Raster Analysis

Measuring cells

Perimeter = # of cells of the perimeter * resolution

  • Perimeter of cells is more estimated than calculated
  • Units are the same as the units in the resolution
  • For vector analysis, we calculate the perimeter based on the coordinates, so it is more precise than in raster analysis

\ Problem: the more complex the polygon, the more grid cells will be at the diagonal to their neighbour, resulting in less accurate results

  • Pythagoras theorem to find the diagonal distance of the cell (L2 = d12 + d22 = 2*d2)
  • Diagonal distance > lateral distance

\ Area = # of cells * area of a cell

-       Area of a cell is calculated by the resolution

\

Measuring distance

There are 2 ways of measuring distance in raster data:

  1. Euclidean distance

       1. For each cell, the distance is calculated to the source cells(s) by calculating the hypotenuse of the square triangle and multiplying the distance by the resolution

  1. Cell distance

       1. Calculate the distance for the origin through the center of neighboring cell       * Useful for calculating the path from different points of the grid (i.e. shortest path)       * Useful for cost distance

\

Cost distance

Calculates the least accumulative cost distance for each cell from or to the least-cost source over a cost surface.Method of calculating cost distance, is by calculating the cost of the links between the cells or the importance of each cell:

*Lateral link = 1 * ((Ci + Cj)/2)

Diagonal link: 1.414 * ((Ci + Cj)/2)

Where:

  • Ci – Value of cell i (seen in Cost_Ras)
  • Cj – value of cell j (seen in Cost_Ras)

\ From the cost distance, derive the accumulative cost path and choose the least accumulative cost path

  • Sum of all the accumulative cost between a destination and the source
  • Keep the path with the lowest value

\ Requires: a source raster, a cost raster, derived distance cost, a destination, and an algorithm for calculating the least cost path

\

Convert raster to vector, vector to raster

Transform a raster file into a vector file (features)

  • All neighboring cells are connected to each other to determine line or polygon limits
  • Points become the center

Converting vector to raster

  • The value of each cell is based on the value of ONE selected field
  • Different methods to determine the value of a cell sharing multiple polygons
  • Errors are common in the converting process

\

Raster analysis – Operators and Functions

The functions associated with raster cartographic modeling can be divided into five types:

  • Local functions   * Works on a cell in a single location
  • Focal functions   * Work on cell locations within a neighbourhood
  • Zonal functions   * Work on cell locations within zones
  • Global functions   * Work on all cells within the raster
  • Applications   * Those that preform a specific applications (i.e. hydrologic analysis function)

\

Raster Overlay

Overlaying data means to put the data on top of each other, allowing us to combine several grids and produce a new raster layer.

  • Combine 2 or more raster files
  • You may need to reclassify one (or more) of the files to create a useful overlay
  • Raster calculator - use arithmetic, logical, and Boolean operators to select areas of interest

\ You can add one grid to another one:

  • The 0-class typically refers to a component that we are not interested in investigating

\ Masking allows selection of pixels from one grid based on pixels from another grid when we are overlaying raster data:

  • Mask grid must be binary (0 or 1)   * 0 = I don’t want   * 1 = I want
  • The mask can be used to ‘clip’ other layers

\ If one grid has a higher resolution compared to another, you should divide the bigger resolution one to match the smaller resolution grid to not lose information (Resampling)