So my question is as follows:

I am a GIS tech for a large vineyard management company. Every year a real struggle of ours is getting a crop estimation (with +- %15 accuracy, company wants it to be +-%5). I currently use 1)Historical Bunch Weights 2)Average amount of producing vines in a vineyard block 3) average Bunch numbers per vineyard block.

This has all been a very slow, painful experience with needing to send guys in the field with the Collector apps I create to manually count bunches, or producing plants vs dead/missing plants.

So my real question is that I am currently trying to use our NDVI's from last year (which I have in Arc already) to somehow extrapolate out the "producing vines" part of the equation. AKA I need to convert the NDVI to a line/point (?) feature that only contains producing vines (or dead/missing vines since that would be a smaller file). I then can send the field guys out to count bunches in pin-pointed regions of high vigor etc to estimate.

Any thoughts?


1 Answer 1


Provided that you have a raster that looks like this:

vine ndvi

what you want is to vectorize certain portions of your ndvi. You can use gdal_polygonize. This answer gives you an example of how to use it. Once you have the polygons (make sure you play with the settings to get the smallest posdible on), calculate centroids. There are other ways to get more accurate results (skip polygon calculation and generate points directly from pixels), but this would require you to give more info about how much area each pixel is covering. I hope this helps.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.