English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Early-stage plant density is an essential trait that determines the fate of agenotype under given environmental conditions and management practices. The useof RGB images taken from UAVs may replace traditional visual counting in fieldswith improved throughput, accuracy and access to plant localization. However,high-resolution (HR) images are required to detect small plants present atearly stages. This study explores the impact of image ground sampling distance(GSD) on the performances of maize plant detection at 3-5 leaves stage usingFaster-RCNN. Data collected at HR (GSD=0.3cm) over 6 contrasted sites were usedfor model training. Two additional sites with images acquired both at high andlow (GSD=0.6cm) resolution were used for model evaluation. Results show thatFaster-RCNN achieved very good plant detection and counting (rRMSE=0.08)performances when native HR images are used both for training and validation.Similarly, good performances were observed (rRMSE=0.11) when the model istrained over synthetic low-resolution (LR) images obtained by down-sampling thenative training HR images, and applied to the synthetic LR validation images.Conversely, poor performances are obtained when the model is trained on a givenspatial resolution and applied to another spatial resolution. Training on a mixof HR and LR images allows to get very good performances on the native HR(rRMSE=0.06) and synthetic LR (rRMSE=0.10) images. However, very lowperformances are still observed over the native LR images (rRMSE=0.48), mainlydue to the poor quality of the native LR images. Finally, an advancedsuper-resolution method based on GAN (generative adversarial network) thatintroduces additional textural information derived from the native HR imageswas applied to the native LR validation images. Results show some significantimprovement (rRMSE=0.22) compared to bicubic up-sampling approach.

Estimates of maize plant density from UAV RGB images using Faster-RCNN detection model: impact of the spatial resolution