Participants

Thomas Darron

  • Automatic farmland biodiversity monitoring of farmland birds through AI
  • BSc Hons Zoology

To prove the success of actions to increase biodiversity, implemented by farmers in response to Defra’s Environmental land policy, surveys need to be conducted. In the summer months, bird identification by sight is both expensive and impractical, this research will focus on recording bird song to survey 10 key species on the farm. Building off a bank of annotated data and aided by data collected using Passive Acoustic Monitors (PAM) to record song on site, we will create and train an AI with the purpose of providing an accurate, automatic and passive identification of a list of key species located on site, through these PAMs. Through R studio coding, the AI and data collected will then be used to create a website, where audio files can be uploaded and the site will be able to accurately identify if any of the key species are present, and which ones.