A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica

Introducing artificial intelligence techniques in agriculture offers new opportunities for improving crop management, such as in coffee plantations, which constitute a complex agroforestry environment. This paper presents a comparative study of three deep learning frameworks: Deep Forest, RT-DETR,...

Descripción completa

Detalles Bibliográficos
Autores principales: Arriola-Valverde, Sergio, Rimolo-Donadio, Renato, Villagra-Mendoza, Karolina, Chacón-Rodriguez, Alfonso, García-Ramirez, Ronny, Somarriba, Eduardo
Formato: Artículo
Lenguaje:Inglés
Publicado: MDPI 2024
Materias:
Acceso en línea:https://repositorio.catie.ac.cr/handle/11554/12711
_version_ 1855487779807428608
author Arriola-Valverde, Sergio
Rimolo-Donadio, Renato
Villagra-Mendoza, Karolina
Chacón-Rodriguez, Alfonso
García-Ramirez, Ronny
Somarriba, Eduardo
author_browse Arriola-Valverde, Sergio
Chacón-Rodriguez, Alfonso
García-Ramirez, Ronny
Rimolo-Donadio, Renato
Somarriba, Eduardo
Villagra-Mendoza, Karolina
author_facet Arriola-Valverde, Sergio
Rimolo-Donadio, Renato
Villagra-Mendoza, Karolina
Chacón-Rodriguez, Alfonso
García-Ramirez, Ronny
Somarriba, Eduardo
author_sort Arriola-Valverde, Sergio
collection Repositorio CATIE
description Introducing artificial intelligence techniques in agriculture offers new opportunities for improving crop management, such as in coffee plantations, which constitute a complex agroforestry environment. This paper presents a comparative study of three deep learning frameworks: Deep Forest, RT-DETR, and Yolov9, customized for coffee plant detection and trained from images with a high spatial resolution (cm/pix). Each frame had dimensions of 640 × 640 pixels acquired from passive RGB sensors onboard a UAS (Unmanned Aerial Systems) system. The image set was structured and consolidated from UAS-RGB imagery acquisition in six locations along the Central Valley, Costa Rica, through automated photogrammetric missions. It was evidenced that the RTDETR and Yolov9 frameworks allowed adequate generalization and detection with mAP50 values higher than 90% and mAP5095 higher than 54%, in scenarios of application with data augmentation techniques. Deep Forest also achieved good metrics, but noticeably lower when compared to the other frameworks. RT-DETR and Yolov9 were able to generalize and detect coffee plants in unseen scenarios that include complex forest structures within tropical agroforestry Systems (AFS)
format Artículo
id RepoCATIE12711
institution Centro Agronómico Tropical de Investigación y Enseñanza
language Inglés
publishDate 2024
publishDateRange 2024
publishDateSort 2024
publisher MDPI
publisherStr MDPI
record_format dspace
spelling RepoCATIE127112024-12-11T23:15:51Z A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica Arriola-Valverde, Sergio Rimolo-Donadio, Renato Villagra-Mendoza, Karolina Chacón-Rodriguez, Alfonso García-Ramirez, Ronny Somarriba, Eduardo Coffea||Coffea||Coffea||Coffea Agricultura de precisión||precision agriculture||agriculture de précision Manejo del cultivo||crop management||gestão da colheita||conduite de la culture Imágen por satélite||satellite imagery||imagem por satélite||imagerie par satellite Costa Rica||Costa Rica||Costa Rica||Costa Rica Sede Central ODS 9 - Industria, innovación e infraestructura Introducing artificial intelligence techniques in agriculture offers new opportunities for improving crop management, such as in coffee plantations, which constitute a complex agroforestry environment. This paper presents a comparative study of three deep learning frameworks: Deep Forest, RT-DETR, and Yolov9, customized for coffee plant detection and trained from images with a high spatial resolution (cm/pix). Each frame had dimensions of 640 × 640 pixels acquired from passive RGB sensors onboard a UAS (Unmanned Aerial Systems) system. The image set was structured and consolidated from UAS-RGB imagery acquisition in six locations along the Central Valley, Costa Rica, through automated photogrammetric missions. It was evidenced that the RTDETR and Yolov9 frameworks allowed adequate generalization and detection with mAP50 values higher than 90% and mAP5095 higher than 54%, in scenarios of application with data augmentation techniques. Deep Forest also achieved good metrics, but noticeably lower when compared to the other frameworks. RT-DETR and Yolov9 were able to generalize and detect coffee plants in unseen scenarios that include complex forest structures within tropical agroforestry Systems (AFS) 2024-12-10T17:06:03Z 2024-12-10T17:06:03Z 2024-11-10 Artículo https://repositorio.catie.ac.cr/handle/11554/12711 openAccess en Remote Sensing https://doi.org/10.3390/rs16244617 27 páginas application/pdf MDPI
spellingShingle Coffea||Coffea||Coffea||Coffea
Agricultura de precisión||precision agriculture||agriculture de précision
Manejo del cultivo||crop management||gestão da colheita||conduite de la culture
Imágen por satélite||satellite imagery||imagem por satélite||imagerie par satellite
Costa Rica||Costa Rica||Costa Rica||Costa Rica
Sede Central
ODS 9 - Industria, innovación e infraestructura
Arriola-Valverde, Sergio
Rimolo-Donadio, Renato
Villagra-Mendoza, Karolina
Chacón-Rodriguez, Alfonso
García-Ramirez, Ronny
Somarriba, Eduardo
A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica
title A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica
title_full A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica
title_fullStr A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica
title_full_unstemmed A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica
title_short A Comparative Study of Deep Learning Frameworks Applied to Coffee Plant Detection from Close-Range UAS-RGB Imagery in Costa Rica
title_sort comparative study of deep learning frameworks applied to coffee plant detection from close range uas rgb imagery in costa rica
topic Coffea||Coffea||Coffea||Coffea
Agricultura de precisión||precision agriculture||agriculture de précision
Manejo del cultivo||crop management||gestão da colheita||conduite de la culture
Imágen por satélite||satellite imagery||imagem por satélite||imagerie par satellite
Costa Rica||Costa Rica||Costa Rica||Costa Rica
Sede Central
ODS 9 - Industria, innovación e infraestructura
url https://repositorio.catie.ac.cr/handle/11554/12711
work_keys_str_mv AT arriolavalverdesergio acomparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT rimolodonadiorenato acomparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT villagramendozakarolina acomparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT chaconrodriguezalfonso acomparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT garciaramirezronny acomparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT somarribaeduardo acomparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT arriolavalverdesergio comparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT rimolodonadiorenato comparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT villagramendozakarolina comparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT chaconrodriguezalfonso comparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT garciaramirezronny comparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica
AT somarribaeduardo comparativestudyofdeeplearningframeworksappliedtocoffeeplantdetectionfromcloserangeuasrgbimageryincostarica