Publicação
Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’
| Resumo: | Today there is a strong demand for fast and reliable vineyard yield estimation methods as it can bring several benefits to the wine industry. Recently, a strong research effort has been made to apply image analysis technologies for the recognition of grapevine yield components (YCs) in ground-based images collected with on-the-go platforms and to automate processing methods for yield estimation. YC detection depends on the magnitude of obstructions/occlusions that varies along the growing cycle, as vines develop and vegetation grows. In this work, grapevine images taken under field conditions were analyzed aiming at evaluating the degree of YC visibility at different phenological stages. Data were collected in 2019 in a spur-pruned vineyard plot of the white cultivar ‘Encruzado’ trained on a vertical shoot positioning trellis system. Images were obtained on-the-go with a RGB camera mounted on an unmanned ground vehicle facing the sunlit side of the canopy. Spurs and YCs (i.e., shoots, inflorescences and bunches) were counted in the image and compared to ground-truth measurements. During winter the absence of leaves enabled an easy detection of the number of spurs left after pruning (mean absolute percentage error (MA%E) <1%). During spring, shoot number was also easy to detect shortly after bud burst, although the detection error was higher (MA%E=31%) than during dormancy. Inflorescences and bunches showed the highest MA%E (≥59%), before flowering, at pea size and at veraison, with a slight decrease at harvest (MA%E=46%). Our results showed that spurs and shoots were easily detected by image analysis, although they were not always well correlated to final yield. In conclusion, as YCs visibility was low in any stage after fruit set, pea size or veraison were the best stages to collect images for yield estimation purposes |
|---|---|
| Autores principais: | Vitorino, G. |
| Outros Autores: | Lopes, C.M. |
| Assunto: | precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| Ano: | 2021 |
| País: | Portugal |
| Tipo de documento: | artigo |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| _version_ | 1865920801585233920 |
|---|---|
| author | Vitorino, G. |
| author2 | Lopes, C.M. |
| author2_role | author |
| author_facet | Vitorino, G. Vitorino, G. Lopes, C.M. Lopes, C.M. |
| author_role | author |
| contributor_name_str_mv | Repositório Científico de Acesso Aberto da ULisboa |
| country_str | PT |
| creators_json_str | [{\"Person.name\":\"Vitorino, G.\"},{\"Person.name\":\"Lopes, C.M.\"}] |
| datacite.contributors.contributor.contributorName.fl_str_mv | Repositório Científico de Acesso Aberto da ULisboa |
| datacite.creators.creator.creatorName.fl_str_mv | Vitorino, G. Lopes, C.M. |
| datacite.date.Accepted.fl_str_mv | 2021-01-01T00:00:00Z |
| datacite.date.available.fl_str_mv | 2023-01-31T01:30:23Z |
| datacite.date.embargoed.fl_str_mv | 2023-01-31T01:30:23Z |
| datacite.rights.fl_str_mv | http://purl.org/coar/access_right/c_abf2 |
| datacite.subjects.subject.fl_str_mv | precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| datacite.titles.title.fl_str_mv | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| dc.contributor.none.fl_str_mv | Repositório Científico de Acesso Aberto da ULisboa |
| dc.creator.none.fl_str_mv | Vitorino, G. Lopes, C.M. |
| dc.date.Accepted.fl_str_mv | 2021-01-01T00:00:00Z |
| dc.date.available.fl_str_mv | 2023-01-31T01:30:23Z |
| dc.date.embargoed.fl_str_mv | 2023-01-31T01:30:23Z |
| dc.format.none.fl_str_mv | application/pdf |
| dc.identifier.none.fl_str_mv | http://hdl.handle.net/10400.5/23321 |
| dc.language.none.fl_str_mv | eng |
| dc.publisher.none.fl_str_mv | ISHS |
| dc.rights.none.fl_str_mv | http://purl.org/coar/access_right/c_abf2 |
| dc.subject.none.fl_str_mv | precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| dc.title.fl_str_mv | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| dc.type.none.fl_str_mv | http://purl.org/coar/resource_type/c_6501 |
| description | Today there is a strong demand for fast and reliable vineyard yield estimation methods as it can bring several benefits to the wine industry. Recently, a strong research effort has been made to apply image analysis technologies for the recognition of grapevine yield components (YCs) in ground-based images collected with on-the-go platforms and to automate processing methods for yield estimation. YC detection depends on the magnitude of obstructions/occlusions that varies along the growing cycle, as vines develop and vegetation grows. In this work, grapevine images taken under field conditions were analyzed aiming at evaluating the degree of YC visibility at different phenological stages. Data were collected in 2019 in a spur-pruned vineyard plot of the white cultivar ‘Encruzado’ trained on a vertical shoot positioning trellis system. Images were obtained on-the-go with a RGB camera mounted on an unmanned ground vehicle facing the sunlit side of the canopy. Spurs and YCs (i.e., shoots, inflorescences and bunches) were counted in the image and compared to ground-truth measurements. During winter the absence of leaves enabled an easy detection of the number of spurs left after pruning (mean absolute percentage error (MA%E) <1%). During spring, shoot number was also easy to detect shortly after bud burst, although the detection error was higher (MA%E=31%) than during dormancy. Inflorescences and bunches showed the highest MA%E (≥59%), before flowering, at pea size and at veraison, with a slight decrease at harvest (MA%E=46%). Our results showed that spurs and shoots were easily detected by image analysis, although they were not always well correlated to final yield. In conclusion, as YCs visibility was low in any stage after fruit set, pea size or veraison were the best stages to collect images for yield estimation purposes |
| dirty | 0 |
| eu_rights_str_mv | openAccess |
| format | article |
| fulltext.url.fl_str_mv | https://repositorio.ulisboa.pt/bitstreams/58001294-ac76-428d-9b5d-16aac0c71031/download |
| id | ul_64f5ed32c573ea48bdb194e252541bcf |
| identifier.url.fl_str_mv | http://hdl.handle.net/10400.5/23321 |
| instacron_str | ul |
| institution | Universidade de Lisboa |
| instname_str | Universidade de Lisboa |
| language | eng |
| network_acronym_str | ul |
| network_name_str | Repositório da Universidade de Lisboa |
| oai_identifier_str | oai:repositorio.ulisboa.pt:10400.5/23321 |
| organization_str_mv | urn:organizationAcronym:ul |
| person_str_mv | Vitorino, G. Lopes, C.M. |
| publishDate | 2021 |
| publisher.none.fl_str_mv | ISHS |
| reponame_str | Repositório da Universidade de Lisboa |
| repository_id_str | urn:repositoryAcronym:ul |
| service_str_mv | urn:repositoryAcronym:ul |
| spelling | engISHSpt_PTToday there is a strong demand for fast and reliable vineyard yield estimation methods as it can bring several benefits to the wine industry. Recently, a strong research effort has been made to apply image analysis technologies for the recognition of grapevine yield components (YCs) in ground-based images collected with on-the-go platforms and to automate processing methods for yield estimation. YC detection depends on the magnitude of obstructions/occlusions that varies along the growing cycle, as vines develop and vegetation grows. In this work, grapevine images taken under field conditions were analyzed aiming at evaluating the degree of YC visibility at different phenological stages. Data were collected in 2019 in a spur-pruned vineyard plot of the white cultivar ‘Encruzado’ trained on a vertical shoot positioning trellis system. Images were obtained on-the-go with a RGB camera mounted on an unmanned ground vehicle facing the sunlit side of the canopy. Spurs and YCs (i.e., shoots, inflorescences and bunches) were counted in the image and compared to ground-truth measurements. During winter the absence of leaves enabled an easy detection of the number of spurs left after pruning (mean absolute percentage error (MA%E) <1%). During spring, shoot number was also easy to detect shortly after bud burst, although the detection error was higher (MA%E=31%) than during dormancy. Inflorescences and bunches showed the highest MA%E (≥59%), before flowering, at pea size and at veraison, with a slight decrease at harvest (MA%E=46%). Our results showed that spurs and shoots were easily detected by image analysis, although they were not always well correlated to final yield. In conclusion, as YCs visibility was low in any stage after fruit set, pea size or veraison were the best stages to collect images for yield estimation purposesapplication/pdfpt_PTGrapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’Vitorino, G.Lopes, C.M.HostingInstitutionOrganizationalRepositório Científico de Acesso Aberto da ULisboae-mailmailto:repositorio@reitoria.ulisboa.ptrepositorio@reitoria.ulisboa.ptDOIIsPartOfDOI: 10.17660/ActaHortic.2021.1314.222023-01-31T01:30:23Z20212021-01-01T00:00:00ZHandlehttp://hdl.handle.net/10400.5/23321http://purl.org/coar/access_right/c_abf2open accessprecision viticulturebunch occlusionmachine visionyield estimationrobotic platform437070 bytesliteraturehttp://purl.org/coar/resource_type/c_6501journal articlehttp://purl.org/coar/access_right/c_abf2application/pdffulltexthttps://repositorio.ulisboa.pt/bitstreams/58001294-ac76-428d-9b5d-16aac0c71031/downloadActa Horticulturae |
| spellingShingle | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ Vitorino, G. precision viticulture bunch occlusion machine vision yield estimation robotic platform Vitorino, G. precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| status | SINGLETON |
| subject.fl_str_mv | precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| title | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| title_full | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| title_fullStr | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| title_full_unstemmed | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| title_short | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| title_sort | Grapevine yield components detection using image analysis: a case study with the white cultivar ‘Encruzado’ |
| topic | precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| topic_facet | precision viticulture bunch occlusion machine vision yield estimation robotic platform |
| url | http://hdl.handle.net/10400.5/23321 |
| visible | 1 |