1

Vector Data Model

• Layers of points, lines, or polygons
• Points – Modis infrared sensor data
• Lines – Streets and highways,
• Polygons – Burn areas, states, lakes

Raster Data Model

• Pixels with values for continuous features
• Vegetation, Elevation, Slope, Temperature, etc.

Multidimensional 3D Data Models

• 3D voxels

British Geological Survey

https://www.bgs.ac.uk/geology-projects/geology-3d/

National Interagency Fire Center: https://maps.nwcg.gov/sa/#/%3F/%3F/39.5379/-119.3257/9

NASA Landsat: https://landsat.gsfc.nasa.gov/landsat-9

Vector Data Models

• Features are assigned a “shape” as:
• points, lines or polygons

​

• Attributes (characteristics) for each

feature are stored as a record in a database

• As “words” such as name (text) or values (Integers) such as area

​

Points, Lines and Polygons

• A Point – is made up of one “node”
• A Line is made up of at least a beginning node, a line (including direction) and an end node

​

• A Polygon is made up of nodes and lines enclosing an area

Real-World Objects in A Vector Data Model

Points

Lines

Polygons

Five Data Layers as Points, Lines and Polygons

Overlaying Point, Line and Polygon Layers

Note: at this scale a city is a point feature

Point, line and polygons data can be “overlayed” using a defined location to accurately position the features on a map

​

You will learn how to do this using coordinate systems, map projections and datums in future Units

https://www.usgs.gov/media/images/map-earthquakes-yellowstone-area-2017

TINs As A Special Type of Vector Data Models

• Triangulated Irregular Network
• Uses points (nodes), lines (edges) and areas (faces) that create triangles to represent volumes or surfaces (often elevations)
• Closer spaced triangles suggest changes in data such as elevation

Triangulated Irregular Network made up of nodes, edges and faces

Example from ArcGIS Help: https://desktop.arcgis.com/en/arcmap/10.3/manage-data/tin/fundamentals-of-tin-surfaces.htm

Geospatial Topology

• Topology is the relationship between point, line, and polygon features based on rules for how they share geometry (adjacency, connectivity)
• polygons such as parcels must share edges, street centerlines and census blocks share geometry, adjacent soil polygons share edges, . . .
• no gaps should exist between polygons, there should be no overlapping features, polygons should close, and so on . . .
• Supports topological relationship queries (analysis) and navigation (connectivity)
• Rarely are maps topologically clean when digitized or imported
• Validation methods provide error detection and cleaning of topologic errors

Topological Errors

Bolstad, 2014

Topology rules allow vector data to be reprojected without errors – map projections and their uses will be covered in later Units

Source: Courtesy of Peter H. Dana

Vectors Summary

• Vectors can represent features as points, lines, and polygons including accurate location and relationships (topology)
• Vectors are far more efficient than Rasters in use of computer memory (information on Rasters coming next…)
• Vectors are not good at continuous coverage that fill areas such as temperatures and elevations

Examples of Raster Data Layers

Elevation data

Soils from USA Soils Hydrologic Group

Imagery – Natural Color Composite, Landsat 8 Satellite data

Examples using Living Atlas data and ArcGIS Pro

Raster Data Model

Landsat 8 - Pixels – 30 x 30 m

One grid cell or pixel and has one attribute (number/text).

​

Its location is defined by its position in the grid.

​

All cells have the same resolution or cell size in ground units (such as meters or feet)

Rows

Columns

Decision Rule – Any Cell Touched or Only in Cells When Feature Is In the Center?

Bolstad, Fig 2-36

Mixed Pixel Decisions

B = Bush

T = Tree

H = House

L = Lawn

D = Driveway

P = Pavement

? = Mixed Pixel

T ? H ?

NAIP Image – US Department of Agriculture

New Hampshire GRANT GIS Clearinghouse

Pixels can only have one “attribute” value. Determining what value is assigned to the mixed pixel can result in some data loss

In remote sensing, a mixed pixel is a pixel whose value represents the average of several different features (grass, trees, buildings) on the ground

Raster Summary

• Rasters are poor at representing points, lines and areas, but good at surfaces
• Raster grids are a natural for scanned or remotely sensed data
• Raster grids suffer from the mixed pixel problem
• Raster data layers generally require much more memory than Vector data formats
• Raster grids often include redundant or missing data and include a “null value”
• Rasters can be easy to understand, easy to read and write, and easy to draw on the screen

Comparison of Features in Vector and Raster Data Models

Bolstad, Fig. 2-19

Data Models Can Be Converted from One Data Model to Another

• Conversion of Vector to Raster and Raster back to Vector data models are possible
• Changing vector to raster is easier than raster to vector
• Both conversions should be done carefully, and accuracy and quality verified
• Why convert from one format to another?
• Some analysis functions work “better” in one data model than another
• Data also are often exchanged or transferred between different GIS packages and computer systems that may only use one data format

Hand drawn Example of Vector to Raster and Raster to Vector

First Raster Uses Center of Pixel, then higher resolution grid and back to Vector

Note loss of some details in the process

Vector and Raster Data Layers Can Be Used Together

Vectors:

States, Counties,

Sub watersheds,

Rivers, Roads,

Rasters:

Landcover

Virginia Department of Forestry – InFOREST Website

Source: Ontario County NY and USGS, public domain

New and Emerging Data Models

• Point Clouds
• More government agencies and industry are

collecting and using LIDAR data and data structures

• Can be used directly (generally as a LAS file format to accommodate interchange of 3-Dimensional data) or converted into formats that can be used by different GIS software applications
• Object Data Models
• This model encapsulates different feature types into one object that can include topological information each feature and relationships between features in a logical model

Homboldt State University

https://youtu.be/8yCy-LA-3wI

Example of an Object Data Model

• Object Data Model for Hydrology
• Basins, sub-basins, streams and gauging stations as “Objects” rather than individual pixels
• Schematic Diagram of logical model showing the relationships and topological constrains for the Objects

Bolstad, Fig2-39

See GeoTech Center website (https://geotechcenter.org) �for additional Model Courses and other curriculum resources. ���������Note: some content is a derivative of other authors���

Author: Ann Johnson

Associate Director

Email: ann@baremt.com

​

2-12-2021 V10