Lecture 11: Spatial Analysis of Raster Data (Link to Powerpoint File)

Lab 12: Raster Data II

Homework 7: Vector and Raster Analysis

Raster Analysis Basics
GISs can display data in various formats but usually can only use data in a specific format (e.g. ArcGIS can only analyze grids).
Raster analysis is based on the cell as the basic unit of analysis
Can perform analysis on individual cells
Can analyze data on a group of cells
Can perform analysis on all cells within a grid
Analysis can operate on single raster grids or multiple raster grids
Data Analysis Environment
Specifies the extent of the analysis area
Specifies the cell size of the output grid
Mask Grid
Can Also be used to define the area of analysis

Raster Analysis Local Operations
Single grid local operations
Compute new values for each grid cell to create a new output grid
Mathematical operations
Reclassification
Multiple grid local operations
Create a new output grid by combining data from multiple grids
Similar to vector based overlay techniques but conceptually simpler
Can use these to compute summary statistics

Raster Analysis Neighborhood Operations
Neighborhood operations operate on a given cell by analyzing the value of surrounding cells
Move through each cell and compute a new value based on it’s value and those of cells in the neighborhood
Often times is based on a cell and all it’s surrounding neighbors
Can also use larger neighborhood areas or even just those neighbors on certain sides
Neighborhood operations are often useful in image processing
Blurring or Average filtering
Sharpening operations
Edge enhancement
Terrain operations for digital elevation data are common as well

Raster Analysis Zonal Operations
Zonal operations operate on groups of cells that share the same values
Zones may be contiguous or non-contiguous
Zonal operations can work on single or multiple grids
Single grid zonal operations
Identify the boundary of zones that contains the same values
Identify the center of zones where similar values exist
Multiple grid zonal operations
Summarizes the cell values for one grid based on the cell values of another grid

Raster Analysis Distance Measures
Distance operations
Calculates the distance from each cell to the nearest cell containing a certain value (e.g. the distance to the nearest stream cell)
Physical distance measures
Uses cells as the basic unit of cell measurement
Basically the same as computing the euclidean distance
Cost distance measures
Calculate distance based on the impedance to traverse each cell
Cost distance for a cell is the sum of the costs for each individual cell
Process can be used to try to calculate the least cost path from one location to another

Raster Data and Spatial Autocorrelation
Spatial autocorrelation is a measure to which there is “structure” in the underlying spatial relationship of the data
Relationship is highly correlated if nearby cells tend to have the same cell values
Relationship is negatively correlated if nearby cells tend to have very dissimilar values
Relationship is random if there is no discernable spatial patterning in the distribution of values