An Intelligent Vision based Pest Detection System Using RCNN based Deep Learning Mechanism

The necessity to control pests and diseases by biological means using computer and internet technology instead of pesticides to protect crops is a primary objective of this work. Research in agriculture is mainly focused towards growth in the productivity and food quality at reduced expenditure and with increased profit. The use of vision based technology for pest monitoring has increased a huge importance in agriculture sector in recent time. A strong demand now also arises for non-chemical control methods for pests or diseases in many countries. However no automatic methods are available which precisely and periodically detect the pests on plants. In fact, in production conditions, greenhouse staff periodically observes plants and search for pests. This manual method is time consuming and not efficient. In our work deep learning based pest detection has been experimented and tried for deployment in real farming field. For this purpose RCNN based detection mechanism using Deep Learning based segmentation has been implemented and tested. The experiment has shown that RCNN based approach shows significant improvement over common pest detection mechanism.