To mitigate ongoing insect biodiversity declines, there is a need for efficient yet accurate monitoring methods. The use of traditional catch-based survey methods is constrained both by costs and need for expertise for manual taxonomic identification. Emerging methods, such as eDNA and robotic sorting, have the potential to reduce workload but still require resource-intensive sample collection in the field. Recently, remote sensing methods such as photonic sensors have shown promise for recording large numbers of insect observations. However, accurately determining species composition in collected data remains a challenge. In this study, we investigated the potential of photonic sensors for quantifying species richness of flying insects in the field and at five sites and compared the results with estimates based on conventional Malaise traps. Firstly, we evaluated two unsupervised clustering methods using a library of measured insect signals from 42 known species. Secondly, we correlated estimated number of clusters in data recorded at five sites with species richness assessment of catches from Malaise traps. This study is based on 84,770 library- and 238,584 field individual insect recordings. Our results demonstrate that both clustering methods perform well and reflect estimates obtained by Malaise traps, indicating the potential of automated insect biodiversity monitoring. This offers the possibility of more efficient but still accurate methods for studying insect biodiversity with broader temporal and spatial coverage.