Precision Livestock Farming Glossary & Technical Terms

A

Accelerometer: A sensor that measures non-gravitational acceleration. In PLF, 3D accelerometers are mounted on collars, ear tags, or leg bands to log movement along three axes, classifying behaviors like grazing, resting, and ruminating.
Acoustic Monitoring: The use of microphones and audio processors to record and analyze animal sounds. Applications include detecting coughing in swine houses or distress chirps in poultry flocks.
AMS (Automated Milking System): Also known as robotic milking. A system where cows voluntarily enter a milking box to be milked by an automated robotic arm, using sensors for sanitation, teat detection, and milk quality analysis.
AST (Audio Spectrogram Transformer): A deep learning model designed for audio classification. It processes spectrograms as images to classify complex vocal patterns, achieving high accuracy in poultry distress detection.

BCS (Body Condition Scoring): A method for evaluating the fat cover of livestock. In PLF, this is automated using 3D depth cameras to measure back fat and tailhead structure, scoring animals to guide feeding regimes.
Bolus (Rumen Bolus): An electronic device swallowed by ruminants that settles in the reticulum. It contains sensors to monitor temperature (fever alerts), pH (acidosis detection), and movement for several years.
BRD (Bovine Respiratory Disease): A serious respiratory infection in cattle. Early detection is achieved using rumen boluses or accelerometers, which flag drops in activity and temperature spikes days before clinical signs appear.

CCC (Concordance Correlation Coefficient): A statistical measure used to evaluate the agreement between two methods (e.g. sensor readings vs. visual observation). A CCC close to 1 indicates high accuracy.
CNN (Convolutional Neural Network): A class of deep neural networks commonly applied to analyze visual or audio data. In PLF, CNNs analyze spectrogram images for cough detection or camera feeds for animal tracking.
Computer Vision: A field of AI that trains computers to interpret visual data. PLF uses computer vision for count tracking, pose estimation, and automated weight prediction.

Digital Twin: A virtual replica of a physical asset, process, or animal. In PLF, digital twins simulate an animal's metabolic state using continuous sensor feeds to optimize health and nutrition.

Edge AI / Computing: Processing data locally on the sensor device rather than transmitting raw files to the cloud. This saves bandwidth and battery in remote rangelands.
EfficientNet: A highly optimized convolutional neural network architecture. In PLF, it is used for fast, accurate image classification tasks, such as identifying diseases from fecal droppings.
Estrus Detection: Identifying when a female animal is in heat. Accelerated movement patterns are flagged by wearable sensors, indicating the optimal breeding window.

Federated Learning: A machine learning technique that trains algorithms across multiple local devices or farms without exchanging raw data, preserving privacy and security.

Heat stress: A physiological condition caused by high temperature and humidity. Measured in PLF using the Temperature-Humidity Index (THI), it guides cooling interventions in barns.

IRT (Infrared Thermography): Using thermal cameras to measure surface heat. In PLF, IRT scans animal eyes, ears, or udders to detect fever or inflammation (e.g. mastitis) non-invasively.
IoT (Internet of Things): A network of physical devices that exchange data over the internet. PLF barns deploy IoT sensors to monitor relative humidity, gas levels, and water flow automatically.

LoRaWAN: A Low-Power Wide-Area Network protocol designed to connect battery-powered devices. It transmits sensor data over long ranges (up to 15km) in remote agricultural areas.
LSTM (Long Short-Term Memory): A recurrent neural network architecture designed to process sequences of data. In PLF, LSTMs analyze time-series accelerometer data to classify daily animal behavior.

MFCC (Mel-Frequency Cepstral Coefficients): Features extracted from audio signals that represent the power spectrum. In PLF, MFCCs are input into neural networks to identify animal coughs and vocalizations.

NB-IoT (Narrowband IoT): A low-power cellular network standard for IoT devices. It is used in areas with cellular coverage to transmit animal location data without private gateways.

PLF (Precision Livestock Farming): The use of advanced technology (sensors, AI, IoT) to monitor and manage livestock individually in real time, improving health, welfare, and productivity.

Random Forest: A machine learning algorithm that uses an ensemble of decision trees. It is applied in PLF to classify animal health status from multi-sensor data arrays.
RFID (Radio Frequency Identification): The use of electromagnetic fields to identify animals via transponders (ear tags or microchips). RFID forms the basis of individual animal tracking in robotic feeders.
Rumination: The process of regurgitating, rechewing, and reswallowing food. Wearable sensors log rumination time as a sensitive health biomarker in ruminants.

SCC (Somatic Cell Count): The number of somatic cells per milliliter of milk. Elevated SCC indicates immune response to mammary infection, tracked in real time by milking robots.

THI (Temperature-Humidity Index): A value combining air temperature and relative humidity. In PLF, THI calculators assess heat stress risk levels and recommend cooling protocols.
TinyML: A field of study enabling machine learning models to run on resource-constrained microcontrollers, processing raw sensor data locally on the animal.

Virtual Fencing: A containment system that uses GPS collars to keep animals inside defined boundaries without physical fences, warning them with audio tones and low-energy pulses.

XAI (Explainable Artificial Intelligence): AI models designed to output explanations for their decisions (e.g. SHAP values). This makes sensor health alerts transparent and interpretable for farmers.

YOLO (You Only Look Once): A family of real-time object detection models. YOLO models are used in PLF camera feeds to count animals, track individuals, and identify behavior anomalies.