Multitemporal NDVI analysis for the identification of different crop species and extraction of their phenological features

Satellite imagery is becoming an increasingly important tool for farmers and decision-making authorities everywhere, allowing them to monitor and evaluate the health and status of their crops. Through the accurate and timely information derived from multitemporal satellite data applications like monitoring and managing of crops, forecasting of crop production, measuring and managing water use and crop needs as well as fighting crop insurance fraud can be easily and accurately issued.

The long archive of Landsat program (dated from 1984) allows us to perform robust time series analysis and examine the crop phenology in order to differentiate specific crop types. In addition, recent satellite data from Landsat 8 give us the ability to provide near real-time estimation of crops health, pinpoint signs of crop stress, monitor vegetation growth as well determine actual rates of evaporation.

Crop yield monitoring rely primarily on vegetation indices, such as the Normalized Differential Vegetation Index (NDVI), in order to monitor crop phenology. By examining and analyzing multitemporal values of the NDVI indices it is feasible to monitor the vegetation growth, the fruit/seed status and the maturity of each crop.

During this study I performed a multitemporal NDVI analysis exploiting the products derived from USGS data repository, in the region of Illinois USA, in order to identify the different crop species and study their phenological features in relation to their growth.

From the above chart we can see that the Soybeans during the summer period are on their highest growth, whereas the Fallow cropland are on their lowest growth. It can also be observed that it is feasible to accurately discriminate all the different types of crops in at least 3 different time periods between the time range of 05/2015–12/2015

Automated classification of the different crop types by exploiting the extracted phenological features and machine learning algorithms

The European Space Agency (ESA) Copernicus program adds huge potential with the Sentinel satellites. This family of satellites includes, SAR (Sentinel-1) and multispectral/superspectral (Sentinel 2/3) with global coverage and high spatial and temporal resolution.

If you merge together Sentinel-2 and Landsat-8 multispectral data you can reach almost daily coverage (excluding of course cloudy days) of the monitored area.

As you probably imagine by now, there are a lot of theoretical and technical issues to cover in order to actually process and extract valuable information from images taken 600Km above ground! And I haven’t even mentioned the needed software (or in some cases services).

It can be quite hard and chaotic where to start learning. I would suggest first start “googling”. A nice place to start is, as usual, Wikipedia. There are a lot of things to learn and understand “strange” concepts..

My personal suggestion, is to take some online courses (free or paid, depending on your budget and willingness to put effort in learning).

As it comes a bit natural (feel free to check my CV, LinkedIn profile), I would suggest to start your “journey” with GEO University online courses. Here experienced instructors (including myself) present online self paced courses related to Earth Observation (or Remote Sensing) and GIS.

We include theoretical, practical and even software tutorials online courses!