Detection of Anomalous Grapevine Berries Using Variational Autoencoders
Grapevine is one of the economically most important quality crops. The monitoring of the plant performance during the growth period is, therefore, important to ensure a high quality end-product. This includes the observation, detection, and respective reduction of unhealthy berries (physically damag...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-06-01
|
| Series: | Frontiers in Plant Science |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpls.2022.729097/full |
| _version_ | 1818208620221825024 |
|---|---|
| author | Miro Miranda Laura Zabawa Anna Kicherer Laurenz Strothmann Uwe Rascher Ribana Roscher Ribana Roscher |
| author_facet | Miro Miranda Laura Zabawa Anna Kicherer Laurenz Strothmann Uwe Rascher Ribana Roscher Ribana Roscher |
| author_sort | Miro Miranda |
| collection | DOAJ |
| description | Grapevine is one of the economically most important quality crops. The monitoring of the plant performance during the growth period is, therefore, important to ensure a high quality end-product. This includes the observation, detection, and respective reduction of unhealthy berries (physically damaged, or diseased). At harvest, it is not necessary to know the exact cause of the damage, but rather if the damage is apparent or not. Since a manual screening and selection before harvest is time-consuming and expensive, we propose an automatic, image-based machine learning approach, which can lead observers directly to anomalous areas without the need to monitor every plant manually. Specifically, we train a fully convolutional variational autoencoder with a feature perceptual loss on images with healthy berries only and consider image areas with deviations from this model as damaged berries. We use heatmaps which visualize the results of the trained neural network and, therefore, support the decision making for farmers. We compare our method against a convolutional autoencoder that was successfully applied to a similar task and show that our approach outperforms it. |
| first_indexed | 2024-12-12T04:47:42Z |
| format | Article |
| id | doaj.art-65bda24057f5426ba5c4651c8b5a7912 |
| institution | Directory Open Access Journal |
| issn | 1664-462X |
| language | English |
| last_indexed | 2024-12-12T04:47:42Z |
| publishDate | 2022-06-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Plant Science |
| spelling | doaj.art-65bda24057f5426ba5c4651c8b5a79122022-12-22T00:37:34ZengFrontiers Media S.A.Frontiers in Plant Science1664-462X2022-06-011310.3389/fpls.2022.729097729097Detection of Anomalous Grapevine Berries Using Variational AutoencodersMiro Miranda0Laura Zabawa1Anna Kicherer2Laurenz Strothmann3Uwe Rascher4Ribana Roscher5Ribana Roscher6Remote Sensing Group, Institute of Geodesy and Geoinformation, University of Bonn, Bonn, GermanyInstitute of Geodesy and Geoinformation, Professorship of Geodesy, University of Bonn, Bonn, GermanyJulius Kühn-Institut, Institute for Grapevine Breeding Geilweilerhof, Geilweilerhof, GermanyInstitute of Bio- and Geosciences IBG-2, Plant Sciences, Forschungszentrum Jülich, Jülich, GermanyInstitute of Bio- and Geosciences IBG-2, Plant Sciences, Forschungszentrum Jülich, Jülich, GermanyRemote Sensing Group, Institute of Geodesy and Geoinformation, University of Bonn, Bonn, GermanyInternational AI Future Lab, Technical University of Munich, Munich, GermanyGrapevine is one of the economically most important quality crops. The monitoring of the plant performance during the growth period is, therefore, important to ensure a high quality end-product. This includes the observation, detection, and respective reduction of unhealthy berries (physically damaged, or diseased). At harvest, it is not necessary to know the exact cause of the damage, but rather if the damage is apparent or not. Since a manual screening and selection before harvest is time-consuming and expensive, we propose an automatic, image-based machine learning approach, which can lead observers directly to anomalous areas without the need to monitor every plant manually. Specifically, we train a fully convolutional variational autoencoder with a feature perceptual loss on images with healthy berries only and consider image areas with deviations from this model as damaged berries. We use heatmaps which visualize the results of the trained neural network and, therefore, support the decision making for farmers. We compare our method against a convolutional autoencoder that was successfully applied to a similar task and show that our approach outperforms it.https://www.frontiersin.org/articles/10.3389/fpls.2022.729097/fullautoencoderdeep learninganomaly detectionviticulturedisease detection |
| spellingShingle | Miro Miranda Laura Zabawa Anna Kicherer Laurenz Strothmann Uwe Rascher Ribana Roscher Ribana Roscher Detection of Anomalous Grapevine Berries Using Variational Autoencoders Frontiers in Plant Science autoencoder deep learning anomaly detection viticulture disease detection |
| title | Detection of Anomalous Grapevine Berries Using Variational Autoencoders |
| title_full | Detection of Anomalous Grapevine Berries Using Variational Autoencoders |
| title_fullStr | Detection of Anomalous Grapevine Berries Using Variational Autoencoders |
| title_full_unstemmed | Detection of Anomalous Grapevine Berries Using Variational Autoencoders |
| title_short | Detection of Anomalous Grapevine Berries Using Variational Autoencoders |
| title_sort | detection of anomalous grapevine berries using variational autoencoders |
| topic | autoencoder deep learning anomaly detection viticulture disease detection |
| url | https://www.frontiersin.org/articles/10.3389/fpls.2022.729097/full |
| work_keys_str_mv | AT miromiranda detectionofanomalousgrapevineberriesusingvariationalautoencoders AT laurazabawa detectionofanomalousgrapevineberriesusingvariationalautoencoders AT annakicherer detectionofanomalousgrapevineberriesusingvariationalautoencoders AT laurenzstrothmann detectionofanomalousgrapevineberriesusingvariationalautoencoders AT uwerascher detectionofanomalousgrapevineberriesusingvariationalautoencoders AT ribanaroscher detectionofanomalousgrapevineberriesusingvariationalautoencoders AT ribanaroscher detectionofanomalousgrapevineberriesusingvariationalautoencoders |