Air Scout thermal images show crops in a new light

Within many fields the variation between areas can be stark, from changes in soil to the impact of shade, with these differences highlighted when extreme weather causes crops to survive outside their natural growing climate. But delving into the reasons why one area yielded 11t/ha at harvest and another only 6t/ha takes more than a few crop walks per year, and more data than is currently in your agronomist’s armoury.

Real-time crop performance data

Brian Sutton, the founder of US company Air Scout, had enough of speculating on the causes of yield fluctuations and wanted to get to the bottom of why it happened year on year. He’s not out to replace his agronomist – far from it. But rather than paying them to walk a small percentage of his farm, he wants to arm these professionals with real-time crop performance data, making them an even more vital part of the farming team.

Getting airborne

Aged 13, Mr Sutton flew his first plane (not solo) and was immediately besotted. He describes his light aircraft as the world’s best off-road vehicle. The bonus now is that he is able to survey his farmland at the same time as getting his flying fix. The 200ha family farm is an hour south of Chicago and when times were tough during the 1980s farm crisis, Mr Sutton went off to study for his pilot’s licence and an aviation mechanic’s qualification.

_{Text continues underneath image}

Brian Sutton with his Piper Super Cub. – Photos: Farmers Weekly

After flying for Trans World Airlines for a number of years, in the mid-1990s he came back to the farm along with his brother when his father suffered ill health, and he was determined to combine his 2 passions in order to benefit the farm. After making some instant changes to the farming practices, such as moving to a no-till system and increasing the size of the farm to its current 600ha, he was sure that imagery gathered from aircraft-mounted cameras would be able to inform agronomic decisions and help boost yields.

NDVI

His first foray into determining if this concept had legs was with near-infrared (NIR) cameras, which produce a normalised difference vegetation index (NDVI) image. This image highlights where in the field there is live green vegetation: healthy plants will soak up the sun and reflect near-infrared light, barren areas will do the opposite.

Once this was downloaded Mr Sutton wanted to see if the image showed anything different to what could be seen with the naked eye. But despite numerous lens, camera and processor changes the NDVI images were telling him little more than the view from the cockpit. In 2010 he bought his first thermal imaging camera, and it was the maps gathered from one of his first flights that highlighted just how accurate these images could be.

Compaction

After processing the image (see image 1), he showed the outcome to his brother who struggled to identify the reason why a centre section of the field, which had a crop of soybeans growing in it, had a very obvious sign of stress as it was much hotter than adjacent areas. A local agronomist identified the cause to be compaction and the penny dropped: the cows had fed on the corn stalks during a wet spell the previous autumn and the soil was suffering from the weight of the animals.

_{Text continues underneath image}

The marked area shows compaction from cows.

That harvest, Mr Sutton combined the field in 2 sections, with the compacted half coming in a resounding seven bushels/acre (0.48t/ha) less than the rest of the field. This experience has led Mr Sutton to now allow cows on to the corn fields only when the soil is completely dry.

Thermal images could help spot disease outbreaks

Further trials on his home farm confirmed that thermal images from flights could help spot disease outbreaks before the agronomists did, broken drainage channels after heavy rainfall, and worrying levels of plant stress. To make the most of his thermal shots, Mr Sutton had to be able to tell a story of the growing season via these images, which meant multiple flights during the year and images that can be compared with previous years’ shots. With this data providing a deep insight into how the soil in each area of the field was performing, Air Scout was born.

How does Air Scout work?

The first flight takes place around drilling when there is bare soil on show. Then up to 14 different images are taken from separate flights throughout the growing season, on average every 14-21 days in the early season and every 7-10 days in the height of the growing period. Mr Sutton uses the bare soil image as a crucial reference point. Here differences in texture and type will be shown, along with broken drainage channels and soggy spots. This image is constantly referred back to when the crop is growing to relate the growing patterns and colours to the soil underneath.

It’s critical to see the crop at these growth stages, as this is when disease is obvious

“There is no predetermined date for the flights as every year has varying weather and drilling dates, so growth stages can vary by weeks from year to year,” Mr Sutton says. “It’s critical to see the crop at these growth stages, as this is when disease is obvious and application timing decisions can be easily made.”

It usually takes a full year of images to build a picture of the farm, but from the initial images and the combine yield map, Air Scout can break fields into zones. There are usually by crop health, disease pressures or variations in the soil type. Once Mr Sutton or one of his pilots has flown over the farm and captured images of each field, these are processed and uploaded to the web platform. The farmer receives a text to say the new images are available.

Air Scout Processor

The Air Scout processor is a key part of the process. It automatically rotates the images so north is up, groups the new images alongside previous ones of the same field, and finally applies the colour palette by setting a central temperature across the field. This is done by using a 10deg variance in both directions to highlight any hotter or cooler areas. Hotter parts of the field mean the plants are under stress and this can commonly be the onset of a disease, pest attack or signs of drought.

_{Text continues underneath image}

A thermal image alongside the equivalent bare soil image.

Cooler areas are generally the best parts of the field as these highlight good evapotranspiration – the transfer of water to the atmosphere by transpiration from plants and evaporation from soil. The processor is essentially what turns the thermal image data into usable information. Farmers can view these images, usually with 24 hours, download them or produce maps for the agronomist with areas marked that need closer inspection.

Fertiliser should be applied at different rates at each growth stage, rather than using a blanket approach

By grouping areas of the field, it is easy to see where fertiliser and money should be prioritised because yield potential remains. Conversely, underperforming areas could see inputs reduced because there is little chance of significantly improving yields come harvest. “Fertiliser should be applied at different rates at each growth stage, rather than using a blanket approach’” explains Mr Sutton.

“We have proved that feeding the bigger, stronger plants with meals that they can finish is better than giving smaller plants too much food which is inevitably wasted. Not only does this hopefully return stronger yields, but can have huge cost savings if there are fluctuations in crop growth, along with striving to be more environmentally friendly.” The information can also be used for drilling as the maps will highlight higher-yielding areas which can accommodate a higher seed rate.

Variable seed rates

This information forms a generic shape file which can be exported to any drill controller and used to apply variable seed rates automatically across the farm. Once a year’s worth of images has been collected, it’s easy for the farmer to see where the yield was lost on the crop and at what time of year it was affected. It is then up to the farmer to stop it repeating.

Other image sources

While drones are becoming increasingly popular for farmers to survey their land, mounting a quality thermal imaging camera, which weighs around 13kg, to a small UAV is not feasible. One light aircraft can cover 400ha a day, which would require many drones with multiple operators and scaled down equipment providing images that wouldn’t be as accurate.

“Also, it’s best to get the entire field in one shot as then every pixel is relative to its neighbour,” Mr Sutton says. “In the time it takes a drone to cover the field, the entire area’s temperature changes.” That is not to say drones don’t have a role. Mr Sutton uses them for detailed shots of areas that have been highlighted by the thermal cameras.

_{Text continues underneath image}

A thermal image (left) shows striking similarities with the combine yield map (right).

For instance, if there is a hot spot in one area of the field, the drone can be programmed from the web platform to visit this exact area of the field by dropping a pin on the location. Once there it can takes pictures and beam live footage back to the user for instant analysis, this is a great way to catch disease early, without walking to the exact spot.

Another way of sourcing an image would be using satellites, but as thermal images aren’t available and the ability to only see on clear days means the image quality is poor, it can’t provide the detail needed for accurate decision making.

Cost

Mr Sutton stresses that Air Scout will not replace paying your agronomist or crop walking, but he says it will make better use of both of them by pinpointing areas that need attention or getting soils in the right shape before the crop is drilled. He reckons that by arming farmers with usable data from across the farm, he can make it easy to spot where disease may break out or if there is particular area that is slow to emerge.

The most cost-efficient way to use Air Scout is by a co-operative of neighbouring farmers working together, so that when the plane is up it can take pictures of everyone’s land, rather than doing separate flights for each farm.

70 cents per acre

By working together, the cost can drop to as low as 70 cents per acre (£1.35/ha), rather than the $ 7 per acre (£13.50/ha) it would cost alone. This covers flights for the entire season’s worth of images. Mr Sutton reckons it would not be too difficult to get an Air Scout setup working in the UK. The critical requirement is a source of self-employed pilots with aircraft, willing to take on the franchise and the varying workload. The raw images could be sent back to America for processing, rather than setting up expensive hardware here, with processed data uploaded to the same web platform.