Difference between revisions of "VerticalAccuracy"

Line

Component

Measure Description

1

Name

absolute linear error at 90 % significance level of biased vertical data

2

Alias

LMAS

3

Data quality element

Absolute or External Positional Accuracy

4

Data quality basic measure

5

Definition

absolute vertical accuracy of the data’s coordinates, expressed in terms of linear error at 90 % probability given that a bias is present

6

Description

See STANAG 2215 ed7 Appendix 2

7

Parameter

8

Data quality value type

Measure

9

Data quality value structure

10

Source reference

NATO STANAG 2215 ed7

ISO 19157 measure n°40

11

Example

12

Identifier

http://dgiwg.org/metadata/qualityMeasure/LMAS

Line

Component

Description

1

Name

absolute linear error at 90 % significance level of biased vertical data (Alternative 2)

2

Alias

ALE

3

Element name

absolute or external accuracy

4

Basic measure

not applicable

5

Definition

absolute vertical accuracy of the data’s coordinates, expressed in terms of linear error at 90 % probability given that a bias is present

6

Description

A comparison of the data (source) and the control (reference) is calculated in the following manner:

1. Calculate the absolute error in the vertical dimension at each point:

δ V i = source V i − reference V i for i = 1 … N

2. Calculate the mean vertical error:

| δ V ¯ | = | 1 N ∑ i = 1 N δ V i |

3. Calculate the standard deviation of the vertical errors:

σ V = 1 N ∑ i = 1 N δ V i 2

4. Calculate the ratio of the absolute value of the mean error to the standard deviation:

ratio = | δ V ¯ | / σ V

5. If ratio > 1 , 4 , then k = 1 , 2815

6. If ratio ≤ 1,4, then calculate k based on the ratio of the vertical bias to the standard deviation of the heights using a cubic polynomial fit through the tabular values as defined in the Handbook of Tables for Probability and Statistics (Reference[ 20 ]).

k = 1 , 643   5 − ( 0 , 999   556 × ratio ) + ( 0 , 923   237 × ratio 2 ) − ( 0 , 282   533 × ratio 3 )

7. Compute LE90 for the source:

LE90 source = | δ V ¯ | + ( k × σ V )

8. Compute absolute LE90:

LE90 abs = LE90 reference 2 + LE90 source 2

7

Parameter

Name: Sample size

Definition: minimum of 30 points is normally used but may not always be possible depending on identifiable control points. For feature level attribution sample 10 % of the feature population.

Value Type: Real

8

Value type

Measure

9

Value structure

-

10

Source reference

ISO 19157 n°41

1. Mapping, Charting and Geodesy, Accuracy (Reference[ 21 ])

2. Handbook of Tables for Probability and Statistics (Reference[ 20 ])

3. NATO STANAG 2215 IGEO (Reference[ 22 ])

11

Example

-

12

Measure identifier

http://dgiwg.org/metadata/qualityMeasure/ALE

Line

Component

Description

1

Name

absolute linear error at 90 % significance level of biased vertical data (Alternative 2)

2

Alias

ALE

3

Element name

absolute or external accuracy

4

Basic measure

not applicable

5

Definition

absolute vertical accuracy of the data’s coordinates, expressed in terms of linear error at 90 % probability given that a bias is present

6

Description

A comparison of the data (source) and the control (reference) is calculated in the following manner:

1. Calculate the absolute error in the vertical dimension at each point:

δ V i = source V i − reference V i for i = 1 … N

2. Calculate the mean vertical error:

| δ V ¯ | = | 1 N ∑ i = 1 N δ V i |

3. Calculate the standard deviation of the vertical errors:

σ V = 1 N ∑ i = 1 N δ V i 2

4. Calculate the ratio of the absolute value of the mean error to the standard deviation:

ratio = | δ V ¯ | / σ V

5. If ratio > 1 , 4 , then k = 1 , 2815

6. If ratio ≤ 1,4, then calculate k based on the ratio of the vertical bias to the standard deviation of the heights using a cubic polynomial fit through the tabular values as defined in the Handbook of Tables for Probability and Statistics (Reference[ 20 ]).

k = 1 , 643   5 − ( 0 , 999   556 × ratio ) + ( 0 , 923   237 × ratio 2 ) − ( 0 , 282   533 × ratio 3 )

7. Compute LE90 for the source:

LE90 source = | δ V ¯ | + ( k × σ V )

8. Compute absolute LE90:

LE90 abs = LE90 reference 2 + LE90 source 2

7

Parameter

Name: Sample size

Definition: minimum of 30 points is normally used but may not always be possible depending on identifiable control points. For feature level attribution sample 10 % of the feature population.

Value Type: Real

8

Value type

Measure

9

Value structure

-

10

Source reference

ISO 19157 n°41

1. Mapping, Charting and Geodesy, Accuracy (Reference[ 21 ])

2. Handbook of Tables for Probability and Statistics (Reference[ 20 ])

3. NATO STANAG 2215 IGEO (Reference[ 22 ])

11

Example

-

12

Measure identifier

http://dgiwg.org/metadata/qualityMeasure/ALE