Home > 1. Data Basics > Scales/Resolution

1.11 Raster Scales / Vector Resolution

  • Exercise Title:  Raster Scales and Vector Resolution

  • Abstract:  A table published by SCANEX.RU is offered as an example of how satellite raster images are referred to "scales" when published.  Scales are widely cited in environmental literature and management documents, and this table allows users to make rough approximations of the pixel sizes of remotely-sensed images that meet these requirements.  Then a method is described for roughly estimating the "resolutions" of vector objects, based on the numbers of data points needed to draw mapped features.  This method is appropriate for natural features only, but might yield too-large resolution values for political or cultural features, due to long, straight lines with few points.

  • Preliminary Reading (in OceanTeacher, unless otherwise indicated):

    • Map Scale

    • Scale (Map) - Wikipedia article

    • Image Resolution and Map Scales - SCANEX.RU website paper, with useful tables relating imaging systems to reporting scales for topographic maps and thematic maps.  Range of scales from 1:10,000 to 1:1,000,000, for pixel sizes ranging from less than a meter to hundreds of meters.

    • Remote Sensing Resources - American Museum article defining satellite image resolution

  • Required Software:

  • Other Resources: 

  • Author:  Murray Brown

  • Version:  3-10-2015
1. Nominal "scales" recommended for images, from the SCANEX paper, cited above.  [Some re-arrangement of the rows for readability; see the source for the sensors identifications for the pixel sizes.]

This table is probably a good general reference, as long as it is not taken to be rigorous, but just suggestive.  Read the original paper to see more than this condensed view.

 

Pixel Size(m)

"Scale"
10K-25K 25K-50K 50K-100K 100K-200K 200K-500K 500K-1M >1M
250-1000              
140              
35-45              
30              
25-30              
23.5              
23.5              
20              
15-30              
15              
10              
5.8              
5              
5              
2-2.8              
0.8              
2.  If you use this very rough tabular representation, then the most honest textual description would read something like, "this image, at a nominal scale of xxx-xxx, was derived from the xxx-m image products of the XXX program".  So what does this mean for us if we're deciding if an existing raster product is "good enough" for our marine project.  It depends on what you're looking for, and is best summarized in this excellent statement:

"For pixel based products, you can determine the appropriate resolution necessary for your applications based upon the smallest feature you want to resolve (i.e. Minimum Mapping Unit).  The pixel size must be half the smallest dimension of the feature in question. For instance if you want to find a car (which would be ~10 feet  x 6 feet), then the smallest dimension of that car is 6 feet, so your pixel size must be no larger 3 x 3 feet to reliably identify and map cars."  [NCSC Digital Coast]
3.  There is a wonderful instructional graphic in the Italian Catalogo Generale delle  Carte e delle Pubblicazioni Nautiche that lists distance scales (on the left) generally associated with the nautical chart scales (on the right).  The last entry, 1:4000 has been cited in the ecological literature as the minimum scale for good coastal zone management.  Larger values would be preferable, such as 1:1,000.  The chart names (translated) used by Chile are included for a verbal point of reference.

NOTE:  ENC = Electronic Navigation Chart

NOTE2:  NM = nautical mile ~ 1.8 km
Serie selezionabile Scala di compilazione standard delle ENCs Chart Type Name (from Chilean Hydrographic Charts Catalog)
200 NM 1:3,000,000 General Overview
96 NM 1:1,500,000
48 NM 1:700,000 Ocean
24 NM 1:350,000
12 NM 1:180,000
6 NM 1:90,000 Coast
3 NM 1:45,000
1.5 NM 1:22,000 Approach
0.75 NM 1:12,000 Port
0.5 NM 1:8,000
0.25 NM 1:4,000
  (>1:4000) Pier
4.  Now we'll look at methods to identify the "resolution" of vector files, such as coastlines or river courses.
5.  These images are both of the widely-used World Vector Shoreline (WVS).  The upper version is from the Global Self-consistent High Resolution Geography collection, cited above.  The lower version is from GEBCO2003.
6.  If you look very closely, you can see they are almost exactly the same.
7.  But there are tiny differences in how GSHHG (red) handles slanted lines, compared to GEBCO (green).  GSHHG uses multiple short lines, where GEBCO uses single lines.  The overall resolution is obviously identical, but you will see that the GSHHG product gets an artificially finer "resolution" score below, because of this.
8.  Right-click on the first line shape (GSHHG in this case), and select ATTRIBUTES > SHOW.
9.  This opens a table that will be the basis of our "resolution" analysis.  It's interesting, but we need more information.
10.  Select MODULES > SHAPES-LINES > LINE PROPERTIES.  Make these choices, and click OK.
11.  This adds 2 new columns to the table:
  • NPOINTS - The number of points in each section of the line shape
  • LENGTH - The length of each section
12.  Now, to see what the "resolution" of the data, in Saga select TOOLS > TABLE > CALCULUS > TABLE CALCULATOR.
  • Enter a new field name such as DEG/POINT to indicate a new measure of the ratio of length (DEG) to data (POINTS)
  • For TABLE, select the GSHHG shape

Then click OK to add this new field to the data table.
13.  Here you see the new field, DEG/POINT, completely populated with values.  You can see what they are, but to make a good analysis, we need to sort them to see the distribution.
14.  Right-click on the label at the top of the new field, and select SORT FIELDS.
15.  Make these choices to sort the new field in ASCENDING order.  Then click OK.
16.  Now you can see all the values sorted:
  • The DEG/POINT values at the top of the table begin at roughly 0.0006
  • The values near the middle (second image) include the MEDIAN at location 80, with a value of roughly 0.0014 DEG/POINT.
  • The values at the bottom of the table increase sharply to a high value of 0.0054 DEG/POINT.
So the majority of the GSHHG global map is built on data with a resolution of around 0.0014 degrees (the MEDIAN).  Near the equator, a degree is about 100 km, so the resolution length scale is about 0.0014 deg X 100 km/deg X 1000 m/km or roughly 140 m.

The full range of resolutions goes from about 41 m to 540 m, but the minima and maxima "tails" in this distribution are very limited, so a simple visual evaluation of the median is quite justified.
16.  Here are the results for the GEBCO data, calculated in the same way:
  • The values at the top of the table sharply increase from the lowest value of 0.0010
  • The values near the middle (second image) include the MEDIAN at location 90, with a value of 0.0025 degrees.
  • The values at the bottom of the table increase sharply to a high value of 0.0057

 
17.  So the majority of the GEBCO global map is built on data with a resolution of around 0.0025 degrees.  Near the equator, a degree is about 100 km, so the resolution length scale is about 0.0025deg X 100 km/deg X 1000 m/ km or roughly 250 m.

The full range of resolutions goes from about 50 m to 570 m, but the minima and maxima "tails" in this distribution are very limited, so a simple visual evaluation of the median is quite justified.
18  As expected, due to the way the GEBCO handles curvy lines (i.e. slightly longer straight lines), the nominal resolution appears to be greater than GSHHG (250 m versus 140 m).  But this objectively derived "difference" is not consequential.  Data users should make similar analyses of their data, in many cases, to see what's actually present and why.
19. A Final Word of Caution:  So the "resolutions" estimated here for these coastlines lie well below the best "scale" value shown in the raster diagram above.  It is possible to make objective comparisons between raster and vector approaches to data resolution versus scales but you must be aware of the differences and honest in your descriptions and claims for your own data or products.