|BEEMAN, KIM - Engineering Design|
|Neel, James - Jim|
Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/14/2011
Publication Date: 2/16/2011
Citation: Clapham, W.M., Fedders, J.M., Beeman, K., Neel, J.P. 2011. Acoustic monitoring system to quantify ingestive behavior of free-grazing cattle. Computers and Electronics in Agriculture. DOI: 10.1016/j.compag.2011.01/009.
Interpretive Summary: The ability to estimate intake of grazing livestock has eluded researchers for decades. Such estimates would facilitate a real understanding of forage/livestock systems that to date rely on limited data and visual observation. We developed a system composed of hardware that records grazing events acoustically and software that can analyze the acoustic record. Acoustic methods have been used previously, however, their widespread use is hindered by a restricted sound frequency response range that limited the potential to automate file processing. We utilized an expanded frequency range that permitted us to automate detection and measurement of bite events. Bite number and bite energy measurements are produced that can be related to animal intake. Our system is capable of analyzing sound files of any length limited solely by hard disk storage capacity. These developments, for the first time, offer the potential to describe livestock grazing behavior and forage intake over extended time periods and allow us to understand the true interactions among grazing livestock, their performance efficiency and forage management. This knowledge will be useful for improving livestock selection for performance on forages and for developing forage and management systems that minimize both production risks and negative impacts to our forage resources and environment.
Technical Abstract: Methods to estimate intake in grazing livestock include using markers, visual observation, mechanical sensors that respond to jaw movement and acoustic recording. In most of the acoustic monitoring studies, the microphone is inverted on the forehead of the grazing livestock and the skull is utilized as a sounding board. The acoustic signals captured with this method are low frequency and confound prehensive bites with mastication. Low signal resolution makes it difficult or impossible to automate the classification of ingestive events. We developed an acoustic monitoring and signal processing system composed of hardware and software components that successfully recorded, detected and quantified the sound energy of prehensive bites from free-ranging cattle grazing triticale pasture. Our acoustic monitoring system utilized a high quality audio flash recorder and a microphone placed adjacent to the jaw of the grazing animal. The microphone placement maximized detection of the signal energy produced with forage shearing during a prehensive bite. We assumed that the forage sheared was equivalent to intake. A 600 Hz high pass filter was successfully used to remove low frequency wind noise from recordings prior to processing. Spectrographic analysis showed that the signal from prehensive bites was rich in frequencies above 17 kHz while processing sounds (chewing, mastication, etc.) were characterized by lower frequencies. High signal resolution was critical to the success of developing parameters and software to automate detection and quantification of prehensive sound wave energy captured during grazing. The software processed files at up to 10 times real time. Logistic regression in conjunction with Monte Carlo simulation was used to adjust for errors in bite detection. The system documented an average of over 9000 prehensive bites animal-1 during one 4.5 h grazing session. As the grazing session progressed, bite rate (bites min-1) declined by an average of 75 percent and mean sound energy bite-1 declined by 33 percent. The software system can process sound wave files of any length, limited only by hard disk space. This acoustic monitoring and processing system represents a significant improvement in our ability to quantify grazing over extended time periods, and improves our ability to gain a mechanistic understanding of forage/grazing systems.