Creating the Connected Farm using sensor and vision data
Feeding the world requires continuous innovation – and as data has become the force multiplier in agriculture – the need for an integrated ecosystem. There is no single solution that solves for creating the Connected Farm because every operation is different and unique in the data it needs to operate. AWS customers have shared their uniqueness of each farm and facility, and their need for an architecture that is flexible and enables an open system of components that can be connected at the software layer. The Connected Farm reference architecture addresses these needs by focusing on ensuring scalability, elasticity, and a responsiveness for each operation’s growing and changing needs.
Early adopters of precision agriculture continue to innovate in the information they collect, often via IoT enabled edge computing devices like soil probes, weather stations, pest traps, and drones. In rural and remote environments, Connected Farm devices can operate in connected, semiconnected, or fully disconnected states. With deployed on the device or AWS IoT Greengrass deployed on a gateway, intermittent connectivity no longer needs to mean lost data. There is no one-size fits all answer to connectivity globally, but there are solutions that work for most workloads. Key considerations when designing a hardware device are latency, on-board compute, connectivity options (Bluetooth LE, Wi-Fi, cellular, satellite), fixed or portable, number of days in a fully disconnected state, and size of data to transmit. Devices capturing soil moisture are sending small data streams that can leverage Amazon Kinesis Data Firehose and Amazon Kinesis Data Streams, whereas a drone is capturing rich visual imagery that is best suited for Amazon Kinesis Video Streams and Amazon Rekognition Video.
Barns & facilities monitoring
Producers and processors are increasingly using IoT enabled edge devices for remote monitoring of facilities through streaming video or auditory capture. Video streaming applications for home security such as Ring allow for a farm manager to see a live feed from a number of facilities. While similar to home security the rural and remote environment creates some inspiring innovations from applications like cattle monitoring, equine health, counting of pigs onto transports to global public private partnerships like FFAR focused on broiler production. Companies like TINE SA have leveraged IoT to enrich an existing data lake to provide context for dairy production and inform breeding programs. Our customers continue to drive innovation daily in this space, and the Connected Farm enables them to use a flexible framework, integrating and ingesting devices and data based on the specific environment and livestock type.
Counting, movement, breeding and birthing are consistent motions in herd management, from agriculture to aquaculture. How this information is collected can be different. Whether it’s submerged cameras and sensors or simply cameras one thing is consistent, it has to withstand the wear and tear of its environment. Animal wearables, like the cattle collar, from New Zealand based Halter Limited have to be designed for cellular low latency connectivity with a minimal footprint that matches the ultra-low power requirements. Minneapolis based Pentair provides water filtration systems equipped with sensors to fish farms and large industrial brewing customers leveraging AWS IoT Core for data collection and AWS IoT Greengrass to account for instability of connectivity. Beewise, a startup from Israel, leverages AWS IoT services to connect its Bee Homes to revolutionize beekeeping and understanding the health of the hive remotely.
Controlled agricultural environments
Recent changes in customer preferences, local supply chains, and regulatory changes have resulted in significant growth in Controlled Agricultural Environments. Startups in the United States, such as Bowery Farms and Parsley, have met consumer demand through highly automated environments. These environments often require a low latency control plane whereby an AWS IoT Greengrass enabled device can leverage on-device models that can be updated from Amazon SageMaker via over-the-air (OTA) updates. IoT edge devices with machine learning models employed on the device can control irrigation or lighting systems, or recognize pests or mold developing within the environment.
The innovation in Agriculture is exciting, and continues to demonstrate that skilled builders and producers not only grow our food, but grow our knowledge. If you would like to get started building your own application, there are 3 great ways to get started. For hardware and devices that have been pre-qualified to work with AWS IoT services please visit the AWS Partner Device Catalog. Interested in building an application that uses cameras as a mechanism for data ingest, a fully deployable AWS CloudFormation Quick Start is available from Onica. Interested in training that will help you build an IoT solution, consider attending an IoT in Agriculture Immersion Day, or IoT Jumpstart.