Map Calculaton using GRASS

Map Calculation with GRASS

Session - II

Workshop on “Introduction to Remote Sensing”, July 7 – 11, 2014, Jawaharlal Nehru Engineering College

In this session● Calculation of vegetation indices● Level slicing of images

Spectral Signature of vegetation

http://ces.iisc.ernet.in/hpg/envis/Remote/introfile8_files/FIG_11.gif

High reflectance of vegetation at the Near infra-red (NIR) region

High reflectance of vegetation at the Near infra-red (NIR) region

Low reflectance of vegetation at the red (R) region

Low reflectance of vegetation at the red (R) region

Vegetation Indices

RVI=NIRR

(Jordan, 1969)

NDVI=NIR−RNIR+R

(Rouse et. al, 1973)

IPVI=NIRNIR+R

(Crippen, 1990)

DVI=NIR−R (Richardson & Everitt, 1992)

Ratio Vegetation IndexRatio Vegetation Index

Normalised Difference Vegetation Index

Infrared Percentage Vegetation Index

Difference Vegetation Index

In our dataset

S. No. Band Part of spectrum

Spatial Resolution

1 B1 Blue 30m

2. B2 Green 30m

3. B3 Red 30m

4. B4 Infra-red 30m

5. B5 SWIR-1 30m

6. B6 Thermal 60m

7. B7 SWIR - 2 30m

Select Aurangabad and Click PERMANENT.

Select Aurangabad and Click PERMANENT.

Start GRASSStart GRASS

Understanding MAPCALC

1. Go to the command console1. Go to the command console

2.Type r.mapcalc

Press Enter

2.Type r.mapcalc

Press Enter

3. Type the name of the new Raster Map RVI

3. Type the name of the new Raster Map RVI

4. Select double(x) from the mapcalc function list.

4. Select double(x) from the mapcalc function list.

5. Type the expression: double (B4S@PERMANENT) / double(B3S@PERMANENT)

5. Type the expression: double (B4S@PERMANENT) / double(B3S@PERMANENT)

6. Click Run6. Click Run

GRASS displays command finishedGRASS displays command finished

7. Click Close.7. Click Close.

8. Check on the RVI layer in Layer Manager

8. Check on the RVI layer in Layer Manager

9. RVI displayed in coloured mode. Need to change the colour table.

9. RVI displayed in coloured mode. Need to change the colour table.

10. Go to the command console. Type r.colors and press Enter

10. Go to the command console. Type r.colors and press Enter

11. Select RVI@PERMANENT 11. Select RVI@PERMANENT

12. Click on the Colors panel and select grey.

12. Click on the Colors panel and select grey.

13. Click Run.13. Click Run.

Computed map displayed in grey.Computed map displayed in grey.

Exercise● Use r.mapcalc to compute

– DVI

– NDVI

– IPVI

● Change the colour table using r.colors

Displays of DVI, NDVI, IPVI

DVIDVI NDVINDVI IPVIIPVI

Segmentation methods using map calculation

Binary and n-nary segmentation / level slicing

Strategy for binary segmentation● Determine a threshold (a number) for an image I● Apply the following function to the image I

newImage= f ( I )={I <τ 1I≥τ 2

How to obtain this threshold?How to obtain this threshold?

The if-then-else function in r.mapcalc

if (<condition>,

<value_if_condition_is_satisfied>,

<value_if_condition_is_not_satisfied>)

● e.g.– r.mapcalc 'newImage = if(b3@PERMANENT < 56, 1, 2)'

Finding the threshold

1. Add B4S@PERMANENT to the layer manager

1. Add B4S@PERMANENT to the layer manager

Water bodiesWater bodies

2. Click on this button (Query raster / vector maps)

2. Click on this button (Query raster / vector maps)

3. Click on various points on the water bodies in the map layer and find the maximum value

appearing here

3. Click on various points on the water bodies in the map layer and find the maximum value

appearing here

4. Go to the command console and type r.mapcalc

4. Go to the command console and type r.mapcalc

5. Enter the name of the new image

5. Enter the name of the new image

6. Type the expression

If (B4S@PERMANENT < 26, 1, 2)

6. Type the expression

If (B4S@PERMANENT < 26, 1, 2)

7. Click Run7. Click Run

Binary segmented image displayed in the map

window. Water body seen here

Binary segmented image displayed in the map

window. Water body seen here

Exercise: It is seen that the second water body is not displayed in the image. Select a better threshold and re-calculate the new image

Exercise: It is seen that the second water body is not displayed in the image. Select a better threshold and re-calculate the new image

Divide an image into three groups● For this, we need two thresholds● We apply the following function

newImage= f ( I )={I<τ1 1τ1≤ I<τ2 2I≥τ2 3

Find these thresholds again by using the Raster query method explained earlier.

Find these thresholds again by using the Raster query method explained earlier.

Using Raster Query tool, we select two thresholds: 31 and 54Using Raster Query tool, we select two thresholds: 31 and 54

1. Go to command window and type r.mapcalc

1. Go to command window and type r.mapcalc

3. Type the expression: if (B4S@PERMANENT < 31, 1, if (

B4S@PERMANENT <56,2,3)). 3. Click Run

3. Type the expression: if (B4S@PERMANENT < 31, 1, if (

B4S@PERMANENT <56,2,3)). 3. Click Run

2. Enter the name of the new raster map as

mewImage2

2. Enter the name of the new raster map as

mewImage2

Image displayed in Map WindowImage displayed in Map Window

Questions● We are able to segment the image in two and three

parts and extract features like ROADS and LAKES● Can we extract other features also?

– Like buildings etc

– Can we make 6 different groups and develop a map?

● Exercise– Identify 5 different thresholds to divide the image into 6

groups

● Knowledge: – The process that we learned here is called LEVEL

SLICING