What's the best mission to retrieve NDVI data?

Agriculture
#1

Hello!

I’m new to this and I’m trying to develop an app for farmers and I need to collect NDVI data from parcels. I’ve read some documentation and the built-in scripts and tried to implement it to my code. However, I’m not sure which is the best mission or script to use. I’ve used the data fusion between Sentinel-1 and Sentinel-2 so I could remove the clouds, but i dont know if the data im getting is accurate. Am i using the right method? If so, how can i check if the color map im getting is the intended?

Thank you so much!

#2

Hi Tiago,

Thanks for the question and welcome to the world of remote sensing! It is a good question, and assuming that you wish to use only free satellite imagery then Sentinel-2 and Landsat 8/9 are the best missions to use when developing an application that requires NDVI.

Sentinel-2 will give you 10m resolution with a revisit time of 5 days, whereas Landsat 8/9 is 30m resolution and has a revisit time of 8 days. In combination, this will give you plenty of data. You can even use Harmonized Landsat Sentinel data. collection for this.

You also mention data fusion using Sentinel-1, so I want to emphasise that Sentinel-1 is a SAR satellite so you can’t generate NDVI from this sensor. There are equivalent indexes generated from SAR satellite images that can be used instead of NDVI though, for example, the Radar Vegetation Index.

Hope that this information helps you out, let us know if you have any other questions you need clarifying!

#3

Thank you so much! What if I want to upgrade to paid satellite imagery? Whats the best way to get accurate NDVI for a field?

#4

Hi Tiago,

The NDVI formula doesn’t change so if you want to utilise paid satellite imagery, as long as the sensor has a Red and NIR band then you can calculate NDVI from this.

We offer data from the following providers:

#5

but what about cloud coverage? isn’t that an issue when calculating the NDVI? that was the reason why I used data fusion, in order to remove the clouds

#6

Naturally, clouds are an issue in optical remote sensing. Unfortunately, there’s not much you can do about them other than filtering the images you are using by the cloud coverage percentage or using cloud masking within your analysis.

You can also experiment with using composites, which utilise multiple cloud masked images to create a mosaic of several acquisitions. However, this may not be the best approach if you require accurate values attributable to a time and date.

I would encourage you to research the different methods used by the community in our custom scripts repository here. There are several scripts that calculate metrics using NDVI here. For example, this script evaluates the NDVI for each scene of the past month and returns the highest NDVI value for every pixel.