Jose A. Jimenez-Berni1
The use of high throughput plant phenotyping tools is becoming widespread in commercial breeding programs and pre-breeding research. Technologies such as LiDAR, multispectral and infrared thermography are now operated routinely onboard ground and aerial platform, providing imagery and data with unprecedented resolution. This data allows performing digital estimation of key physiological traits such as biomass, plant vigour, nutrition status, water stress or disease susceptibility. In modern gene-assisted selection, having this phenotyping capability is paramount to discover genetic markers and candidate genes of interest in the breeding programs.
However, the use of these phenotyping tools is still insignificant in agronomical applications. Even though agronomists have a long trajectory of adopting remote sensing technologies, such as satellite and aerial imagery or soil moisture monitoring, there is still a degree of frustration when it comes to interpreting the data obtained. Turning pretty pictures into valuable knowledge to assist agronomical decisions remains a challenge. The use of the phenomics paradigm, with a strong focus on the genetic and environmental interactions, together with new sensor technologies and tools that integrate crop simulation models, machine learning and big-data analysis, can alleviate some of these limitations, complementing existing remote sensing applications in making better informed agronomic decisions.
We present here different examples of how phenomics tools can be translated into an agronomic context, providing high throughput physiological measurements that reveal the effects of abiotic and biotic stresses in crops. The current limitations in terms of data processing, sensor calibration and confounding effects will be also discussed.
Plant phenomics, LiDAR, infrared thermography, hyperspectral, decision support tools, machine learning.