Satellite Data in Agriculture
In one of our previous blogs — “Kisan Talks — Knowing the Indian Farmer”, we talked about how a large proportion of modern day farmers have started utilising technology and particularly smartphones in making informed decisions about their cultivation practices. To cater to this progressive mindset of farmers, a challenging aspect is providing accurate data and more so providing ways to interpret that data to produce observations and insights, which were originally indiscernible to the human eye. The solution to this challenge has come by the way of satellites. One of the earliest satellites launched by NASA for agricultural use was Landsat 1 in 1972. Since then around 300 satellites have been launched providing high, medium and low resolution spatial imagery. Satellites provide images at different frequency, resolutions and spectral bands. Information for some of the most widely used ones are shown in the following table:
Frequency of a satellite refers to the number of days in which fresh images are obtained from the satellite. Spatial resolution represents the amount of detail with which a farm can be represented in an image. For example, 30m resolution means that each pixel stands for a 30m x 30m area on the ground.
Lets understand how satellite data can be helpful for a typical farmer :
- Progressive Farm Analysis — Because a satellite has a designated frequency as explained above, it will circle and cover an individual piece of land multiple times in a crop season. This allows farmers to get a sequence of snapshots of the farm’s performance over the months.
- Accurate understanding — Satellite imagery provides a more detailed and more accurate picture of any given field. It highlights things across many different criteria, from crop type, crop health, fertiliser requirement, irrigation requirements and other problem areas.
- Better targeted scouting efforts — With images being taken and analysed in regular intervals, farmers are able to detect trouble areas, and optimise the scouting efforts in a much more targeted way.
- Predict agricultural output — Constant data driven feedback on trouble areas and real time farm performance can also help in predicting end of season yield and anticipate, mitigate the effects of food shortages and famines well in advance.
Other than these, there are many more advantages of spatial data such as crop area estimation, deriving basic soil information, cropping system studies, experimental crop insurance, etc.
Transforming satellite data into useful insights — Vegetative Indices (NDVI & EVI)
Satellite imagery in its raw form may not be of very much use unless it is processed into valuable information. This processed information in agricultural scenarios can be called vegetation indices. A satellite image consists of several bands like Red, Green, Blue, Infrared etc. and vegetation index is developed by performing a transformation on two or more bands. Historically there have been many multi spectral vegetative indices, but the two most widely used ones are — Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI). Before diving deep into each of them, let’s first understand the science behind this process.
Green plants absorb sunlight and use this solar radiation as a source of energy to perform photosynthesis. Other than absorption, plants also reflect or re-emit some part of this sunlight into specific frequencies. Plants reflect most of the light in the near-infrared bands as opposed to soil, water or snow. Optical sensors of satellites capture this reflected radiation across various bands which is then transformed by mathematical formulas to generate vegetative indices.
Understanding NDVI :
The Normalized Difference Vegetation Index is the most widely used index since its introduction in the 1970s. It is a simple indicator of photosynthetically active biomass or, in layman’s terms, a calculation of vegetation health. NDVI helps to differentiate vegetation from other types of land cover and determine the current health of a farm. Mathematically, NDVI can be expressed by the given formula
Where NIR represents the amount of light reflected by a plant in the Near Infrared region and Red represents the same in the visible red region. Since NDVI is a ratio of two bands, the resultant value always ranges from -1 to 1. Negative values correspond to areas with water surfaces, man made structures, rocks, clouds, snow. Barren soil usually falls within 0.1- 0.2 range while plants will always have positive values between 0.2 and 1. Healthy, dense vegetation canopy should have NDVI value above 0.5, and sparse vegetation will most likely fall within 0.2 to 0.5. However, this is just a rule of thumb and in actuality the NDVI varies from crop to crop, farm to farm, and is highly influenced by crop stage and farming practices.
The images below represent sample NDVI plots for a few crop farms. Notice the colour difference in each of the examples.
The color in an NDVI plot is in accordance with the general rule of thumb explained above. A general palette and the guidelines for vegetation type can also be seen below:
Many would argue that if we already had NDVI as a solution, why do we need EVI? Is it a better vegetative index than NDVI? If yes, then why is it not more widely known than NDVI?
EVI stands for Enhanced vegetation index. It is an optimised index aimed to work better at high biomass areas. While NDVI is highly chlorophyll sensitive (NDVI saturates after a certain crop stage) and depends on canopy type, canopy density, time of the day satellite image was taken, EVI is more responsive, flexible and hence is used to correct NDVI. EVI used additional wavelengths of the spectrum — NIR, Red, Blue to correct the inaccuracies in NDVI. Despite the benefits, NDVI still supersedes EVI because EVI is very complex to calculate in practicality. The additional wavelengths in EVI coupled with canopy background adjustment, aerosol correction (Mathematics not dealt into because of complexities) are very convoluted for a typical usage and thus, NDVI is the preferred industry standard since it is also able to perform the job at hand.
GramworkX Kisan App integrates NDVI — Our solution
GramworkX recently launched a new version of Kisan App in which we provided our own implementation of NDVI as a service to farmers at an affordable rate. The whole solution is easy to understand and intuitive for farmers. A four step summary of the solution is given below:
- Farmers draw their farm coordinates on the app by simple touch and draw polygons. The land acreage and other basic details are shown to the farmer.
- The latest NDVI image is shown for the drawn plot in the app
- The crop progress, crop health across different parts of the farm are visualized for the plot
- Real time feedback and advisory is provided to assist the user in his decision making.
The sequence of the steps can be better understood by the following images given below:
How will this NDVI solution benefit farmers?
As explained earlier, NDVI is a useful tool for farmers that can help them use a data-driven feedback loop to determine how to optimize their farm upto the highest extent.
- Through observing, analyzing, and interpreting the regular snapshots, farmers can monitor the real time growth of the crops, separate high health vs low health segments of the farm, detect areas of concern within the field faster and establish normal growing crop patterns for specific areas of the farm.
- We provide both generic recommendations as well as crop phenology based recommendations. These recommendations are constantly validated on ground and hence are highly accurate.
- Farmers can see historic NDVI since the time their subscription started on the app itself. This removes the need of any physical log book entries, eradicating any scope of human error.
- The recommendations are easy to understand and we connect manually with farmers, field officers as well to provide necessary support.
Future of Satellite Farming and GramworkX’s vision with NDVI service
While there are multiple benefits to this technology, there are also delimits of applying remote sensing and vegetation indices in agriculture. Some of the key challenges are below:
- Cloud coverage can play a pivotal role in the quality of information available in a satellite image. High cloud coverage can obscure the image, rendering it less effective to give farm predictions.
- Remote sensing is a fairly expensive method of analysis especially when measuring or analysing smaller areas
- It is a data intensive process. Since satellites are gathering so much data regularly, special training as well as highly specialised infrastructure is required to process the data at high rates.
Although satellite based farming has some challenges of its own, they don’t impede the pace at which remote sensing is changing the face of precision agriculture. Many progressive farmers are coming forward everyday to try this new technology and see where they can belong in this ecosystem. For them, from an on-ground perspective it is very difficult to handle a lot of information quickly especially when you have a large number of farmers working under you. Remote sensing data does provide many advantages over conventional methods, particularly in terms of timely decision making mechanisms, big picture understanding and cost optimizations.
This is where GramworkX comes as a facilitator for promoting satellite farming among farmers and equipping them with tools to enhance their agricultural practices. We have already started gaining traction for our NDVI service and in future, we plan to introduce more vegetative indices in our app. Moreover, we will also line up other remote sensing based solutions like soil moisture estimation, flood mapping, NPK mapping and many more in nearby future.
Want to know more about us and our NDVI solution? — Connect with us at firstname.lastname@example.org
Contributors — Yash Agrawal