Optical High Throughput Phenotyping of Crops and Produce
Accurately measuring a plant’s physical characteristics is critical to understanding. However, the nature of the knowledge and its application is highly stakeholder-dependent. For instance, in plant breeding programs, the objectives are to quantify value characteristics (e.g., size, shape, weight, yield, storage capability, count, taste, processing characteristics) within the context of the plant’s genetics. This enables specific genes to be identified that are more preferential than others or may convey desirable traits (disease or pest resistance) to the final crop, thus aiding breeder selection. Conversely, in a commercial setting, the same sensing modalities can be leveraged at different points in the stakeholder’s workflow to enable the grading of produce into different value-added categories. In this case, the necessary scanning speed – or throughput – needs to be approximately two orders of magnitude faster than for plant breeding (e.g., 1-2 million sweet potatoes scanned per day). In this talk, I will overview different imaging and optical sensing technologies and machine learning models we have developed for various stakeholders and crops (corn, sweetpotato, and hyperaccumulators). Key to their development is the integration of low-cost and high-technology readiness level (TRL) sensing (often a cell phone) with appropriate measurement protocols (usually no more than three steps) and algorithms to enable the desired measurement in a short timeframe. Additionally, more advanced sensing techniques will be discussed within the context of new sensor development, whose TRL is lower but will be advanced in the context of commercial high throughput phenotyping.
Michael Kudenov
North Carolina State University on November 1, 2022 at 1:00 PM in MRC 454
Join Zoom Webinar
Dr. Kudenov completed his BS degree in Electrical Engineering at the University of Alaska Fairbanks in Fairbanks, AK in 2005. Upon graduation, his personal interest in astronomy and photography led him to obtain his Ph.D. in Optical Sciences at The University of Arizona (UA) in Tucson, AZ in 2009. Following his Ph.D., he remained as an Assistant Research Professor at the UA until departing for North Carolina State University in 2012. Research performed at the UA included visible and infrared imaging polarimetry, spectroscopy, 3D profilometry, interferometry, active learning, and lens design.
His current research is focused on developing novel imaging systems, interferometers, detectors, and anisotropic materials related to polarization and spectral sensing, for wavelengths spanning ultraviolet through the thermal infrared. He is particularly interested in developing novel anisotropic materials and detector technologies that better enable snapshot systems, which are capable of maximizing the spatial, spectral, and/or polarimetric information contained within a single image. Applications include biomedical imaging, remote sensing, food safety, 3D Imaging, and atmospheric monitoring.
ASSIST is developing leading-edge systems for high-value applications such as healthcare and IoT by integrating fundamental advances in energy harvesting, low-power electronics, and sensors with a focus on usability and actionable data.
