Agritech Drones: Vegetation Indices And Aerial Imagery
Will it be a feast or famine this harvest? Precision farming and Vegetation Indices enable farmers to gain a better understanding of their produce. Drones, as part of an agritech solution, are opening up new possibilities for crop cultivators everywhere.
Farm field health maps produced from visual and multispectral aerial imagery provide insight to enable agricultural managers to make crucial, well-informed decisions.
Farming is filled with uncertainty, but data helps to reduce this and enables more intelligent planning. As the Department for International Trade points out: “UK innovation is improving one of the fastest-growing areas of agritech – precision technology. It’s already widely used in the UK to improve the efficiency of farm operations, including targeted fertiliser and agrochemical applications.”
A drone fitted with an RGB camera and/or a modified NIR (Near Infrared) camera returns data that enables agriculture professionals to effectively analyse trends in plant health right down to individual stems. This can then be used to modify seeding, cultivation, fertiliser and nutrition planning to optimise results.
Vegetation Indices down on the farm – the science bit
A Vegetation Index (VI) is a spectral calculation of two or more bands of light that assesses vegetative properties.
To determine the density of green on an area of land, the distinct colours (wavelengths) which make up visible and near-infrared sunlight which are reflected by the plants can be captured for analysis. Many different wavelengths make up the spectrum of light. When sunlight strikes objects, certain wavelengths are absorbed and other wavelengths are reflected.
Chlorophyll gives leaves their green colour. This pigment strongly absorbs visible light and then uses it in photosynthesis. This means that measuring the amount of light absorbed by the leaves can be used to make a calculation of plant growth based on individual and overall photosynthesis levels across an area.
On the other hand, the cell structure of leaves strongly reflects near-infrared light. If we assume that the more leaves a plant has the bigger and healthier it is, more of these wavelengths are reflected to give a good indicator of crop condition. This means the farmer can make comparisons of the photosynthetic activity across his or her farm to estimate on yield.
There are lots of VIs – well over 100. Each index is a calculation of a certain combination of sensor-measured reflectance factors (water content, chlorophyll content, pigment, etc.) to reveal particular characteristics of the vegetation being analysed.
Some of the most common VIs are:
- NDVI – Normalised Difference Vegetation Index
- ENDVI – Enhanced Normalized Difference Vegetation Index
- VARI – Visible Atmospherically Resistant Index
- GRVI – Green Ratio Vegetation Index
- GLI – Green leaf index
- NDRE – Normalised Difference Red Edge
- SAVI – Soil-Adjusted Vegetation Index
The use of VIs is not a new thing. Scientists have used remote sensing to monitor fluctuations in vegetation since the 1960s. But the advent of drone technology has opened up the application to farmers across the globe, large or small.
Until recently, data to produce Vegetation Indices and field health mapping was captured from satellites. But to be truly effective, precision agriculture requires high spatial resolution (a large number of pixels per digital image). That is not always possible from space.
At present, satellites may typically return images with a spatial resolution of around 1 metre. To return optimal and extremely precise results – down to individual plant level – a spatial resolution as low as 10 centimetres can be captured by a drone-based camera or sensor. That opens up a whole new field of possibilities for agricultural managers.
A drone has other advantages over a satellite. A drone can fly at any time and doesn’t need to orbit the earth. Drones can generally obtain images under different levels of sunlight and their imagery is not as affected by cloud cover as that from a satellite. Plus, satellite images are typically only received every one or two weeks. A drone can be flown whenever it is required – daily if necessary – to provide near real-time data.
Go for the green – use of Vegetation Indices
ENDVI, NDRE, Ndvi and VARI are three of the most widely adopted and easy to understand Vegetation Indices.
NDVI analyses the red and near-infrared (NIR) bands of imagery taken from crops to evaluate a Vegetation Index value. NDVI detects differences to emphasise the green colour of a healthy plant, it is commonly used as an indicator of chlorophyll content in several different types of crops, including corn and wheat.
ENDVI is a close equivalent to NDVI that uses red, blue and green visible light, as opposed to red light alone.
NDRE is like NDVI but it also utilises the edge of the red band, which penetrates farther down through the crop. It’s sensitive to chlorophyll content, changes in leaf area, and the effect of soil in the background. NDRE is used for determining the relative nitrogen content of crops in the field.
VARI is also a tweaked NDVI but has minimal sensitivity to atmospheric effects, allowing for vegetation to be estimated in a wide variety of environments.
NDVI has been around a lot longer than other Vegetation Indices and so is the most prevalent. It is readily accessible and easy to interpret so it is widely used.
NDVI essentially shows you the amount of photosynthesis occurring in a plant. On an NDVI map, green can mean a plant is healthy and red can mean it is unhealthy. Although of course weeds can also be healthily growing quite happily amongst the desired crop, so adjustments are made to suit.
With NDVI you still need to get your boots dirty and take to the field to validate the data captured from the skies. But as an entry-level indicator, NDVI can be coupled with other indices such as soil and weather measurements to build accurate pictures of crop status.
So that’s quite a lot of technical information already. But a few illustrations of just how VI mapping is developing and being used on farms to produce results demonstrates the potential for results.
Case study USA – regional corn & soybean company
Peterson’s Farm Seeds in North Dakota began trialling UAVs in 2015 to monitor and return data on corn and soybean crops.
Using a DJI Phantom 4, the company operate a multispectral camera along with an RGB camera plus analytical software to determine seed and fertiliser ratios – known as variable rate application. This enables Peterson’s to analyse its crops and modify the prescription (mix of inputs – seed, fertiliser, water etc).
The drone is set on a pre-programmed flight plan and hovers over the desired area to capture multiple images. Once complete, the data is stored on an SD card. This avoids the need for internet connectivity – farmland doesn’t tend to come equipped with WiFi routers mounted on fences just yet. A map can then be generated within minutes in the field itself, so to speak.
The NDVI returns are layered with other data including soil conditions and are plotted with historical yield returns. The software then can then generate a plan using sophisticated algorithms for the next day’s prescription.
The bottom line in North Dakota is that yields have increased and seed, fertiliser and other inputs have become more economical and efficiently used.
Case study – John Deere & big green drone data
John Deere, the giant agricultural machinery manufacturer, is no longer just known for its iconic big green tractors. Through the John Deere Operations Centre, the global corporation is also at the forefront of pioneering precision practices.
In the spring of 2020, PrecisionAg periodical reported on how US agricultural dealer 4 Rivers Equipment had partnered with John Deere to integrate data solutions in everyday farming workflows.
Incorporating DJI drones and Pix4D software, 4Rivers is introducing farmers in the firm’s core regions of Wyoming, New Mexico and Texas to how analysis of Vegetation Indices can enable crop management.
The drone operations combine DJI Phantom 4 Multispectres and Pix4DFields – a recent release which is essentially Pix4DMapper tweaked for mapping agricultural landscapes.
Zonation maps are produced using NDRI and NDRE to enable zoned variable rate technology to reduce over applying seed, fertiliser, nitrogen etc into areas that don’t need it.
Case study Northern Ireland – Grasscheck & the green pastures of NI
“The whole visible universe is but a storehouse of images and signs to which the imagination will give a relative place and value; it is a sort of pasture which the imagination must digest and transform.” Charles Baudelaire – poet
Grasscheck is a project which has been operating in Northern Ireland for over 20 years.
The Grasscheck project brings together farmers, the Department of Agriculture, Environment and Rural Affairs (DAERA), the College of Agriculture, Food and Rural Enterprise (CAFRE) and the Centre for Innovation Excellence in Livestock (CIEL) in Northern Ireland.
The Grasscheck programme of regular monitoring pasture health offers insights for improvement of grass growing and grass quality. This brings financial benefits for livestock farmers based on identifying optimum grazing conditions.
The project has, until recently, been largely land-based but has begun exploring ways in which drones (and other emerging precision farming techniques) can support in the collection of farming data. The trials which completed in December 2019 utilised drones to measure and monitor NDVI reflectance. The results and evaluation report are due to be published in 2020.
Getting started using drones with Vegetation Indices
For those just getting started in crop and pasture analysis, a mid-range drone with an RGB (Red, Green and Blue) sensor is all that is needed to begin. As the human eye sees in RGB, it is our primary sense for interpreting the world around us. That means the agritech sector can easily start assessing crops by using RBG imagery.
This article from Farmers Weekly is an easy-to-understand overview of how to introduce drones into crop management and to begin using NDVI. Do check in with us though as the drone industry, tech developments and regulations have, and are, changing rapidly at the moment.
As we said, Vegetation Indices such as NDVI use the Near-Infrared (NIR) wavelength. That requires an NIR sensor and an agritech software package to analyse fields and yields with a far more powerful insight. If you want to use the NDVI index for true NDVI, you need an NIR camera.
However, if you don’t have an NIR camera, you can still use a Vegetation Index to highlight crop variability. The VARI index (Visible Atmospherically Resistant Index) was developed explicitly for use with the visible spectrum (RGB imagery).
What drones for Vegetation Index analysis?
A DJI Phantom 4 RTK fitted with a suitable sensor or the DJI Phantom 4 Multispectral will open up a whole new field of possibilities.
The Parrot Bluegrass was developed specifically with agritech in mind. The rugged quadcopter can map multiple fields and generate NVDI maps via Pix4D software.
Where next for you?
As the UK’s leading provider of complete drone solutions, Coptrz can discuss how you can explore revolutionising your crop and pasture monitoring. Get ahead of the rapidly accelerating agritech curve to begin realising the efficiency, yield and cost-saving benefits utilising drones can bring to the agricultural sector.
The post Agritech Drones: Vegetation Indices And Aerial Imagery appeared first on COPTRZ.
