The reticuloruminal function is central to the digestive efficiency in ruminants. For cattle, collar- and ear tag-based accelerometer monitors have been developed to assess the time spent ruminating on an individual animal. Cattle that are ill feed less and so ruminate less, thus, the estimation of the time spent ruminating provides insights into the health of individual animals. pH boluses directly provide information on the reticuloruminal function within the rumen and extended (three hours or more) periods during which the ruminal pH value remains below 5.6 is an indicator that dysfunction and poor welfare are likely. Accelerometers, incorporated into the pH boluses, have been used to indicate changes in behaviour patterns (high/low activity), utilised to detect the onset of oestrus. The paper demonstrates for the first time that by processing the reticuloruminal motion, it is possible to recover rumination periods. Reticuloruminal motion energy and the time between reticuloruminal contractions are used as inputs to a Support Vector Machine (SVM) to identify rumination periods with an overall accuracy of 86.1%, corroborated by neck mounted rumination collars.

Monitoring cattle behaviour is core to the early detection of health and welfare issues and to optimise the fertility of large herds. Accelerometer-based sensor systems that provide activity profiles are now used extensively on commercial farms and have evolved to identify behaviours such as the time spent ruminating and eating at an individual animal level. Acquiring this information at scale is central to informing on-farm management decisions. The paper presents the development of a Convolutional Neural Network (CNN) that classifies cattle behavioural states (‘rumination’, ‘eating’ and ‘other’) using data generated from neck-mounted accelerometer collars. During three farm trials in the United Kingdom (Easter Howgate Farm, Edinburgh, UK), 18 steers were monitored to provide raw acceleration measurements, with ground truth data provided by muzzle-mounted pressure sensor halters. A range of neural network architectures are explored and rigorous hyper-parameter searches are performed to optimise the network. The computational complexity and memory footprint of CNN models are not readily compatible with deployment on low-power processors which are both memory and energy constrained. Thus, progressive reductions of the CNN were executed with minimal loss of performance in order to address the practical implementation challenges, defining the trade-off between model performance versus computation complexity and memory footprint to permit deployment on micro-controller architectures. The proposed methodology achieves a compression of 14.30 compared to the unpruned architecture but is nevertheless able to accurately classify cattle behaviours with an overall $F1$ score of 0.82 for both FP32 and FP16 precision while achieving a reasonable battery lifetime in excess of 5.7 years.

Cloud cover remains a significant limitation to a broad range of applications relying on optical remote sensing imagery, including crop identification/yield prediction, climate monitoring, and land cover classification. A common approach to cloud removal treats the problem as an inpainting task and imputes optical data in the cloud-affected regions employing either mosaicing historical data or making use of sensing modalities not impacted by cloud obstructions, such as SAR. Recently, deep learning approaches have been explored in these applications; however, the majority of reported solutions rely on external learning practices, i.e., models trained on fixed datasets. Although these models perform well within the context of a particular dataset, a significant risk of spatial and temporal overfitting exists when applied in different locations or at different times. Here, cloud removal was implemented within an internal learning regime through an inpainting technique based on the deep image prior. The approach was evaluated on both a synthetic dataset with an exact ground truth, as well as real samples. The ability to inpaint the cloud-affected regions for varying weather conditions across a whole year with no prior training was demonstrated, and the performance of the approach was characterised.

This paper reports on the use of estimates of individual animal feed intake (made using time spent feeding measurements) to predict the Feed Conversion Ratio (FCR), a measure of the amount of feed consumed to produce 1 kg of body mass, for an individual animal. Reported research to date has evaluated the ability of statistical methods to predict daily feed intake based on measurements of time spent feeding measured using electronic feeding systems. The study collated data of the time spent eating for 80 beef animals over a 56-day period as the basis for the prediction of feed intake. A Support Vector Regression (SVR) model was trained to predict feed intake and the performance of the approach was quantified. Here, feed intake predictions are used to estimate individual FCR and use this information to categorise animals into three groups based on the estimated Feed Conversion Ratio value. Results provide evidence of the feasibility of utilising the ‘time spent eating’ data to estimate feed intake and in turn Feed Conversion Ratio (FCR), the latter providing insights that guide farmer decisions on the optimisation of production costs.

Collar-based activity sensors are in common use as a means of detecting oestrus to optimise farm fertility and, hence, productivity. Recently, the same acceleration-derived signals have been processed to detect the time spent ruminating and eating, which, together, give an insight into animal welfare. Here, the use of neck-mounted accelerometers to provide a quantifiable measure of the time period that an individual animal exhibits signs of heat stress is reported. Heat stress has a significant impact on both animal welfare and productivity. Cattle studied during elevated temperatures were found to exhibit signs of exaggerated breathing motions, an indicator of heat stress, for 8 h on average per day, exceeding the time that cattle spend feeding and is similar to daily rumination times. No similar cases were recorded in the cooler conditions of a Scottish winter. The approach offers a cost-effective measure of heat stress and a potential tool to quantify its impact more generally.

Monitoring and classification of dairy cattle behaviours is essential for optimising milk yields. Early detection of illness, days before the critical conditions occur, together with automatic detection of the onset of oestrus cycles is crucial for obviating prolonged cattle treatments and improving the pregnancy rates. Accelerometer-based sensor systems are becoming increasingly popular, as they are automatically providing information about key cattle behaviours such as the level of restlessness and the time spent ruminating and eating, proxy measurements that indicate the onset of heat events and overall welfare, at an individual animal level. This paper reports on an approach to the development of algorithms that classify key cattle states based on a systematic dimensionality reduction process through two feature selection techniques. These are based on Mutual Information and Backward Feature Elimination and applied on knowledge-specific and generic time-series extracted from raw accelerometer data. The extracted features are then used to train classification models based on a Hidden Markov Model, Linear Discriminant Analysis and Partial Least Squares Discriminant Analysis. The proposed feature engineering methodology permits model deployment within the computing and memory restrictions imposed by operational settings. The models were based on measurement data from 18 steers, each animal equipped with an accelerometer-based neck-mounted collar and muzzle-mounted halter, the latter providing the truthing data. A total of 42 time-series features were initially extracted and the trade-off between model performance, computational complexity and memory footprint was explored. Results show that the classification model that best balances performance and computation complexity is based on Linear Discriminant Analysis using features selected through Backward Feature Elimination. The final model requires 1.83 ± 1.00 ms to perform feature extraction with 0.05 ± 0.01 ms for inference with an overall balanced accuracy of 0.83.