Snail Recognition Using YOLO

Abstract

Many species of snails inhabit different areas in the world. Some species have made their way to farmlands and the urban regions, surviving through eating plants and breeding unnoticed making them a cause for concern and a known threat to some crops. A study on snail detection has been previously conducted, but recognizing individual species for their risk has not yet been pursued. This study aims to develop a Tiny-YOLOv4 snail recognition system using a Raspberry Pi. The model focuses on four snail species subject to an input image processed through the system. The outputs show the image with the relevant bounding boxes and labels and notify a user through email for any recognitions. The system produced an overall accuracy of 92%, proving successful in the study's objectives and providing a basis for future literature.