Skip to content

When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.


Introducing the Plant Phenomics & Precision Agriculture Collection

We are very excited to launch our new Collection on Plant Phenomics and Precision Agriculture. Our Guest Editors- Malia Gehan, Guillaume Lobet and Sierra Young- have curated a diverse group of research articles selected from the pool of submissions we received in response to our call for papers. Here we highlight some of the articles included in the Collection at launch- but more will be added in time, so do keep checking back!

We start our tour underground. Measurement of belowground plant traits requires specialised technological solutions to overcome the challenges posed by the physical properties and variability of substrates. Adu and colleagues measured root architecture traits of young pot-grown cassava plants, with the aim of identifying appropriate yield predictors for mature field-grown plants [1]. They found strong evidence for the feasibility of their phenomic approach as a means for exploring yield components.

Other belowground structures were studied by Kudenov and colleagues [2]. They demonstrated the potential of interactance spectroscopy for probing the internal tissues of sweet potatoes in order to identify defects that affect culinary value and storage quality. The technique was effective to depths of approximately 5 mm, and could represent a significant improvement over current destructive methods for defect monitoring.

Necrotic tissue RGB cross-sectional images from

Machine learning has found a wide range of applications in plant phenomics and precision agriculture, including image-based plant classification. However, all such applications require appropriate training data for the particular crop and task involved. To address this bottleneck, Beck and colleagues developed an embedded robotic system that can automatically generate and label plant images for machine learning applications [3]. The system presented in their Collection article could dramatically increase the efficiency of machine learning-based phenomics.

Full view of the EAGL-I system from

Image segmentation is another key area of research in plant phenomics, underpinning many approaches to analysis of crop growth and development. Li and colleagues described a new end-to-end segmentation system based on convolutional neural networks to support high-throughput phenotyping of maize plants [4]. This system allows for identification of individual plant shoots within a field and extraction of their phenotypes, which could help facilitate progress in varietal selection.

Overview of the image acquisition system from

Another segmentation approach based on near-infrared spectroscopy is presented by Colorado and colleagues in their Collection article [5]. They demonstrate its application for the measurement of aboveground biomass in rice crops. When compared with an existing k-means-based method, the new technique showed a 13% improvement in the strength of the correlation between image-derived and actual biomass.

The UAV system from

Crop mixtures, an important component of many agroecological land-use practices, represent a challenge for image analysis. Focussing on a model grass-legume mixed pasture system, Ball and colleagues used RGB imaging to predict growth and the level of interspecies facilitation [6]. They found that high-throughput phenotyping provides a valuable tool for exploring interactions between species in such complex agricultural environments.

Pot-level separation of plant biomass from

Herbivory is a major challenge to crop growth, and methods for measuring herbivores’ effects on plant health status are therefore vital for screening for resilient crop varieties. The results of a study by Horgan and colleagues suggest that changes in plant reflectance properties caused by planthopper feeding could provide the basis for high-throughput approaches to screening for resistance to sap-sucking insects in cereals [7]. They discuss how such measurements could be integrated in bulk phenotyping tests.

The nutrients available to a crop are another key determinant of performance. Understanding the nutritional composition of any organic soil amendments (e.g. manure, compost) is therefore of critical importance. In their Collection article, Towett and colleagues present a machine learning-based method for quantifying nutrients in organic soil amendments using X-ray fluorescence and mid-infrared spectroscopy [8]. They suggest that portable spectrometers coupled with machine learning algorithms could even be developed as a low-cost tool for use by smallholder farmers.

Several papers in the Collection focus on the potential of unmanned aerial vehicles (UAVs) for phenotyping and surveillance of crop status. In their study, Grüner and colleagues tested whether UAV-based multispectral and textural monitoring could predict aboveground biomass and nitrogen fixation in grass-legume mixtures [9]. Their results showed great promise for this approach, with strong evidence for the importance of including textural information in prediction models.

Orthomosaic of the experimental field from

Meanwhile, Cao and colleagues used UAV hyperspectral remote sensing to model the chlorophyll content of rice canopies [10]. Chlorophyll content is an important indicator of growth status and photosynthetic capacity, particularly under different environmental conditions. The authors used an inversion model based on machine learning, which produced encouraging results supporting the use of such technologies for chlorophyll monitoring.

The UAV hyperspectral imaging system from

Automated monitoring of plant growth in space and time poses many technical difficulties. In recent years, three-dimensional point clouds derived from laser scanners and depth cameras have been used for measuring plant structure, but tracking points over time during plant growth is challenging. In their paper, Chebrolu and colleagues explore the feasibility of a non-rigid registration approach, finding that it could successfully model plant growth in four dimensions and prove useful in automated trait analysis [11].

4D registration of a point cloud pair for maize and tomato from

Last but not least in the first batch of papers included in the Collection, Jacques and colleagues highlight the issue of variability in the properties of materials commonly used in phenomic applications [12]. They show that the production location of a gelling agent used in Arabidopsis seedling phenotyping has a significant effect on the results obtained, affecting the comparability of studies performed using product from different sources.

More articles will be added to the Collection over the coming months, so please do check back for updates!


  1. Adu MO, Asare PA, Yawson DO, Nyarko MA, Abdul Razak A, Kusi AK, et al. (2020) The search for yield predictors for mature field-grown plants from juvenile pot-grown cassava (Manihot esculenta Crantz). PLoS ONE 15(5): e0232595.
  2. Kudenov MW, Scarboro CG, Altaqui A, Boyette M, Yencho GC, Williams CM (2021) Internal defect scanning of sweetpotatoes using interactance spectroscopy. PLoS ONE 16(2): e0246872.
  3. Beck MA, Liu C-Y, Bidinosti CP, Henry CJ, Godee CM, Ajmani M (2020) An embedded system for the automated generation of labeled plant images to enable machine learning applications in agriculture. PLoS ONE 15(12): e0243923.
  4. Li Y, Wen W, Guo X, Yu Z, Gu S, Yan H, et al. (2021) High-throughput phenotyping analysis of maize at the seedling stage using end-to-end segmentation network. PLoS ONE 16(1): e0241528.
  5. Ball KR, Power SA, Brien C, Woodin S, Jewell N, Berger B, et al. (2020) High-throughput, image-based phenotyping reveals nutrient-dependent growth facilitation in a grass-legume mixture. PLoS ONE 15(10): e0239673.
  6. Colorado JD, Calderon F, Mendez D, Petro E, Rojas JP, Correa ES, et al. (2020) A novel NIR-image segmentation method for the precise estimation of above-ground biomass in rice crops. PLoS ONE 15(10): e0239591.
  7. Horgan FG, Jauregui A, Peñalver Cruz A, Crisol Martínez E, Bernal CC (2020) Changes in reflectance of rice seedlings during planthopper feeding as detected by digital camera: Potential applications for high-throughput phenotyping. PLoS ONE 15(8): e0238173.
  8. Towett EK, Drake LB, Acquah GE, Haefele SM, McGrath SP, Shepherd KD (2020) Comprehensive nutrient analysis in agricultural organic amendments through non-destructive assays using machine learning. PLoS ONE 15(12): e0242821.
  9. Grüner E, Wachendorf M, Astor T (2020) The potential of UAV-borne spectral and textural information for predicting aboveground biomass and N fixation in legume-grass mixtures. PLoS ONE 15(6): e0234703.
  10. Cao Y, Jiang K, Wu J, Yu F, Du W, Xu T (2020) Inversion modeling of japonica rice canopy chlorophyll content with UAV hyperspectral remote sensing. PLoS ONE 15(9): e0238530.
  11. Chebrolu N, Magistri F, Läbe T, Stachniss C (2021) Registration of spatio-temporal point clouds of plants for phenotyping. PLoS ONE 16(2): e0247243.
  12. Jacques CN, Hulbert AK, Westenskow S, Neff MM (2020) Production location of the gelling agent Phytagel has a significant impact on Arabidopsis thaliana seedling phenotypic analysis. PLoS ONE 15(5): e0228515.

Leave a Reply

Your email address will not be published. Required fields are marked *

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Related Posts
Back to top