UNMANNED AERIAL SYSTEM (UAS) FOR BIRD DAMAGE CONTROL IN WINE GRAPES

Santosh Bhusal

Pest bird damage in an agricultural setting is a long-standing problem. Past studies showed that most of the currently adopted bird deterrence measures do not meet growers' expectations regarding efficacy, reliability, and cost. Therefore there is a need for a more practical approach for bird deterrence. In this study, an Unmanned Aerial System (UAS) was investigated as an alternative approach to protect wine grapes from the pest bird. The UAS-based bird deterrence techniques investigated in the past needed human operators to locate birds' flocks and fly UASs to desired trajectories. This study integrated artificial intelligence (AI) technology with the UAS towards developing a fully autonomous bird detection and deterrence system. Also, the performance of a UAS-based bird deterrence system to minimize the pest bird activities and the fruit damage was studied. Results showed that the automated bird counting algorithms achieve an F1-score of 85% for counting birds. It was also found that super-resolution and deep learning techniques significantly increased bird detection (> 90%) accuracy. Also, this study showed a significant decrease (p-value = 0.001) in the number of birds entering into a field in the presence of the UAS, leading to decreased wine grape damage when UAS was used (p-value < 0.0001). The use of an on-board auditory scare device and erratic flight movement of the UAS proved to be more effective in deterring pest birds. The utilization of the graphical processing unit for vision algorithms significantly improved the frame rate (frames per second > 30) for continuous surveillance of incoming birds, enabling opportunities towards automated bird deterrence. Finally, two bird deterrence missions that can be dynamically created and executed in a fully automated manner were investigated and evaluated towards deterring incoming birds and birds that have already settled in the field (~90% success rate). These research findings are expected to provide a basis for developing a commercially viable autonomous system for deterring pest birds from vineyards. When commercially adopted, this study's outcome is expected to make a huge, positive impact on minimizing production loss in vineyards and improve the long-term sustainability of the wine grapes industry.

UNMANNED AERIAL SYSTEM (UAS) FOR BIRD DAMAGE CONTROL IN WINE GRAPES

Files and links (1)

Abstract

Metrics

Details