Contrasted NDVI: The story behind this user-requested feature

In 2021, agriculture is as technologically advanced as oil and gas production or even the airspace industry. OneSoil's role is to provide farmers with data about how their fields are doing so they can bolster their results. We process satellite data using algorithms we've developed in-house and share the results with farmers via our apps in an easy-to-understand and convenient format.

OneSoil evaluates plant conditions using the NDVI index, which is a numeric index that indicates the quantity and quality of vegetation in the field. The NDVI index is calculated based on how plants reflect sunlight (healthy plants actively absorb the red light and reflect near-infrared light).

Contrasted NDVI is a layer in our mobile and web apps that shows the difference in NDVI values more clearly through bright and contrasted colors (Learn more about the correlation between your plants' health and the NDVI index in our blog post.)

Now, back to business: what did it take to deliver this feature requested by our users? For that, OneSoil's team members weigh in. Meet Valerie Matskevich, Philip Kondratenko, and Eugene Voronets.

Our feature release wasn’t as intense as an epic blockbuster battle. But we know our users had definitely been waiting for it no less than another Batman movie. And it was farmers themselves who suggested that we should add contrasted NDVI.

Let's see what the fields near Balaton, Minnesota, look like in GIS software.

However, even with this filter on, a farmer wouldn't understand a thing about his or her fields. Besides, the filter doesn't distinguish fields from other objects and it can be pretty hard to understand what's really going on in a particular area. Here's when the OneSoil app comes in handy.

As you can see, our images give a better idea of what's going on in the fields. Things are pretty good in the areas colored dark green. In red zones, something suspicious is happening. (Check out this blog post to find out how we turn satellite images into NDVI indexes.)

Standard NDVI gives you a general idea of what's going on. At the same time, Contrasted NDVI is the same picture, but things are a little bit exaggerated. In a picture with NDVI between 0.7 — 0.8, the whole field will be green. Contrasted NDVI gives more nuance, and farmers can notice things they wouldn't have seen otherwise.

It took only a month from the time Valerie shared user input with the team to the feature’s release. And just for a little context, a month is pretty fast for a company that creates all of its features from scratch.

In a standard NDVI image, the field is colored from brown to green to indicate values from −1 to 1 (0 is the worst NDVI value possible, 1 is the best, and negative values indicate extraneous objects in the image, such as snow, water, rocks, etc.). Contrasted NDVI uses a wider palette (from red to blue), and its values are calculated differently. Instead of ranging from 0 to 1, it covers the difference between the minimum and maximum values in the specific field.

We didn't expect any special challenges with this (and didn't face any), but the advantages for users were huge.

Eugene continues.

We used to store different versions of satellite data (for example, standard NDVI and contrasted NDVI) in one place. But, by the time we started developing Contrasted NDVI, this was no longer a viable option. With the new architecture in place, we store a single raw image version. We color it on the fly depending on the user’s request.

It’s a great feature that can be developed almost in no time. If it hadn’t been for the architecture, we’d have been ready with it a lot faster.

Sure, sometimes we still receive questions like "What is Contrasted NDVI?" or "What should I do with it?" For many of our users, Contrasted NDVI is still something new. However, we're not the first to do it, and our competitors have it in their apps, too.

Moreover, we offer this feature for free, unlike our competitors. And finally, it's great to know our users like it. On the release day, several people sent us their positive feedback: "Thank you, it's a great feature!"

As you see, there was nothing extra challenging in developing this new feature. As Eugene Voronets puts it, "It wasn't rocket science." Nevertheless, the advantage it gives users is clear.

Our apps also include very complicated features that require months of hard work to make, such as Cloud-free NDVI or the Crop rotation chart. What makes this feature special is that Contrasted NDVI was requested by our users, it was easy to do, and it adds real value to our products. It seems like sometimes we need to look at things in a whole new light.