Accuracy Assessment - PowerPoint PPT Presentation

1 / 19
About This Presentation
Title:

Accuracy Assessment

Description:

Validating an Electronic Land Cover Map ... Students use the 'MUC System' to identify land cover types in their classified image. ... Or their Landsat image, ... – PowerPoint PPT presentation

Number of Views:191
Avg rating:3.0/5.0
Slides: 20
Provided by: larr145
Category:

less

Transcript and Presenter's Notes

Title: Accuracy Assessment


1
Accuracy Assessment
Validating an Electronic Land Cover Map
AA_Vis_1
2
Using the software MultiSpec, students follow
protocols to prepare a clustered image of their
GLOBE Study Site
3
Students examine the true color And
false-color Infrared images provided by GLOBE
to identify land cover types.
AA_Vis_2
4
Students use the MUC System to identify land
cover types in their classified image.
5
Using either a local map,
6
Or their Landsat image,
7
Students gather data from 90 x 90 m Qualitative
Land Cover Samples to check the accuracy of their
map
8
At each of these sites, the MUC value is
determined, a GPS reading is taken at the center,
and photographs are taken in each cardinal
direction. These are Validation Data
Site 1 MUC 0222
Site 2 MUC 4522
.
Site 3 MUC 93
Site 5 MUC 52
Site 4 MUC 4522
Site 6 MUC 72
Site 7 MUC 4522
Site 8 MUC 92
9
The Accuracy Assessment Data Work Sheet is filled
out for all sites.
Landing Woods 0222 0222 v
Rachel Carson_1 4522
4522 v
Burnham Plaza 91
93 v
Rachel Carson_2 4522
4522 v
Dunes Development 91
52 v
Wells Beach 72
72 v
Rachel Carson_3 4522
4522 v
Wells Corner 91
92 v
10
The Difference/Error Matrix is constructed
Validation Data (What we actually found)
0222 4522 91 92 93 72
52 Total
0222
4522
91
Map Data (What we expected to find)
92
93
72
52
Total
11
Map and Validation data are entered on the Matrix
Validation Data (What we actually found)
0222 4522 91 92 93 72
52 Total

0222



4522
91



Map Data (What we expected to find)
92
93
72

52
Total
12
Totals are calculated for rows and columns
Validation Data (What we actually found)
0222 4522 91 92 93 72
52 Total

0222
1



4522
3
91



3
Map Data (What we expected to find)
92
0
93
0
72

1
52
0
1
1
1
3
0
1
1
8
Total
13
Overall Accuracy is calculated
Map Data
Sum of Major Diagonal Tallies
Overall Accuracy -----------------------------
--------- x 100 Overall
Accuracy ----- x 100 63
Total Number of Samples
5
8
14
More Qualitative sites are visited, and the
Matrix is developed for all classes on the map
Validation Data
0222 4522 91 92 93 72
52 62 63 Total
12
0222
12
6
6
4522
91
14
8
4
2
Map Data
92
5
4
18
9
93
8
1
9
72
3
3
3
2
5
52
2
62
3
5
4
3
63
7
12
6
13
16
14
3
3
6
6
Total
79
15
Overall Accuracy is Calculated
Validation Data
Map Data
57
Overall Accuracy --------- x 100 72
79
16
Producer Accuracy is Calculated
Producer Accuracy tells us how well we did in
identifying Land cover types in our map.
Correctly Identified
8
Producer Accuracy ---------------------
-------- x 100 -------- 62
13
Column Total
17
User Accuracy is Calculated
Validation Data
User Accuracy tells us how accurate someone else
would be using our data in the field
Correctly Identified
8
User Accuracy -------------------------
---- x 100 -------- 57
14
Row Total
18
REMOTELY SENSED FOREST VEGETATION MAPSRASTER
MAPS CREATED FROMLANDSAT TM IMAGERY
July 1996 Landsat TM Image
Pawtuckaway Study Area, NH, USA
Forest Cover Type Map from Classification of July
1996 Landsat TM Image Pawtuckaway Study Area, NH,
USA
Legend
Class Names White Pine White Pine-Hemlock Other
Conifer Hemlock Oak Red Maple Mixed
Forest Nonforest Beech Other
Scale
Kilometers
AA_Vis_13
19
ERROR MATRICES ANDESTIMATES OF ACCURACY
AA_Vis_15
Write a Comment
User Comments (0)
About PowerShow.com