Page 26 - AI Ready – Analysis Towards a Standardized Readiness Framework

AI Ready – Analysis Towards a Standardized Readiness Framework

4�3�3 Smart Shrimp Farm Aquaculture

This use case [39] applied cameras with night vision and sensors with underwater capabilities
to capture shrimps and other water parameters such as pH value, turbidity, and oxygen
concentration. Since the images are captured underwater, pre-processing tools such as
robotflow [55] are used to enhance the quality. A special model YOLO is deployed for unique
object counting. In addition, to predict the length of the shrimps, the ArUco marker [54] is used
for measurements. This use case referred to available standards such as BAP 1000, ASC shrimp
standard, Global GAP Aquaculture standards, ISO 9001: 2015, and ISO 22000: 2018.

4�3�4 Soil Moisture Testing

This use case [40] used soil sensors with a limited measurement range to detect soil parameters
and water usage. The data collected undergoes the process of conversion with the protocol of
Message Queuing Telemetry Transport (MQTT) [56], quantization/aggregation/range-checking,
network, and model, and finally reaches users. Random Forest (RF) and Multivariate Adaptive
Regression Splines (MARS) as classification models are applied. Transmission Control Protocol/
Internet Protocol (TCP/IP) with error handling capabilities are used. The analysis is achieved
using edge data, edge board and sensors, and edge storage. Given that the sensors and
facilities are deployed outside, ruggedness is considered under the standard of IP65. The edge
data processing board is open-sourced.

4�3�5 IoT-based Crop Monitoring

In this use case [31], edge data such as pH value, Nitrogen, Phosphorus, Potassium (NPK) levels,
electric conductivity of the soil, weather parameters, and leaf wetness are captured. Sensors use
solar panels to harvest energy. By combining the data, it is possible to not only manage pests
and diseases but also plan pesticide usage and irrigation schedules.

4�3�6 Agriculture: Crop Monitoring and Planning

In this use case, all sensors are connected to IoT to collect temperature, moisture, crop health,
yield, minerals, soil health, and carbon level data. Based on the information from the greenhouse
station, camera, and weather station, the actuators controlled remotely such as sprinklers could
be used to manage irrigation and potentially crop planning mapping between crop, fertilization,
and pesticide usage. Storage security is an important consideration in this use case. OneM2M
[22] standard is applied in the use case.

4�3�7 Smart Irrigation

In modern agriculture, the integration of advanced technologies involves a diverse array of
actors and systems working together to enhance efficiency and yield optimization. Agricultural
farmers use both traditional methods and modern technology for irrigation, pesticide usage,
and farm management. Sensors are used to monitor temperature, humidity, soil moisture, fluid
levels, and mineral content in the roots, feeding data into low latency, high throughput networks
such as edge networks. AI and ML systems collect this data and infer actionable insights aligning
with policies, which are then executed by actuators such as automated irrigation systems,
tractors, and dispensers for pesticides and fertilizers. Backend cloud storage supports this
ecosystem, while dashboards provide farmers with information. Local conditions, such as water

19