Normalized Difference Vegetation Index (NDVI) is a ratio of the red light to near-infrared light that our sensor sees. Plants are uniquely efficient at reflecting light in a way that increases the value of this ratio. Higher NDVI values mean more chlorophyll is detected by the sensor. Both leaf area and photosynthetic activity increase NDVI values. Normally, these two values are conflated into a measure of “vigor,” but basically it is a measure of chlorophyll density from whatever cause.
TerrAvion’s NDVI is translated to an 8-bit scale which has pixel values from 0-255, however our NDVI values do not span the whole range. 0 corresponds to the start of the useful range of NDVI values, not zero in a reflectance calculation and 255 is off the scale. The start of vigor, i.e. the sensor is seeing living growing chlorophyll is about 50 on our scale. Non-living materials will have values below 50 and plants that are growing will measure more based on how vigorous they appear to the sensor. Each NDVI value is the same vigor consistently throughout the season, this allows comparisons about photosynthetic potential between locations and through time.
All Channels are Flat Field corrected.
Channels 4,5,6 are normalized to channel 6.
NDVI is calculated from channels 4 and 6.
CIR is generated from channels 4,5, and 6.
Geometry Buffer is the buffer TerrAvion put in place on top of the geometry so that extra imagery is contained in the GeoTIFF in case surrounding information is needed for analysis.
Alpha blending also known as feather is contained in the geotiff, so that tiles derived from the geotiff blend together well when there are overlapping fields.
The Thermal channel 7 scale is in 0.1 Kelvin. For example, pixel intensity 3000, is 300.0 Kelvin, which is 80.33 (°F)
You can roll your own scale and categorization. We provide the NDVI’s as downloadable palletized .tif files that can be downloaded from the interface. These files are geo-registered in WGS 84 Web Mercator Auxiliary Sphere (EPSG: 3857) and they contain NDVI values from 0-255 in Band 1, with corresponding mapping to the NDVI function of that pixel. These files can be downloaded into GIS programs such as ArcGIS or QGIS. There you can stretch the NDVI histogram, classify, or run pixel math to unlock more information. If you are going to do this at scale, contact us about an API for the program you are using.
For more information, contact firstname.lastname@example.org