| Description: |
Insects play an important role within ecosystems, occupying almost all trophic levels and providing crucial services such as pollination and decomposition. In addition, they can directly and negatively impact humans as crop pests, parasites, and disease vectors. Thus, it is essential to understand the spatial and temporal dynamics of insect populations and communities. Various forms of remote sensing, such as LiDAR, have been used to capture data on insects but often are costly and require advanced expertise. The objective of this thesis is to evaluate a cost-effective methodology that balances information loss. This thesis aims to outline the discrimination required within a range of agricultural and conservation questions and to ascertain the preliminary specifications of a machine that can meet these discrimination requirements. The proposed method combines an infrared sensor powered by a Raspberry Pi in conjunction with machine learning for automated classification. |