Implementation and Evaluation of a High Throughput Phenotyping System to Resolve Morphometric Traits under Multiple Environmental Conditions

Abstract:

Challenges involving the measurement of plant traits over large areas can considerably limit functional genomics studies. Several methods for plant phenotyping using a variety of technologies have been proposed, including those that benefit from remote sensing approaches that include the measurement of spectral responses (e.g., based on spectroradiometers) or 3D reconstructions of plant canopies or other structures based on light detection and ranging or structure from motion. Some of these technologies are limited when representing small plant structures, such as thin blades, due to their limited ability to spatially resolve those structures. In addition, technologies can be strongly affected by changes in structure due to environmental conditions (e.g., effects of wind on results from structure from motion). Here, we report on the design, implementation and evaluation of a mobile high-throughput phenotyping system and phenotyping workflow based on a multi-sensor platform that can be used to capture and represent hard-to-resolve morphometric traits. The platform, incorporating a high-density depth scanner and cameras (RGB, multispectral and thermal) was used for data acquisition over plots planted with graminoids and forbs. Ground measurements based on traditional methods and using the implemented phenotyping platform were conducted simultaneously to data acquisition by a multi-sensor unmanned aerial system for data integration and comparisons. 3D models of plant canopy were created and morphometric descriptors were extracted and evaluated to assess system performance. This work presents the integration of the high throughput phenotyping system, including platform and sensor requirements and control, data acquisition and processing, and analyses of results for target/vegetation characterization and to support information extraction and plant morphometric analyses