Plant Science & Agricultural Research
Plants have characteristic signatures in the VNIR (400-1000nm), NIR (900-1700nm), and SWIR (1000-2500nm) spectral regions and show a strong and complex fluorescence pattern. Various hyperspectral modalities can detect and measure the effects of chemical treatments, infection, and other stressors on basic plant physiology, as is often demonstrated on standard model plants such as Arabidopsis, soy beans or corn.
Many phases of agricultural research benefit from the nondestructive optical measurements. Plant phenotyping is a process of identifying individual plants or groups of plants that have favorable traits. Hyperspectral imaging allows high throughput monitoring and measuring in almost all phases of plant improvement research from individual laboratory measurements to greenhouse and to outdoor field studies.
Using hyperspectral imaging
Hyperspectral imaging is one of the widely used modern tools to study plants in different settings:
Plants and plant parts can be explored on a microscopic scale to study plant physiology and plant pathology using hyperspectral fluorescence or hyperspectral transmission and reflection optical arrangements. Middleton Spectral Vision’s MacroPhor™ Hyperspectral Fluorescence Imaging system combine hyperspectral and fluorescence imaging to create a powerful instrument for the analysis of plants.
Different growing conditions, different treatments, phenotyping and other experiments can be conducted in greenhouse settings on whole plants with VNIR, NIR or SWIR hyperspectral cameras, scanning the plants periodically using broadband illumination. This can be done from different directions, under well-controlled environmental conditions during the life cycle of the plant. MSV has different software and hardware components for scanning small to large whole plants. Special designs are available for scanning plants from side-to-side or looking down on the top of the leaves.
Under real outdoor growing conditions in test plots the different plants are studied over the whole growing season using VNIR, NIR and other cameras. The hyperspectral measurements can be performed with cameras moved above the crop with a fixed structure or mounted to a vehicle or mounted to a tower. MSV’s InSight-Field™ system has unique approaches and equipment for monitoring crops from a tower perspective over a long time and analyze the whole dataset as a function of environmental parameters or comparing different plots.
KemoQuant™ is a unique chemometric package utilizing fast algorithms developed at Sandia National Laboratories. Based on a licensing arrangement MSV is able to utilize this package at all levels for laboratory, greenhouse and field measurement analysis.
Laboratory Applications – Including Hyperspectral Fluorescence Imaging
Hyperspectral imaging can be used to discover or measure many different phenomena in plants and plant materials.
Many plant biology experiments in the laboratory rely on various imaging technologies to track growth of small or emerging plants or to study small samples of plant materials in a highly controlled environment. Middleton Spectral Vision brings the powerful capabilities of fluorescence imaging and hyperspectral imaging to the research lab to help scientists discover new insights into plant processes.
Specialized Imaging Systems
Middleton Spectral Vision can design and build bench-top systems with a range of imaging technologies.
Hyperspectral imaging systems combine cameras, spectrographs, lenses and illumination to capture VNIR (400-1000 nm), NIR (950-1700 nm) and SWIR (1000-2500 nm) regions of the spectrum. Components are selected based on research goals to enable the observation and measurement of specific chemical processes or structures in the plant material that would otherwise be invisible.
Our laboratory scanning systems integrate optical components with data collection software and mechanized sample handling to optimize both the effectiveness and efficiency of the data capture process.
- The macroPhor™ System offers both fluorescence and (VNIR) reflectance imaging from a single platform designed specifically for the needs of plant research.
- The Via-Spec Measurement Stages can be configured with different cameras to capture various spectral regions. These systems have several illumination options and provide both reflectance and transmission configurations from the same instrument.
Middleton Spectral Vision also provides the software tools that you need to make sense of your data, including multivariate analysis techniques. The powerful KemoQuant™ software implements Multivariate Curve Resolution for discovering unique interpretable spectral components that utilize the full wavelength range of the hyperspectral data.
Learn more about our solution for discovering unique spectral components from your data at the KemoQuant Software product page.
References related to the MCR analysis approach:
- Pedroso, M. C.; Sinclair, M. B.; Jones, H. D. T.; Haaland, D. M., Hyperspectral Confocal Fluorescence Microscope: A New Look into the Cell. Microscopy Today 2010, 18, (05), 14-18.
- Howland D. T. Jones, David M. Haaland, Michael B. Sinclair, David K. Melgaard, Mark H. Van Benthem, and M. Cristina Pedroso, “Weighting hyperspectral image data for improved multivariate curve resolution results,” Journal of Chemometrics, 2008; 22: 482-490.
Hyperspectral Imaging of Whole Plants
Capturing information that you might otherwise be missing
Hyperspectral imaging can be used to monitor a wide range of plant processes. Middleton Spectral Vision can bring the power of hyperspectral imaging into your greenhouse, customizing the system to meet your research goals.
Middleton Spectral Vision will configure the right combination of cameras, lenses, and illumination sources to maximize the information in the wavelength ranges that best reveal the plant processes and features the researcher is interested in.
Beyond the optical components, MSV can configure and build a complete system with the right level of automation for the situation and software to coordinate the data collection.
Please contact Middleton Spectral Vision to discuss your particular configuration needs.
Making good use of your data
To make the most of your plant image data, MSV has developed several software packages for visualizing your hyperspectral data. The KemoQuant Analysis Software is a particularly powerful analysis package built around a high-speed implementation of Multivariate Curve Resolution (MCR) licensed from Sandia National Labs. MCR analysis is useful to discover unique spectral components that differentiate various plant characteristics, and the software provides tools for storing and reusing spectral components for future analyses.
The images below show an example of a spectral component derived from images of potato plants, then applied to an image of a corn plant. The plots in Figure 1 depict two spectral components resembling Chlorophyll and Carotenoid along with a third component as found by KemoQuant and recognized as resembling Anthocyanin. The hyperspectral images were captured with a VNIR camera (400 – 1000 nm). Figure 2 shows a composite image representing the relative concentrations of each of the three components observed (full spectral signatures) at each pixel. Figure 3 shows the same three spectral components in the image of a corn plant captured with the same camera.
Hyperspectral Imaging Applied to Phenotyping of Crops in the Field
Hyperspectral imaging produces high resolution spatial information along with spectral data to enable precise monitoring of experimental crops, allowing you to visualize what your eyes (or ordinary cameras) can’t see: a plant’s early response to stresses from disease, pests, or drought… responses that help identify resistant phenotypes or genotypes, or to predict yield further in advance of harvest time.
To improve understanding of the related processes in the field, a range of sensors can be combined to collect complementary data, such as weather, soil moisture, or thermal emissions of the canopy.
Middleton Spectral Vision has all the tools and capability needed to deploy, integrate, and automate your data collection. We also have unique algorithms and software tools to help you make the most of your data.