The aim of this project was to undertake a scoping study to assess the suitability of digital shape and pattern recognition techniques to differentiate different, but related insect species with the overall goal of eventually developing an automatic detection system for Emergency Plant Pests (EPPs) to be deployed into traps. Three pairs of insect species from each of the following Orders, Diptera, Coleoptera and Lepidoptera, were used as model systems. A dataset of over 400GB of images of the model insects was acquired using hyperspectral cameras in the near infrared, visible and UV. The effectiveness of existing state of the art recognition methods was tested on the insect imagery dataset and compared with performance on two published standard datasets. The algorithm achieved an average of approximately 90% recognition for the target model insects investigated using shape and ‘colour’ image descriptors. A subsequent analysis identified the three most significant bands (out of 77 spectra outside the visible range) to use for classification and achieved recognition rates of up to 95% for pseudo-colour data. We have also commenced the development of methods to recover descriptors using any number of bands and including the use of texture as a queue in addition to shape and spectral response. We expect that this additional information, together with the flexibility of selecting any number of bands, will increase significantly the recognition rate of the model insects investigated. A preliminary benchmark and evaluation of the whole system comprising our descriptor, classifier and band selection steps has also been undertaken. The recognition rates obtained in our experiments indicate that the development of an automatic detection system to deploy in a ‘smart trap’ is feasible. The next phase of this work will require expertise in trap design, wireless technology and sensors networks to develop a prototype trap with an auto-reporting semi-automated system for field testing.