Growing annually at a compounded annual growth rate (CAGR) of 31% from 2019, the commercial value of drones in agriculture is expected to reach $5bn by 2025 according to a report released by Meticulous Research.

Globally, usage of drones in agriculture production is exploding, particularly in North America, Europe and Asia Pacific as farmers contend with high labour costs and a demand for high quality crops that also increase yield. Coupled with increased automation and favorable legislation, this environment has resulted in new farming methods, or what is in reality an entirely new sub-sector of agriculture, precision farming. We now have applications being developed and focused on day-to-day crop monitoring, soil and field analysis, field mapping, as well as crop spraying and health assessment.

Livestock Monitoring

Growth in precision farming has fueled developments across the agriculture sector with new innovations being developed for livestock monitoring, smart greenhouses, irrigation and fish farming, with farmers using a multitude of drone types (fixed-wing and rotor) as well as passive and active sensors to achieve results.

From a livestock perspective, early adopters are operating low cost consumer drones with basic passive RGB camera sensor drones for monitoring and tracking purposes, which produce usable stills and video footage.

The ease of utilising an elevated point of view to locate individual livestock to observe movement, behavior of larger groups, as well as daily infrastructure management related to water, fencing and other assets has improved productivity and reduced labour costs for farmers.

Much of these routine and daily tasks, are undertaken via a simple video stream to a mobile device, while stills could later be studied on a desktop application for analysis, if you’re looking for more detailed data on the health of your pasture for example for signs of overgrazing. While these applications save time and resources for a farmer, they are not smart, but limited, and there is a whole new field of applications being developed for livestock farming, especially around their health and wellbeing through the use of thermal camera imaging.

Thermal Imaging

More than 60% of cattle losses in the US are health related, a statistic which can be reduced and understood by incorporating a thermal imaging solution, AI and machine learning in the future. (2017, North American Meat Institute)

One of the most reliable and traditional indicators of ill health and the presence of an infectious disease is body surface temperature screening. Observed temperature elevations can also be used to detect heat stress in livestock, as well as changes in blood flow patterns which can be attributed to identifying inflammation areas.

The challenge is that a diagnosis requires labour and time, and results in a loss of productivity, as well as increased stress on the animal, and a high probability of transmitting the disease to new livestock herds.

By using lightweight thermal cameras mounted on drones, farmers can examine livestock remotely and without having to restrain the animals; thereby saving time, and reducing stress and the risk of transmission. Typically, thermographic sensors are able to detect infrared radiation from animals in the wavelength of between 8-12 μm. While, the method is effective, it is primarily a strategy best suited to filtering large groups and a confirmed diagnosis will still need to be achieved by a ground level inspection of the animal.

Challenges and Controls in Body Temperature Readings

Despite the successes, elevated temperature readings can be misleading and proper controls should be implemented, which consider differences between skin and body temperature, environmental conditions, and the required thermal resolution and spatial resolution of the thermographic data. Typically, farmers will look for a 1-degree temperature differentiation between animals that are exhibiting fever symptoms and those that are not.

However, skin surface temperatures are naturally affected by absorption of sunlight and environmental factors. Best practice would be to conduct inspections during twilight periods before sunrise and during periods of overcast weather. Scanning of the thorax dorsal or rump are the best locations for thermographic scanning as an actionable data set requires a sufficient number of pixels (10 in a line) across the area of interest. Screening efficiency is also dependent on the size of the sensor as it determines the size of the image footprint that can be used. Choosing a correct altitude to fly at depends on the size of the herd, pixel sensor density, and the focal length of the lens; however, the most effective altitudes range between 20 and 100m.

More applications

While identifying and screening animals efficiently is arguably one of the most effective and promising methods for thermal imaging, utilizing thermal for day to day farm management is can also be more efficient and valuable than basic RGB camera sensor drones. Especially for tasks such as locating lost animals in dense bush, trespassers, pasture health and overgrazing, electric fence line breaks and more.

Future Applications

One of the most groundbreaking current studies, a 3-year research project in the UK set to end in 2021, involves the use of autonomously controlled drones, AI and machine learning. In the pilot project, drones will be eventually used to locate, identify via facial recognition and then measure for vital information like weight, size and physical activity all remotely. Essentially, the team uses one observer drone at a 100m elevation, which identifies the cow and position with a stereo camera and sends the data to three worker drones. These worker drones will then conduct the inspection and transmit the data back to the farmer and his computer station.

As food demand continues to increase and land shrinks, farmers increasingly not only have to mechanise their farms, but automate and digitize their operations to remain competitive and resilient while balancing the need for ecological harmony and diversity. In light of these modern demands, drone-based solutions can provide the most effective tools and platforms, which can be introduced into the food chain and operational toolbox for tomorrow’s farmer.