Toward Sustainable Electronic Beehive Monitoring: Algorithms for Omnidirectional Bee Counting from Images and Harmonic Analysis of Buzzing Signals

摘要

Two algorithms are presented for omnidirectionalbee counting from images to estimate forager traffic levels. Anaudio processing algorithm is also presented for digitizing beebuzzing signals with harmonic intervals into A440 piano notesequences. The three algorithms were tested on samplescollected through four beehive monitoring devices deployed atdifferent apiaries in Northern Utah over extended periods oftime. On a sample of 378 images from a deployed beehivemonitoring device, the first algorithm for omnidirectional beecounting achieved an accuracy of 73 percent. The secondalgorithm for omnidirectional bee counting achieved anaccuracy of 80.5 percent on a sample of 1005 images with greenpads and an accuracy of 85.5 percent on a sample of 776images with white pads. The note range detected by theproposed audio processing algorithm on a sample of 3421.52MB of wav data contained the first four octaves, with thelowest note being A0 and the highest note being F#4.Experimental results indicate that computer vision and audioanalysis will play increasingly more significant roles insustainable electronic beehive monitoring devices used byprofessional and amateur apiarists.