Fotosynthese is hét fundamentele proces waarmee gewassen groeien. Het verbeteren van fotosynthese zou de opbrengst van gewassen kunnen verhogen. Hiervoor is het belangrijk om te weten wanneer fotosynthese de biomassa productie limiteert. Om deze limitaties te detecteren, hebben we technieken nodig om accuraat de biomassa te kunnen vastleggen. Dit project zal nieuwe technieken ontwikkelen om accurater biomassa van gewassen te kunnen vastleggen. Dit wordt gedaan met behulp van nieuwe kunstmatige intelligentie en sensortechnologieën. Met de nieuwe technieken kan fotosynthese met biomassa productie vergeleken worden, en krijgen boeren en veredelaars nieuwe mogelijkheden om de biomassa van hun gewassen te kunnen vastleggen.

The methods will involve deep learning models to predict biomass over time in the field. Subsequently, I will examine the correlations between photosynthesis and biomass, over time. Timeframes during which photosynthesis correlates highly with biomass production will provide starting points for future research to study when photosynthesis limits biomass production.

Photosynthesis is the fundamental process with which crops convert solar energy into biomass. Understanding when photosynthesis limits biomass production, will provide essential insights into how photosynthesis can be improved to increase crop production. Currently, determining when such photosynthetic limitations occur is difficult, due to the dynamic nature of photosynthesis. It will require to compare photosynthetic activity with biomass production, over time. Methods have been developed to track photosynthesis over time, but methods to accurately track biomass over time are still poor. The proposed research aims to develop new deep learning-based methods for determining biomass over time, to study when photosynthesis limits biomass production

The most innovative aspects of my project will be the design of new biomass prediction methods using the state-of-the-art in deep learning models and sensor technology. Deep learning and sensor technology have advanced immensely within the past decade. Considering the widespread use of crop prediction models, bringing the models up to speed with these major scientific advancements will be a significant contribution. I will apply deep learning techniques that have been applied with great success in other scientific domains, but have yet to be used for crop biomass models 31,33,36,43,45–47. In addition, I will evaluate the use of wearable plant sensors for predicting biomass. Wearable plant sensors are an emerging type of small plant sensor that can continuously measure unique plant or environmental parameters due to direct placement on the plant 38,39,51. Due to how fast sensor technology is progressing, wearable plant sensors are rapidly becoming more effective and affordable data collection tools. Considering the widespread use of crop biomass models, it is important to study how these emerging plant sensors can be used to predict biomass more accurately. The main impact of the project will be that the methods to track biomass production over time will provide a crucial first step towards being able to identify when photosynthesis limits biomass production in field crops, a long-term quest in plant science. The more immediate and applied impact of the project will be providing growers and breeders with new methods to determine crop biomass over time.