huong dan su dung tien xu ly anh ve tinh bang ILWIS

Looijen PRE-PROCESSING IMAGES • This phase includes the standard operations of geometric correction, image referencing, image enhancement and, if needed, the application of filters • Al

Pre-processing imagery

Joan Looijen

TOPICS

• Why land cover mapping?

• Land cover classification

• Data requirements

• Pre-processing imagery

(FC & NDVI)

• Image classification

• Data collection

• Land cover mapping

• Accuracy assessment

J Looijen

PRE-PROCESSING IMAGES

• This phase includes the standard operations of

geometric correction, image referencing, image

enhancement and, if needed, the application of filters

• All data are georeferenced to UTM WGS 84,

Zone 48 N

COLOUR COMPOSITES

• Besides geo-referencing and radiometric

corrections we need to make sure that the objects

of interest can be distinguished

• The spectral information stored in the separate

bands can be integrated by combining them into a colour composite

• Many combinations of bands are possible

• The spectral information is combined by displaying each individual band in one of the three primary

colours: red, green and blue

J Looijen

ELECTRO-MAGNETIC SPECTRUM

REFLECTION CURVES VEGETATION

• If in land cover classification and mapping you will

mainly concentrate on the mapping of vegetation

cover, including agricultural fields, you have to select those band combinations that give you the ideal

reflectance

• This differs per satellite image

J Looijen

FALSE COLOUR COMPOSITES

• Human eye cannot see the high reflection in the near infrared and middle infrared radiation

• Since the human eye is very sensitive for different

shades of red, the infrared band is displayed in red

• False Color Composite (FCC):

 the red color is assigned to the near-infrared band,

 the green color to the red visible band and

 the blue color to the green visible band

REFLECTANCE CURVES - SENSORS

J Looijen

FC LANDSAT

• Which band combinations (RGB) would you

select to create a FC based on Landsat TM 5?

• Which band combinations (RGB) would you

select to create a FC based on Landsat 8 OLI?

VNREDSAT 1

• Which band combinations (RGB) would you select

to create a FC based on VNREDSat 1 imagery?

J Looijen

VNREDSat 1

Similar to?

EXERCISE FC

• Creation and use of a false colour composite & DEM

J Looijen

NDVI

RATIO IMAGES

• Ratio images are often useful for discriminating

differences in spectral variations

• Which spectral ratio to use depends upon the

particular reflectance characteristics of the features involved and the purpose of the land cover map

• For example a near-infrared / red ratio image might

be useful for differentiating between areas of

stressed and non-stressed

 Drought

 Pollution

J Looijen

NDVI

• If you would like to know the presence and

condition of the green vegetation, preparation of a Normalized Difference Vegetation Index (NDVI) maybe useful

• The NDVI is calculated as:

NDVI = (NIR — VIS)/(NIR + VIS)

• NDVI values range from -1 to 1

NDVI

How would you estimate the NDVI values for:

 No vegetation (bare soil, rocks):

 Dense green vegetation:

 Water , clouds, snow:

J Looijen

EXERCISE

• Creation of a NDVI