Optimizing the Visibility of Agricultural activities on the Farm using a sound analytics platform

Kanthraj, Manoj (2023) Optimizing the Visibility of Agricultural activities on the Farm using a sound analytics platform. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (200kB) | Preview

Abstract

Norway's agricultural sector is significant culturally and economically, and it thrives despite severe challenges. The production of dairy products is at the top, followed by those of sheep and pigs, with salmon aquaculture also on the rise. Governmentally approved strict criteria for biodiversity, emissions reduction, and sustainability (including livestock welfare) are used to cultivate crops including grains, potatoes, and organic vegetables. Sound-based analytics are used by cutting-edge AI technologies for monitoring farm machinery. The method uses supervised classification to pick out relevant audio data from the noise and then uses one-class classification to identify unique content. Over 20 distinct farming activities may be identified using multi-class classification, with assistance from farm boundary detection. With the Urban Sound dataset, we can see that Random Forest achieves 90% accuracy in just 22.83 seconds, which is significantly faster than XGBoost's time of 326.69 seconds. XGBoost can identify both human voices and garbage in 0.81 seconds, but Random Forest takes 1.27 seconds to do it with just 85% accuracy. XGBoost can identify agricultural tasks with 97% accuracy in about 0.04 seconds, while Decision Trees can only get to 95% accuracy. These changes show how technology is improving life in Norway's rural areas.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Heeney, Sean UNSPECIFIED
Uncontrolled Keywords:	Agriculture; Machine Learning; Fast Fourier Transform; Mel Frequency Cepstral Coefficients; XGBoost; Ensemble; Space Complexity
Subjects:	H Social Sciences > HD Industries. Land use. Labor > Specific Industries > Agriculture Industry Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Tamara Malone
Date Deposited:	20 Aug 2024 16:56
Last Modified:	20 Aug 2024 16:56
URI:	https://norma.ncirl.ie/id/eprint/7054

Actions (login required)

View Item