Crop Health Assessment Using Satellite Imagery And Remote Sensing
How Fasal is bringing 360-degree visibility into crop care by augmenting early detection with powerful prediction.
Fasal at its core is an innovation-driven company where we all observe and identify problems, ask a lot of questions, discuss, debate, brainstorm, and come up with solutions to help our farmers. Our approach to innovation has always been incremental and outcome-driven and we believe in bringing visible and sustainable change to the ecosystem we operate in.
When we started Fasal a few years ago, we were determined to change the way in which farmers make day to day decisions while growing a crop. We understood weather, farm, soil, crop, crop life cycle, disease, pest, irrigation, markets, and everything in between to build a deep tech solution for crop care that helps farmers make data-driven and informed decisions to minimize the cost of cultivation, improve quality of crop and crop yield. We also saved about 3 billion liters of freshwater that is used in agriculture and we brought the concept of FASAL WATER CREDIT ™, one of its kind, to reward farmers who save fresh water and practice sustainability.
Also read —
Last week, we introduced our flagship product — a plug and play IoT sensing device — ‘Fasal Kranti’ to the world that we were working on for more than a year now to make the IoT sensing system affordable, available, and accessible to every farmer in the world. With ‘Fasal Kranti’, any farmer, anywhere can use an IoT sensing system at his farm to practice precision farming and sustainable agriculture without worrying about its installation and maintenance. The Better India wrote a nice encouraging story on this —
Today, I am going to talk about a new initiative that we are working on for some time now under Fasal Infinity Lab to further improve our digital crop care technology. This new initiative at Fasal is the beginning of a new gateway to take precision farming to the next level. We believe that detection and prediction are two faces of a coin and each is equally important when it comes to precision and accuracy. At Fasal, we are continuously working on our precision, predictions, and accuracy but now we are expanding our wings in early detection. This is very important from a farming perspective. We are super serious about precision farming and we constantly build technology to achieve the same.
As the title says we are working on crop health assessment using satellite imagery that involves various steps like collecting huge satellite data, selecting data sources and its types, studying and analyzing health properties and parameters of crops, and then deciding how to build a robust, scalable and user acceptable technology that can be used and integrated inside Fasal app.
A good question to ponder upon is — why is a precision agriculture pioneer looking for satellite health assessment of crops? As I shared earlier our innovations are always outcome-driven. When we built our prediction technology to assist farmers in protecting crops from damage due to disease and pests attack by alerting them in advance, we were able to save up to 50% in spray cost. Because preventive sprays are many times cheaper than reactive sprays and they do not cause any harm to crop quality. With satellite imagery and remote sensing, we want to go further and save more money by informing farmers that they only need to spray in a fraction of their farms and not the entire farm. And that is possible with the help of augmenting early detection with prediction.
Remote sensing is a type of satellite information that collects electromagnetic (EM) radiation emitted and reflected from the planetary, atmospheric, and marine ecosystems of the globe in order to identify and track a region’s physical characteristics without physical contact. Usually, this data collection technique includes aircraft-based and satellite-based sensor systems that are either labeled as passive or active sensors.
Passive sensors respond to external stimuli, gathering radiation that is reflected or emitted by an object or the surrounding space.
Active sensors use internal stimuli to collect data, emitting energy in order to scan objects and areas whereupon a sensor measures the energy reflected from the target.
Types of remote sensing
Drone Imagery
- Best way to capture a high-resolution image.
- Costly and inefficient to implement.
Satellite Imagery
- Large area cover
- Easy accessibility and availability of data
- Inefficient during cloud
Remote Sensing Data Types
LIDAR Remote Sensing
LiDAR is a remote sensing device that is involved. An active system implies that to quantify items on the ground, the device itself produces energy — in this case, light. Light is released from a quickly fired beam in a LiDAR device. You can picture light strobing rapidly from the source of laser light. This wave moves to the surface and returns from objects like structures and branches of trees. The reflected light energy then falls to where it is registered on the LiDAR sensor.
Hyperspectral Imagery
Recent developments in remote sensing and geographic data have paved the way for hyperspectral sensor growth. Hyperspectral remote sensing, also described as imaging spectroscopy, is a fairly new concept that researchers and scientists are currently investigating in terms of the analysis and tracking of rocks, planetary flora, and objects and backgrounds created by humans.
Hyperspectral remote sensing combines imaging and spectroscopy in a single system which often includes large data sets and requires new processing methods. Hyperspectral data sets are generally composed of about 100 to 200 spectral bands of relatively narrow bandwidths (5–10 nm), whereas multispectral data sets are usually composed of about 5 to 10 bands of relatively large bandwidths (70–400 nm).
Hyperspectral Imagery from Satellites
As a first step, we use satellite data sources to get Hyperspectral images of our fields. And since we know how our data looks like and what are different band images that we are getting, we figure out the health of the plant by looking at the optical properties of different plants, and thus we set few parameters to identify the healthy and unhealthy region, cloudy or non-cloudy, burnt or polluted, etc. Let’s see what all parameters help in determining the healthy and unhealthy regions.
NDVI
- The NDVI (Normalized difference vegetation index) is an indicator of a plant’s health based on how a plant reflects different light waves.
- The cell structure of a plant reflects the near-infrared waves. So a healthy plant, the one with a lot of chlorophyll and good cell structure, actively absorbs red light and reflects near-infrared.
What the NDVI can tell you at different stages of the season
At the beginning of the season, the NDVI index helps to understand how the plant has survived through the previous season.
- If the NDVI is lower than 0.15, most probably all the plants died in this part of the field. Typically, these figures correspond to plowed soil without any vegetation.
- 0.15−0.2 is also a low value. This may indicate that plants started wintering in the early phenological phase, before tillering.
- 0.2−0.3 is a relatively good value. Probably, the plants entered the tillering stage and have resumed vegetation.
- 0.3−0.5 is of good value. Nevertheless, you should keep in mind that high NDVI values can indicate that plants wintered at a late phenological stage. If the satellite image was taken before the resuming of vegetation, then it is necessary to analyze the zone after the resuming of the vegetation also.
- Above 0.5 is an abnormal value for the post-wintering period. It is better to check this field zone yourself.
In the middle of the season, the NDVI index helps to understand how plants grow and develop. If the index values are medium to high (0.5−0.85), most likely there are no major issues at this part of the field. If the index is low, probably there are specific issues, like a lack of moisture or nutrients. It is better to check this part of the field yourself.
Challenges in NDVI
- NDVI loses sensitivity when a plant reaches a certain development threshold.
- The NDVI index also depends on the weather: if clouds are overhanging the satellite image will be unclear.
Overcoming NDVI Challenges
NDVI is one of the parameters which helps us to understand the plant health but conditions may change with the place, season, calamity, etc., and determining the health of the plant becomes more and more tricky. Many studies have been done to understand the plant’s health in different conditions and different varieties of plants. And these studies gave us a few more parameters as listed below. These parameters help us to differentiate the health of a plant with more variety of results.
With the help of remote sensing, satellite imagery, and advanced analytics, we are going to provide pixel by pixel health analysis of farm and crop with respect to disease, pest, irrigation, yield estimate, etc. We are bringing 360-degree visibility into crop care by augmenting early detection with powerful prediction and this is just the beginning. We are currently running a private beta of this advanced solution and it will be available to our farmers very soon on the Fasal app. Stay tuned!
