Variable-time cameras with image recognition for inexpensive, large-scale monitoring of plant pollination events
Plant-pollinator interactions (PPI) are a major driver of plant evolution, and are studied by research groups all over the planet. Traditionally, PPI are tediously documented through direct observation, which is now increasingly replaced with digital cameras. However, commercial cameras lack the technical flexibility required for such experiments. Here, we aim at developing an inexpensive, weatherproof Raspberry Pi-based camera with image-recognition guided frame-rate adjustment to monitor both long and short-time PPI. Study system is the South African daisy Gorteria diffusa, which has orange flowers with black spots, and is pollinated by the bee-fly Megapalpus capensis: Female flies look for nectar, and mistake spots for conspecifics. Males, however, mistake spots for females, and copulate with them. Population-wide documentation of long-time preferences of females, and short-time copulatory behaviour of males may answer long-standing evolutionary questions. An affordable camera that can handle these scenarios would be of interest for the entire PPI research community.