Postdoctoral Research Assistant Institute for Plant Sciences / Purdue University
Disclosure: Disclosure information not submitted.
Hyperspectral Imaging (HSI) is a prominent methodology for plant phenotyping for both field operations and controlled environments. As a way to boost new overall agricultural technology, Purdue University developed the Controlled Environment Phenotyping Facility (CEPF) to enable researchers and to facilitate the demand for automated High Throughput Phenotyping (HTP). The CEPF has three different imaging/sensing systems: RGB cameras, an HSI set, and an X-ray Computed Tomography Scanner. The goal of this report is to demonstrate the progress made by CEPF regarding the HSI system capabilities. Two hyperspectral cameras, of which one covers the Visible and Near-Infrared (VNIR) waveband, and another that includes the Short-Wave Infrared (SWIR) waveband, are integrated to collect plants’ light reflectance between 400 to 2500 nanometers. Different spectroscopic methods, such as vegetation indices, having been leveraged in analyzing spectral data. Pixel level colormaps (2.05 mm per pixel), are provided to facilitate better visualization of the processed data. The CEPF data pipeline can process plant images in batches, resulting in an average result for each instance, which is easier to manipulate and analyze. This simpler processed data, for example, would include the mean value for several vegetation indices (including NDVI), some morphological features, and the reflectance value for the plant at each light band, as previously mentioned. Purdue researchers are using these results for different scenarios. For example, detecting stress on wheat due to soil salinity, correlating the accumulation of heavy metals on kale and basil plants, and showing signs of nutrient deficiency on corn and soybean plants. However, there is still space for improvement and advancement, especially regarding nutrients predictive models and the interaction between VNIR and SWIR data. Nevertheless, the CEPF has shown to be a powerful platform for researchers who need access to HTP sensing capabilities and processed, information reach data set.