Terahertz Imaging: A Diagnostic Technology for Prevention of Grass Seed Infestation

One of the most significant problems the Australian Sheep and Lamb industry faces today is Grass Seed Infestation (GSI), which occurs when seeds accumulate in the sheep's fleece and penetrate the skin, causing infection. Meat & Livestock Australia estimates that the yearly losses caused due to GSI are around AUD$47.5M (in Australia alone). Recently, we have demonstrated that terahertz spectroscopy and imaging can be utilised for early detection of GSI. This is possible because terahertz waves can penetrate through sheep wool and have the appropriate wavelength for identifying the seed. Moreover, terahertz waves have non-invasive and non-ionizing properties and are ideal for non-contact and standoff detection. To implement the technology, it is necessary to have a fast raster scan and enhance the signal quality. In this project, we aim to implement an image processing technique (e.g., background subtraction) to reduce the average time required for raster scans and enhance the signal-to-noise ratio. We will develop low-cost machine learning-enabled terahertz signal-processing solutions for a sheep auto drafter.