Dataset Information

Testing the ability of unmanned aerial systems and machine learning to map weeds at subfield scales: a test with the weed Alopecurus myosuroides (Huds).

ABSTRACT:

Background

It is important to map agricultural weed populations to improve management and maintain future food security. Advances in data collection and statistical methodology have created new opportunities to aid in the mapping of weed populations. We set out to apply these new methodologies (unmanned aerial systems; UAS) and statistical techniques (convolutional neural networks; CNN) to the mapping of black-grass, a highly impactful weed in wheat fields in the UK. We tested this by undertaking extensive UAS and field-based mapping over the course of 2?years, in total collecting multispectral image data from 102 fields, with 76 providing informative data. We used these data to construct a vegetation index (VI), which we used to train a custom CNN model from scratch. We undertook a suite of data engineering techniques, such as balancing and cleaning to optimize performance of our metrics. We also investigate the transferability of the models from one field to another.

Results

The results show that our data collection methodology and implementation of CNN outperform pervious approaches in the literature. We show that data engineering to account for 'artefacts' in the image data increases our metrics significantly. We are not able to identify any traits that are shared between fields that result in high scores from our novel leave one field our cross validation (LOFO-CV) tests.

Conclusion

We conclude that this evaluation procedure is a better estimation of real-world predictive value when compared with past studies. We conclude that by engineering the image data set into discrete classes of data quality we increase the prediction accuracy from the baseline model by 5% to an area under the curve (AUC) of 0.825. We find that the temporal effects studied here have no effect on our ability to model weed densities. © 2019 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

SUBMITTER: Lambert JP

PROVIDER: S-EPMC6767585 | biostudies-literature | 2019 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Testing the ability of unmanned aerial systems and machine learning to map weeds at subfield scales: a test with the weed Alopecurus myosuroides (Huds).

Lambert James Pt JP Childs Dylan Z DZ Freckleton Rob P RP

Pest management science 20190521 8

<h4>Background</h4>It is important to map agricultural weed populations to improve management and maintain future food security. Advances in data collection and statistical methodology have created new opportunities to aid in the mapping of weed populations. We set out to apply these new methodologies (unmanned aerial systems; UAS) and statistical techniques (convolutional neural networks; CNN) to the mapping of black-grass, a highly impactful weed in wheat fields in the UK. We tested this by un ...[more]

PMID: 30972939

Similar Datasets

Project description:BackgroundHerbicide resistance in agrestal weeds is a global problem threatening food security. Non-target-site resistance (NTSR) endowed by mechanisms neutralising the herbicide or compensating for its action is considered the most agronomically noxious type of resistance. Contrary to target-site resistance, NTSR mechanisms are far from being fully elucidated. A part of weed response to herbicide stress, NTSR is considered to be largely driven by gene regulation. Our purpose was to establish a transcriptome resource allowing investigation of the transcriptomic bases of NTSR in the major grass weed Alopecurus myosuroides L. (Poaceae) for which almost no genomic or transcriptomic data was available.ResultsRNA-Seq was performed from plants in one F2 population that were sensitive or expressing NTSR to herbicides inhibiting acetolactate-synthase. Cloned plants were sampled over seven time-points ranging from before until 73 h after herbicide application. Assembly of over 159M high-quality Illumina reads generated a transcriptomic resource (ALOMYbase) containing 65,558 potentially active contigs (N50 = 1240 nucleotides) predicted to encode 32,138 peptides with 74% GO annotation, of which 2017 were assigned to protein families presumably involved in NTSR. Comparison with the fully sequenced grass genomes indicated good coverage and correct representation of A. myosuroides transcriptome in ALOMYbase. The part of the herbicide transcriptomic response common to the resistant and the sensitive plants was consistent with the expected effects of acetolactate-synthase inhibition, with striking similarities observed with published Arabidopsis thaliana data. A. myosuroides plants with NTSR were first affected by herbicide action like sensitive plants, but ultimately overcame it. Analysis of differences in transcriptomic herbicide response between resistant and sensitive plants did not allow identification of processes directly explaining NTSR. Five contigs associated to NTSR in the F2 population studied were tentatively identified. They were predicted to encode three cytochromes P450 (CYP71A, CYP71B and CYP81D), one peroxidase and one disease resistance protein.ConclusionsOur data confirmed that gene regulation is at the root of herbicide response and of NTSR. ALOMYbase proved to be a relevant resource to support NTSR transcriptomic studies, and constitutes a valuable tool for future research aiming at elucidating gene regulations involved in NTSR in A. myosuroides.

Dataset Information

Testing the ability of unmanned aerial systems and machine learning to map weeds at subfield scales: a test with the weed Alopecurus myosuroides (Huds).

Background

Results

Conclusion

Publications

Testing the ability of unmanned aerial systems and machine learning to map weeds at subfield scales: a test with the weed Alopecurus myosuroides (Huds).

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets