Multispectral NDVI measurements

Projects & Applications - Agriculture

Multispectral imaging is an important technology for remote sensing in aerial applications. The data from NASA's Landsat satelite family have resulted in both a lot of academic research as well as numerous practical applications. Global vegetation monitoring has been greatly influenced by these satelites and their multispectral sensors that captured wavelengths from around 500 to 1000nm.

The Normalized Differential Vegetation Index (NDVI) is a method to segment regions of vegetation in an aerial images. The key notion of this index is the fact that vegetation responds in a unique way to light of different wavelengths. The green color of foliage is caused by a pigment called chlorophyll, which is also responsible for the photosynthesis process that transforms CO₂ to O₂. There are several types of chlorophyll, each with a slightly different absorbance response.

Chlorophylls have high absorbance in blue and red regions, but low in the green and infrared (not shown in the absorbance plot). These differences in absorbance (and thus also in reflection) can be exploited to measure the amount of vegetation in an area.

The NDVI value is computed by:

NDVI = (NIR – VIS) / (NIR + VIS)

It always results in a value between -1 and +1 and for visualization the NDVI output is sometimes mapped to greyscale images ranging from 0 to 255.

Several variants to the NDVI have been proposed. The Enhanced Vegetation Index (EVI) is a version that takes also the blue channel into account. Another variant, the Soil Adjusted Vegetation Index (SAVI) was developed for situations where the vegetative cover is low and soil surface is exposed to the measuring system. Yet other variants exist, and each method stems from the NDVI in an attempt to improve on its results for certain applications.

The Condor-1000 multispectral cameras have channels that are similar in wavelength to those of the multispectral sensors of Landsat. The Condor-1000 MS2 cameras can provide exactly the VIS and NIR channels needed for NDVI, while the Condor-1000 MS4 cameras can provide all bands needed for the Extended Vegetation Index. The Condor-1000 MS5-VNN-285 camera comes closest to Landsat because it splits the near infrared in two separate channels in addition to the full resolution of red, green and blue channels.

You may also want to take a look at NASA's Landsat website.