Enhancing Remote Sensing Image Interpretation Through Band Ratioing and Vegetation Indices

Essay, Pages 6 (1482 words)

Views

It is sometimes difficult to interpret, classify or even identify the correct surface materials or land use in a remotely sensed image due to the differences in the brightness values of the objects. That kind of condition is usually caused by topographic slope and aspect ,shadows, or seasonal changes in sunlight illumination angle and intensity. Thus, the other useful and effective way to process satellite images especially in areas where topographic effects are important is by band ratioing. Band ratios are the mathematical operations of dividing pixels one band by the corresponding pixels in another band to create a spectral index and to produce images with relative band intensities.

Don't use plagiarized sources. Get your custom paper on

“ Enhancing Remote Sensing Image Interpretation Through Band Ratioing and Vegetation Indices ”

Get high-quality paper

NEW! smart matching with writer

Therefore, spectral indices are the transformations of bands that accentuates the spectral features to easily being recognised as they appear distinct from other image features. For instance, the type of index are the vegetation indices, water and also burned areas. Thus, the band rationing can enhance the spectral differences between bands and reduce any effects of the topography.

Additionally, the most commonly used vegetation indices is Normalized Difference Vegetation Index or in short, NDVI. It is an index of plant greenness or photosynthetic activity. The indices are based on the observation that different surfaces reflect different types of light. Photosynthetically active vegetation absorbs most of red light while reflects much of near infrared light. Meanwhile, the dead or stressed vegetations reflect more red light and less near infrared light. Next, the indices that can be used to differentiate water from the dry land or rather most suitable for water body mapping is the Normalized Difference Water Index (NDWI).

It uses near infrared and green bands which can boost the water information efficiently in most cases. Indeed, water bodies usually have a low radiation and strong absorbability in the visible infrared wavelengths range. In fact, the outputs from the NDVI and NDWI indices can be categorised as Combined Mangrove Recognition Index or called as CMRI. It mostly has been used to assess the mangrove vegetations using the information from the greenness and the water content.

The objective of this experiment is to find the value of NDVI, NDWI and also CMRI using the previous result from the atmospheric correction experiment. Firstly, this experiment also used ENVI as a main tool to find the values of NDVI, NDWI and CMRI. The students need to use the previous files of the reflectance values for each band at the available band lists box. Next, the formula to find the vegetation, water and mangrove indices were written as references. Thus, the formula for the NDVI which is (b5-b4)/(b5+b4) was written in the band math expression which the band 5 and band 4 are referring to the reflectance values of each band. After that, the enhancement subset will be displayed. The value for the index displayed at the cursor location when the cursor is moved at that image.

The steps were repeated for the NDWI and NDWI. In fact, the mathematical expression for the NDWI is (b3-b5)/(b3+b5) where reflectance values of each band were chosen meanwhile the formula for CMRI is NDVI-NDWI. The NDVI is labelled as b1 and NDWI as b2 to prevent any misunderstanding in choosing the files. Finally, the images for all three indices and the data at the cursor location were captured and displayed at the result part. Generally, band ratio, the comparison of at least two bands is an easy way to take advantage of the properties of the objects as each object has a spectral signature. It is widely used in remote sensing because it can be used to enhance the differences between bands by suppressing other features which are not involved. Therefore, the idea of certain wavelengths of spectrum of sunlight being absorbed and reflected when the sunlight strikes objects as can be seen through a prism play a big role in this technique. In addition, the bands covered by this sensor are band 3 (green) , band 4 (red) and band 5 ( near infrared) .

NDVI, the vegetation indices, helps to compare images over time to look for ecologically significant changes, compensate for differences both in illumination within an image due to slope and aspect or time of day or season and are used by researchers to determine the density of green on a patch of land. They observe the wavelengths of the visible and near-infrared sunlights that are reflected by the plants (NASA, 2000). Chlorophyll, for example, strongly absorbs about 0.4 to 0.7 µm of visible light to do photosynthesis meanwhile the cell structure of the leaves strongly reflects about 0.7 to 1.1 µm of near-infrared light. On the other hand, the expression used to calculate the index is (b5-b4)/(b5+b4) whereas the band 5 is infrared and band 4 is red.

So, the actual formula for the NDVI is (NIR-RED)/(NIR+RED). It is calculated on a per-pixel basis as the normalized difference between the near infrared and red band value for a cell. Thus, NDVI can be applied for any image that has a red and a near infrared band. From the result above, the data value is positive because the cursor was put at the white area which contains vegetation at the mangrove areas. In fact, it will turn negative if the cursor is moving to the dark area which is probably water as the water reflects very little NIR light. . The negative values typically do not have any ecological meaning. The low NDVI values signify that there is only a little difference between the red and NIR signals, little photosynthetic activity, or just very little NIR light reflectance. Therefore, the higher values of the output means there are huge differences between the red and near infrared radiation recorded by the selected sensor.

It also identifies that the area is associated with highly photosynthetically-active vegetation. Plant photosynthetic activity, total plant cover, biomass, plant and soil moisture, and plant stress are the examples of factors that affect the NDVI values. However , there are limitations of NDVI which might affect the NDVI values such as the atmospheric conditions,scale of the imagery, vegetation moisture, soil moisture, overall vegetative cover and differences in soil type. So, it is important to control these factors. After all, the output of NDVI is a new image file or layer (Jgillan, 2013).

Next, the opposite result of NDVI is NDWI. It is the appropriate water absorption index. These water indices are used to delineate land from open water and to achieve the signature differences between water and other objects through analysing the signature features of each ground target among different spectral bands. Similarly, the values of NDWI also ranges from -1 to +1. The positive values signifies the presence of extensive water bodies and the negative for vegetation cover. Generally in NDWI, the values of data that is greater than 0.5 represents water bodies. Vegetation indeed has much smaller values which distinguishes them easily from the water bodies as the build-up features usually having positive values lies between 0 to 0.2. In fact, the formula to calculate the NDWI is (GREEN-NIR)/(GREEN+NIR).

Thus in this experiment, the same formula applied, (b3-b5)/(b3+b5) as band 3 represents green and band 5 is NIR band . In comparison, the data in the result shows positive value as the cursor is placed at the bright area and it will turn to negative if it moves to the dark area, the vegetation. GAO (1996) and Huang etal. (2009) presented a modified Normalised Difference Water Index (NDWI) which is known as NDWI GAO to minimise the errors of NDWI on soil water monitoring using Near-infrared (NIR) as earlier and middle infrared (MIR) instead of green band of the TM data (as cited by Sanhu, 2013).

Lastly, CMRI produces better accuracy to map the mangrove areas in the Kilim compared to other vegetation indices. Basically, the index either uses greenness and wetness index values collected from satellite data considering both high and low tide conditions, or using leaf water content and overall health condition, the greenness. Additionally, mangroves exhibit substantial tolerance to a wide range of soil salinity and have high water content in their leaves which enables them to thrive properly under high saline conditions. The values of CMRI in this experiment range from -2 to +2.

These positive values are higher compared to NDVI values and lower than NDWI values when negative. It can be seen that the values of CMRI are up to 1 and above, and -1 and below. When the cursor is placed at the plants area, the value turns positive and vice versa if it moves to the water bodies . Thus, a simple algorithm was used to get the CMRI. The outputs of the NDWI must be subtracted from NDVI values at pixel level. Subtraction was found to increase the upper and lower range of the overall output if the outputs are negatively related. Thus, it also eventually increases the scope of distinction between two classes with near-similar spectral signatures.