Monitoring monarch predation using Raspberry Pis and computer vision

Insects often experience low survival rates throughout their development. In particular, monarch butterflies (Danaus plexippus) experience high rates of death as eggs and caterpillars. For species of management concern, like monarchs, understanding cause of death and related factors can be used to create more effective management strategies. However, arthropod death data is difficult to acquire through human observation alone, which may bias behavior or even entirely miss predation events. Therefore, we aimed to design a method to monitor monarch survival using video surveillance. We used weatherproof, portable camera systems built from Raspberry Pi elements to record up to 24 hours of consecutive video. Then, we developed a computer vision model that can identify predation events to assist in data processing. Together, these strategies provide a feasible method for collecting and analyzing detailed predation data that can be used to inform management recommendations for monarch populations.