Applications of computer vision techniques in precision viticulture
Precision viticulture is a technique that aims at improving grapevine production and quality while reducing the environmental impact by optimising resource use. For its implementation, the correct, georeferenced, precise measurement of the vine status which represent the inter- and intra-field varia...
Guardado en:
| Autor principal: | |
|---|---|
| Otros Autores: | |
| Formato: | text (thesis) |
| Lenguaje: | eng |
| Publicado: |
Universidad de La Rioja (España)
2017
|
| Materias: | |
| Acceso en línea: | https://dialnet.unirioja.es/servlet/oaites?codigo=122699 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| id |
oai-TES0000022825 |
|---|---|
| record_format |
dspace |
| institution |
DialNet |
| collection |
DialNet |
| language |
eng |
| topic |
Image analysis non-invasive sensing technologies flower number estimation yield estimation canopy status assessment grapevine Vitis vinifera L Análisis de imagen sensores no-invasivos estimación del número de flores estimación de la producción evaluación de la canopy vid Vitis vinifera L |
| spellingShingle |
Image analysis non-invasive sensing technologies flower number estimation yield estimation canopy status assessment grapevine Vitis vinifera L Análisis de imagen sensores no-invasivos estimación del número de flores estimación de la producción evaluación de la canopy vid Vitis vinifera L Millán Prior, Borja Applications of computer vision techniques in precision viticulture |
| description |
Precision viticulture is a technique that aims at improving grapevine production and quality while reducing the environmental impact by optimising resource use. For its implementation, the correct, georeferenced, precise measurement of the vine status which represent the inter- and intra-field variability is mandatory. The development of the geo-positioning systems and sensing technologies, capable of monitoring vine status in a non-invasive, fast and reliable way has stimulated the development and implementation of precision viticulture. Image analysis techniques are currently of increased interest to agricultural monitoring. Their low costs and wide range of applications make them ideal for crop status evaluation.
The main goal of this PhD thesis is to provide new reliable, objective and simple methodologies for vineyard status monitoring using image analysis. To this end, different procedures have been developed to do so: i) assessment of flower number per inflorescence; ii) estimation of the yield before harvest; and iii) evaluation of canopy status. The use of different capturing procedures (manual, smartphone based and on-the-go) was also taken into account, tested and analysed.
The algorithm developed for the assessment of flower number per inflorescence provided estimations with over 90% precision for all the studied varieties. When an improved version of this algorithm was implemented for use in an Android smartphone, the precision rose to the 94%. The new version identified 84% of the flowers present in the image correctly. The number of flower that were visible per image (not all the flowers are visible in the image due to occlusions) was used to estimate the total flower number using a non-lineal model with a root mean square error (RMSE) of 37.1.
The yield assessment before harvest was carried out using two approaches: firstly, a series of vine images were captured manually using a white screen as background, resulting in a classification performance of 98% for clusters and 92% for leaves, this allowed the assessment of the yield with R2 = 0.73. Not all the berries are visible in a vine image due to occlusion from clusters or other parts of the vine. Secondly, the use of a Boolean model was used to reduce the error associated to the occlusion and segmentation errors, resulting in an error in the yield estimation of RMSE = 203g per vine from images captured on-the-go.
Canopy status assessment was carried out with a multi-site experiment conducted in New Zealand, Croatia and Spain. The comparison between the reference method (point quadrat analysis) and the results obtained by analysis of manually captured images (taken on the field using a white screen as background) yielded a determination coefficient over 0.90 on every evaluated site and R2=0.93 when all the data was analysed together. The following experiment was carried out using a modified all-terrain vehicle (ATV) for the automatic image capture at a speed of approximately 7 km/h. This setup permitted high sampling rate data capture and thus vine status map generation. The correlations obtained for the canopy porosity and exposed leaves showed a R2>0.85 and R2>0.71 respectively. Finally, the pruning wood weight is a classic vine vigour indicator. The use of manually captured images (with white screen as background) resulted in RMSE=87.7g and R2=0.91. When the images were captured with a modified ATV the precision slightly dropped to RMSE=115.7 and R2=0.85 but with a significant reduction in the capturing effort.
The results show how computer vision can provide valuable information on vineyard status for precision viticulture. The low cost of the sensor, its non-destructive and fast capturing process offers a great advantage over classical manual reference methods. Image analysis showed high precision in the assessment of flower number per inflorescence, yield estimation and canopy status assessment. The possibility to capture the images on-the-go greatly increases its applicability reducing the effort for data capturing and allowing map generation |
| author2 |
Tardáguila Laso, Javier (null) |
| author_facet |
Tardáguila Laso, Javier (null) Millán Prior, Borja |
| format |
text (thesis) |
| author |
Millán Prior, Borja |
| author_sort |
Millán Prior, Borja |
| title |
Applications of computer vision techniques in precision viticulture |
| title_short |
Applications of computer vision techniques in precision viticulture |
| title_full |
Applications of computer vision techniques in precision viticulture |
| title_fullStr |
Applications of computer vision techniques in precision viticulture |
| title_full_unstemmed |
Applications of computer vision techniques in precision viticulture |
| title_sort |
applications of computer vision techniques in precision viticulture |
| publisher |
Universidad de La Rioja (España) |
| publishDate |
2017 |
| url |
https://dialnet.unirioja.es/servlet/oaites?codigo=122699 |
| work_keys_str_mv |
AT millanpriorborja applicationsofcomputervisiontechniquesinprecisionviticulture |
| _version_ |
1718346663565721600 |
| spelling |
oai-TES00000228252018-05-13Applications of computer vision techniques in precision viticultureMillán Prior, BorjaImage analysisnon-invasive sensing technologiesflower number estimationyield estimationcanopy status assessmentgrapevineVitis vinifera LAnálisis de imagensensores no-invasivosestimación del número de floresestimación de la producciónevaluación de la canopyvid Vitis vinifera LPrecision viticulture is a technique that aims at improving grapevine production and quality while reducing the environmental impact by optimising resource use. For its implementation, the correct, georeferenced, precise measurement of the vine status which represent the inter- and intra-field variability is mandatory. The development of the geo-positioning systems and sensing technologies, capable of monitoring vine status in a non-invasive, fast and reliable way has stimulated the development and implementation of precision viticulture. Image analysis techniques are currently of increased interest to agricultural monitoring. Their low costs and wide range of applications make them ideal for crop status evaluation. The main goal of this PhD thesis is to provide new reliable, objective and simple methodologies for vineyard status monitoring using image analysis. To this end, different procedures have been developed to do so: i) assessment of flower number per inflorescence; ii) estimation of the yield before harvest; and iii) evaluation of canopy status. The use of different capturing procedures (manual, smartphone based and on-the-go) was also taken into account, tested and analysed. The algorithm developed for the assessment of flower number per inflorescence provided estimations with over 90% precision for all the studied varieties. When an improved version of this algorithm was implemented for use in an Android smartphone, the precision rose to the 94%. The new version identified 84% of the flowers present in the image correctly. The number of flower that were visible per image (not all the flowers are visible in the image due to occlusions) was used to estimate the total flower number using a non-lineal model with a root mean square error (RMSE) of 37.1. The yield assessment before harvest was carried out using two approaches: firstly, a series of vine images were captured manually using a white screen as background, resulting in a classification performance of 98% for clusters and 92% for leaves, this allowed the assessment of the yield with R2 = 0.73. Not all the berries are visible in a vine image due to occlusion from clusters or other parts of the vine. Secondly, the use of a Boolean model was used to reduce the error associated to the occlusion and segmentation errors, resulting in an error in the yield estimation of RMSE = 203g per vine from images captured on-the-go. Canopy status assessment was carried out with a multi-site experiment conducted in New Zealand, Croatia and Spain. The comparison between the reference method (point quadrat analysis) and the results obtained by analysis of manually captured images (taken on the field using a white screen as background) yielded a determination coefficient over 0.90 on every evaluated site and R2=0.93 when all the data was analysed together. The following experiment was carried out using a modified all-terrain vehicle (ATV) for the automatic image capture at a speed of approximately 7 km/h. This setup permitted high sampling rate data capture and thus vine status map generation. The correlations obtained for the canopy porosity and exposed leaves showed a R2>0.85 and R2>0.71 respectively. Finally, the pruning wood weight is a classic vine vigour indicator. The use of manually captured images (with white screen as background) resulted in RMSE=87.7g and R2=0.91. When the images were captured with a modified ATV the precision slightly dropped to RMSE=115.7 and R2=0.85 but with a significant reduction in the capturing effort. The results show how computer vision can provide valuable information on vineyard status for precision viticulture. The low cost of the sensor, its non-destructive and fast capturing process offers a great advantage over classical manual reference methods. Image analysis showed high precision in the assessment of flower number per inflorescence, yield estimation and canopy status assessment. The possibility to capture the images on-the-go greatly increases its applicability reducing the effort for data capturing and allowing map generationLa viticultura de precisión permite mejorar la calidad y producción de la uva, al mismo tiempo que optimiza el uso de los recursos, reduciendo el impacto ambiental. Para su correcta implementación es necesaria la medida precisa y georreferenciada del estado del viñedo, de forma que se represente la variabilidad intra e inter parcela. Los recientes progresos en sistemas de geo-posicionamiento y sensores capaces de monitorizar el viñedo de forma rápida, no invasiva y precisa han impulsado el desarrollo e implementación de la viticultura de precisión, aunque su uso comercial es limitado. Entre los diferentes tipos de sensores disponibles, destacan los basados en análisis de imagen, que están experimentando un fuerte desarrollo en los últimos años gracias a su bajo coste y amplio rango de aplicaciones. Debido a sus características, el análisis de imagen es una tecnología clave para la viticultura de precisión y su implantación comercial. El objetivo principal de este trabajo es el desarrollo de nuevas metodologías de monitorización del viñedo mediante el análisis de imagen. Con esta finalidad se han desarrollado y evaluado nuevas técnicas para: i) estimación del número de flores por inflorescencia; ii) predicción de la cosecha; y iii) evaluación del estado de la “canopy”. Para ello se han utilizado diferentes métodos de adquisición de imagen, incluyendo la captura manual, el uso de “smartphones” y la utilización de plataformas móviles que realizan la adquisición de forma automática. La precisión del algoritmo para el conteo de flores por inflorescencia fue superior al 90% en todas las variedades evaluadas. Con el fin de facilitar el uso de esta metodología en el viñedo, se desarrolló una versión mejorada del algoritmo compatible con “smartphones“ de sistema operativo Android. La aplicación fue capaz de identificar correctamente el 84% de las flores presentes por imagen, obteniendo una precisión del 94% y un error cuadrático medio (RMSE) de 37,1 en la estimación del número total de flores por inflorescencia. La predicción de la cosecha se realizó mediante dos enfoques distintos: a partir de imágenes capturadas de forma manual utilizando un fondo blanco o con una plataforma móvil capaz de realizar la captura de forma automatizada. En el primer caso se logró la clasificación correcta del 98% 6 y 92% de los píxeles correspondientes a racimos y hojas respectivamente, obteniéndose la estimación de la producción con alta precisión (R2=0,73). En el segundo caso se utilizó el modelo Booleano para mejorar la precisión de la estimación frente a oclusiones y errores de segmentación, obteniéndose un error (RMSE) de 203g por cepa. La capacidad de medida del estado de la “canopy” mediante análisis de imagen se ha evaluado con experimentos ejecutados en Nueva Zelanda, Croacia, y España, de forma que se pudo valorar la robustez del sistema frente a diferentes variedades y sistemas de manejo. Se obtuvo un coeficiente de determinación superior a 0,90 para la relación entre el método de referencia (“point quadrat analysis”) y el algoritmo de análisis de imágenes (capturadas manualmente utilizando un fondo blanco) para cada uno de los experimentos y de R2=0,93 cuando todos los datos se analizaron de forma conjunta. Para aumentar la aplicabilidad comercial de la metodología, se modificó un “quad” de forma que la captura de las imágenes se realizara de forma automática y continua a una velocidad en torno a 7 km/h. Con esta metodología se pudo evaluar la porosidad del viñedo (R2>0,85) y hojas expuestas (R2>0,71), y gracias a la alta densidad de muestreo se pudieron realizar mapas representativos de la variabilidad del viñedo. Finalmente, también se evaluó la capacidad de estimar el peso de la madera de poda, que es un indicador del vigor del viñedo. Mediante el análisis de las imágenes capturadas de manera manual se obtuvo una estimación (R2=0,91) con un error (RMSE) de 87,7g por cepa. Cuando la captura de imágenes se realizó de forma automatizada y en continuo, la precisión descendió ligeramente (RMSE=115,7; R2=0,85), pero con una importante reducción en el esfuerzo requerido para la obtención de las imágenes. Los resultados obtenidos muestran que el análisis de imagen es una tecnología de gran interés para la viticultura de precisión. El bajo coste de los sensores, la captura rápida y no destructiva y la alta precisión y variedad de los parámetros que pueden ser medidos representa importantes ventajas frente a los métodos clásicos. Los algoritmos desarrollados permiten la estimación del número de flores por inflorescencia, predicción de la producción y evaluación de la “canopy” con gran precisión. La posibilidad de captura de imágenes desde plataformas móviles reduce el esfuerzo de captura y permite la generación de mapas, facilitando el uso de estas técnicas a nivel comercial en el sector vitícola.Universidad de La Rioja (España)Tardáguila Laso, Javier (null)2017text (thesis)application/pdfhttps://dialnet.unirioja.es/servlet/oaites?codigo=122699engLICENCIA DE USO: Los documentos a texto completo incluidos en Dialnet son de acceso libre y propiedad de sus autores y/o editores. Por tanto, cualquier acto de reproducción, distribución, comunicación pública y/o transformación total o parcial requiere el consentimiento expreso y escrito de aquéllos. Cualquier enlace al texto completo de estos documentos deberá hacerse a través de la URL oficial de éstos en Dialnet. Más información: https://dialnet.unirioja.es/info/derechosOAI | INTELLECTUAL PROPERTY RIGHTS STATEMENT: Full text documents hosted by Dialnet are protected by copyright and/or related rights. This digital object is accessible without charge, but its use is subject to the licensing conditions set by its authors or editors. Unless expressly stated otherwise in the licensing conditions, you are free to linking, browsing, printing and making a copy for your own personal purposes. All other acts of reproduction and communication to the public are subject to the licensing conditions expressed by editors and authors and require consent from them. Any link to this document should be made using its official URL in Dialnet. More info: https://dialnet.unirioja.es/info/derechosOAI |