Creating NDVI from multispectral imagery?

I have LISS IV image data in three bands (band4, band3 and band2). I wanted to know the required steps to obtain the NDVI image from the given data using ERDAS IMAGINE 9.2.

• Hi! Welcome to GIS.SE What do you mean by 'obtain NDVI value'? What have you tried so far? – Devdatta Tengshe Apr 2 '13 at 9:46
• not sure may be useful for you shibayan.webs.com/NDVI%20by%20ERDAS.pdf – Sunil Apr 2 '13 at 10:27
• What do you mean by "NDVI Image"? What bands are in it, where did you get it, what software do you have availible? – til_b Apr 2 '13 at 11:59

To begin, you need to know the which spectral bands are which in your base image. NDVI is calculated from reflectance rather than radiance or DN. Therefore you will need to make sure your imagery has been converted to express reflectance. The equation to calculate NDVI is as follows:

(near infrared - red)/(near infrared + red)

If you are using LISS IV imagery, this translates to the following band combinations:

(band 3 - band 2)/(band 3 + band 2)

Luckily Erdas has a handy tool that quickly calculates NDVI from these bands (Figures 1 & 2). The resulting NDVI will have values ranging from -1 to 1, where high positive values represent healthy green vegetation.

To check the results of your NDVI calculation choose the Inquire tool (Figure 3). You can see the tree I selected with the cross hairs displays a pixel value of 0.402. To further investigate the distribution of values within your NDVI, select Metadata > histogram (Figure 4). Here you can see the pixel values range from -0.84 to 0.65 with the majority pixel values centered about 0. From an ecological perspective, this makes sense given the sparse vegetation within the semiarid desert this NDVI is derived from.

Figure 1

Figure 2

Figure 3

Figure 4

• Thanks for the help.I tried doing the same exercise on LISS IV bands. Band 4- NIR, Band 3- R but i get a error massage of Invalid layer list. I don't know what went wrong. – user17950 May 8 '13 at 10:15

@Aaron provided the procedural answer that you should follow although, you should be changing the Sensor type to SPOT XS/XI. Please take note that the IRS-P6 LISS IV sensor does not have a blue band and, in the context of sensors that have R,G,B as the first 3 bands, bands 1 (green) and 2 (red) are reversed. On the LISS IV sensor Band 4 is the mid-infrared not near-infrared.

You need to look at the sensor specifications and select the correct bands for NDVI. It would help tremendously if you took the time to read a few basic articles on NDVI and research your sensor so you can make better informed decisions on your analysis.

The spectral ranges of LISS IV are: B1=0.52-0.59nm (green), B2=0.62-0.68 (red), B3=0.77-0.86 (near-IR), B4=1.55-1.70 (mid-IR), pan=0.62-0.68

Since band 2 is the red band and band 3 the near-infrared (NIR) these are the bands that your want to ratio to derive NDVI.