Search

There is a need to classify eating behaviors of grazing animals to better understand time budgets and intake levels of extensively managed livestock. The objective of this work was to compare spectral and motion-based sensing for classification of animal grazing versus standing behaviors within extensive production systems. The sensor was an open-source setup with a TTGO-T-Beam microprocessor (Shenzhen Xin Yuan Electronic Technology Co, China), soldered with an AS7265x Spectral Triad (SparkFun Electronics, Niwot, CO) and generic 9 degree of freedom inertial motion unit. The sensors were attached to a mesh face mask and placed on 5 horses for 10 minutes of data collection. Within the data collection period, animals were allowed to graze for 5 minutes, and required to stand for 5 minutes. The order of grazing and standing was randomized among animals. Data were collected by the microprocessor at 100 hertz and averaged over a 20 second period. Average data were transmitted via long range radio (LoRa) to a base station. Data were matched by timestamp to ground truth observations of animal standing vs grazing behaviors. A random forest algorithm was used to explore predicting animal grazing versus standing behaviors based on motion data (average pitch, yaw, and roll of animals over the experimental period) individually, based on the top features selected from the spectral data individually, or based on all collected data together. The model using the motion data only returned an out of bag estimated error rate of 26.5%. The model using the top features selected from the spectral data yielded a much lower out of bag estimated error rate (8.82%), which was not improved by leveraging all data together. By comparing the spectral data with GPS coordinates attached to each observation, we were able to clearly distinguish the spatial area used for grazing compared with that used for standing. One potential impact of this preliminary exploration is in low-cost, low-power, more precise behavioral classification for grazing livestock. Because the spectral sensed data generated more reliable ability to distinguish grazing and standing behaviors, we may be able to rely on lower frequency observations, or lower data acquisition rates to achieve acceptable classification accuracies. Similarly, given the success of NIR in analyzing feed composition values, there is likelihood that spectral sensing could be used not only to identify grazing, but to character forage quality grazed. [ABSTRACT FROM AUTHOR]

The continued success of livestock production in grazing systems depends on the ability to manage lands sustainably. Non-point-source emissions, particularly defecation/urination events (DUE), are a major cause of soil degradation, runoff into water sources, and nutrient management issues. Monitoring of these emissions from livestock through wearable sensing could support more precision management, but is complicated by the need for individualized sensors, variable signal strength across topographies, and unreliable data transmission in remote locations. The objective of this study was to explore opportunities to use CO2 sensing as an indicator of DUE from cattle on pasture using data obtained from an open-source sensor. The use of tail sensors was implemented to monitor the DUE of horses at various times of the day, ranging from 8 am to 5 pm. The tail sensors were developed using a neoprene base attached to velcro straps that were placed around the tailhead. The sensors themselves were equipped with a SparkFun ESP32 Thing Plus microprocessor connected to a SparkFun CO2 sensor and a lithium ion rechargeable battery that was charged via micro USB (SparkFun Electronics, Niwot, CO). Sensors were placed on 4 horses for 6 hours each collection period and data collection occurred over 9 days. Horses were observed in real time and DUE were recorded at the time of occurrence. Data were collected from the sensors via a Dragino LoRa gateway and a mobile hotspot for wireless connection. Data analysis relied on visual data exploration and targeted assessment of associations using random forest regression, support vector machines, and extreme gradient boosting. The visual data exploration revealed that DUE occurred in much lower frequency than non-events, and so random oversampling and undersampling were used for each classification approach to attempt to improve accuracy. Although CO2 peaks often corresponded to DUE, they also frequently occurred in the absence of DUE. All classification algorithms showed poor accuracies (0.50 to 0.51), which were only marginally improved by over- (< 0.51) and undersampling (< 0.69). This preliminary assessment revealed considerable noise in sensing CO2 emissions in production settings, which may preclude usefulness in DUE sensing. [ABSTRACT FROM AUTHOR]