Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 6
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Towards a non-destructive foliar nutrient estimation in &apos, Clementina de Nules&apos, mandarin through hyperspectral imaging-based leaf age discrimination

  • Munera, S|||0000-0003-3064-1186
  • Rodriguez-Ortega, Alejandro|||0000-0003-4781-6368
  • Acosta, Maylin
  • Rodriguez-Carretero, Isabel
  • Blasco, Jose
  • Quinones, Ana
[EN] BackgroundNon-destructive foliar nutrient estimation is a critical component of sustainable fertilisation strategies in citrus orchards. Diagnosis relies on analysing spring flush leaves at a specific stage, but field sampling often includes mixed ages, causing inaccurate estimations and poor fertilisation decisions. This study explored visible/near-infrared hyperspectral imaging (500-980 nm) to discriminate leaf age and estimate nutrient concentrations in Clementina de Nules mandarin.MethodsHyperspectral images were analysed using partial least squares discriminant analysis (PLS-DA) to classify leaf age and partial least squares regression (PLS-R) to predict nutrient concentrations. Eleven nutrients were evaluated, and optimal wavelengths identified to reduce dimensionality and enable portable sensor development.ResultsThe PLS-DA model achieved 99.4% accuracy in distinguishing young from mature leaves, addressing a critical challenge for reliable diagnostics. Among the eleven nutrients, PLS-R models moderately predicted phosphorus, potassium, calcium, iron, and copper, with test set R2 values of 0.60-0.69 and prediction-to-deviation ratios (RPD) of 1.5-1.8, allowing differentiation of high and low nutrient levels. Accuracy was Powered by Editorial Manager (R) and ProduXion Manager (R) from Aries Systems Corporation maintained while reducing dimensionality by selecting ten optimal wavelengths per nutrient.ConclusionThis study improves the reliability of non-destructive nutrient diagnosis in citrus by addressing leaf age variability and provides a starting point for future in-field, real-time monitoring systems. The approach supports optimised fertilisation strategies aligned with precision agriculture and European Green Deal sustainability goals, offering cost-effective, field-deployable solutions for citrus management., This study receives funding from MCIN/AEI/https://doi.org/10.13039/501100011033 and the EU "NextGenerationEU"/PRTR through projects TED2021-130117B-C31 and C32, GVA-IVIA through projects 52203 and 52204, and the ERDF of the Generalitat Valenciana 2021 2027. Maylin Acosta thanks the Professional Excellence Scholarship from IFARHU-SENACYT under contract No. 270-2021-020. Sandra Munera thanks the Juan de la CiervaFormación postdoctoral contract (FJC2021-047786-I), cofunded by MCIN/AEI and the EU "NextGenerationEU"/PRTR.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Tesis doctoral (DoctoralThesis)

Predicción de macro y micronutrientes en hojas de cítricos y caqui utilizando métodos ópticos no destructivos

  • Acosta Tello, Maylin Oristela
[ES] El conocimiento del estado nutricional de los cultivos permite corregir o ajustar cualquier exceso o deficiencia nutricional en los mismos, a lo largo de su ciclo vegetativo, asegurando un alto rendimiento en la producción y una óptima calidad del fruto. Tradicionalmente, para la realización del diagnóstico nutricional se ha utilizado el análisis de la ionómica de diferentes órganos de la planta, especialmente las hojas, por su facilidad de muestreo y por ser el órgano fotosintético de excelencia en las plantas. Por ello es necesario implementar estrategias sostenibles que nos permitan ajustar la dosis de fertilización según las necesidades del cultivo con el mínimo riesgo de contaminación. El objetivo de esta tesis doctoral es desarrollar métodos y modelos que permitan el diagnóstico nutricional en cultivos mediante métodos ópticos no destructivos, como la espectroscopia y la imagen hiperespectral en el rango Vis-NIR, en combinación con técnicas de quimiometría. De este modo, el primer bloque centrado en el cultivo de caqui cv. 'Rojo Brillante', comprende los estudios publicados en dos artículos científicos. En el primero de estos artículos se estudió el potencial de la espectroscopia Vis-NIR (430-1040 nm), con el objetivo de predecir macros y micronutrientes utilizando modelos de regresión PLS. Los resultados mostraron que es posible predecir de forma precisa macronutrientes como, fósforo (P), calcio (Ca) y magnesio (Mg), con un coeficiente de determinación en la predicción (R2P) de 0,78 a 0,63. En los micronutrientes, el boro (B) y el manganeso (Mn) fueron los que obtuvieron mejores coeficientes de predicción, con R2P de 0,79 y 0,69, respectivamente. En el segundo artículo se ha evaluado, para la estimación de la concentración de nutrientes, el uso de imágenes hiperespectrales en el rango entre 500 y 980 nm. Los resultados mostraron la predicción de los macronutrientes como N, P, potasio (K), Ca y Mg con R2P de 0,80 a 0,62 y, para los micronutrientes, solo en el B se obtuvo un valor aceptable para la estimación (R2p = 0,69). Además, utilizando el método de reducción de variables de influencia en la proyección (VIP) se obtuvo una predicción fiable para los nutrientes de N (R2P = 0,76) y B (R2P = 0,61). En el segundo bloque, se ha estudiado otro cultivo emblemático en la Comunidad Valenciana por su importancia social y económica, como son los cítricos. En este caso, se desarrollaron herramientas de estimación del cv. 'clementina de Nules', descritas en otros dos artículos científicos. En el tercer artículo se ha estudiado la capacidad de la espectroscopía para determinar la concentración de nutrientes en las hojas de los cítricos en un ciclo vegetativo completo. Los resultados mostraron una predicción con un R2P de 0,70 a 0,65 para el P, K Ca y B. Utilizando el coeficiente de regresión de ponderado (BW) se determinó un subconjunto de bandas importantes para determinar la concentración de P, K y B. Los resultados mostraron que las bandas de mayor relevancia para estos nutrientes se situaron en la región del visible (430-750 nm), asociada a la absorción de pigmentos fotosintéticos. Finalmente, en el cuarto artículo se ha estudiado el potencial de la imagen hiperespectral para discriminar entre hojas jóvenes y hojas de ciclos vegetativos anteriores, lo que mejoraría el diagnóstico dado que las tablas de referencia en este cultivo están realizadas en hojas de la brotación de primavera. Partiendo de esa hipótesis, se obtuvo que es posible realizar la discriminación entre ambos tipos de hojas. Posteriormente se realizó la predicción de concentración de nutrientes de hojas jóvenes, utilizando 49 bandas espectrales, obteniendo mejores resultados para los nutrientes P, K, Ca, hierro (Fe) y Mn con R2P de 0,69 a 0,60. Además, se realizó la predicción de estos nutrientes minimizando el número de bandas a diez, con el BW y se obtuvo un R2P de 0,67 a 0,57., [CA] El coneixement de l'estat nutricional dels cultius permet corregir o ajustar qualsevol excés o deficiència nutricional en estos, al llarg del seu cicle vegetatiu, assegurant un alt rendiment en la producció i una òptima qualitat del fruit. Tradicionalment, per a la realització del diagnòstic nutricional s'han utilitzat l'anàlisi de la ionómica de diferents òrgans de la planta, especialment les fulles, per la seua facilitat de mostreig i per ser l'òrgan fotosintètic d'excellència en les plantes. Per això és necessari implementar estratègies sostenibles que ens permeten ajustar la dosi de fertilització segons les necessitats del cultiu amb el mínim risc de contaminació. L'objectiu d'esta tesi doctoral és desenvolupar mètodes i models que permeten la predicció del diagnòstic nutricional en cultius mitjançant mètodes òptics no destructius, com l'espectroscòpia Vis-NIR, en combinació amb tècniques quimio mètriques. D'esta manera, el primer capítol de publicacions es centra en el cultiu del caqui cv. Rojo Brillante, comprés per dos articles (I i II). En el primer d'estos articles es va estudiar el potencial de l'espectroscòpia Vis-NIR (430-1040 nm), amb l'objectiu de predir macros i micronutrients utilitzant models de regressió PLS. En este estudi es van aplicar tractaments diferencials per als nutrients de N (0 %, 33 %, 50 % i 100 %) i per a K2O (0 %, 50 % i 100 %) de la demanda del cultiu. Els resultats van mostrar que, sí que és possible predir de manera precisa macronutrients com, fòsfor (P), calci (Ca)i magnesi (Mg), amb un coeficient de determinació en la predicció (R2P) de 0,78 a 0,63. En els micronutrients, com el bor (B) i el manganés (Mn) van ser els que van obtindre millors coeficients de predicció, amb R2P de 0,79 i 0,69, respectivament. En el segon article s'ha avaluat, per a l'estimació de la concentració de nutrients, l'ús d'imatges hiperespectrales en un rang (500-980 nm). Els resultats van mostrar la predicció dels macronutrients com a nitrogen (N), P, potassi (K), Ca i Mg amb R2P de 0,80 a 0,62 i, per als micronutrients, només en el B es va obtindre un valor acceptable per a l'estimació (R2p 0,69). A més, utilitzant el mètode de reducció de variables d'influència en la projecció (VIP) es va obtindre una predicció fiable per als nutrients de N (R2P 0,76) i B (R2P 0,61). En el segon capítol, s'ha estudiat un altre cultiu emblemàtic a la Comunitat Valenciana d'importància econòmica, com són els cítrics. En este cas, es van desenvolupar ferramentes d'estimació del cv. 'clementina de Nules' compreses en dos articles (III i IV). De tal manera, que en el tercer article s'ha estudiat la capacitat de les tècniques espectromètriques per a determinar la concentració de nutrients en un cicle vegetatiu complet. Els resultats van mostrar una predicció amb un R2P de 0,70 a 0,65 per al P, K, Ca i B. Utilitzant el coeficient de regressió de pes (BW) es va determinar un subconjunt de bandes més influents per als nutrients P, K i B. Els resultats van mostrar que les bandes de major importància, per a estos nutrients, es situen a la regió del Vis (430-750 nm), el qual està associada a l'absorció de pigments fotosintètics. Finalment, en el quart article s'ha estudiat el potencial de les HSI per a discriminar fulles joves de fulles de cicles vegetatius anteriors, la qual cosa milloraria el diagnòstic atés que les taules de referència en este cultiu estan realitzades en fulles de la brotada de primavera. Posteriorment es va realitzar la predicció de concentració de nutrients de fulles joves, utilitzant 49 bandes espectrals, obtenint millors resultats per als nutrients P, K, Ca, ferro (Fe) i Mn amb R2P de 0,69 a 0,60. A més, es va realitzar la predicció d'estos nutrients minimitzant el nombre de bandes a deu, amb el BW i es va obtindre un R2P de 0,67 a 0,57., [EN] Knowledge of the nutritional status of crops allows for correcting or adjusting any nutritional excess or deficiency throughout their vegetative cycle, ensuring high yields in production and optimal fruit quality. Traditionally, the analysis of the ionomics of different plant organs has been used for nutritional diagnosis, especially the leaves, due to their ease of sampling and being the photosynthetic organ par excellence in plants. These analyses are carried out by expensive conventional laboratory methods that are destructive, polluting, time-consuming and costly. Therefore, it is necessary to implement sustainable strategies that allow the fertilisation dose to be adjusted according to the crop's needs with the minimum risk of contamination. This doctoral thesis aims to develop methods and models for nutritional diagnosis prediction in crops using non-destructive optical methods, such as Vis-NIR spectroscopy, combined with chemometric techniques. Thus, the first chapter of the publications focuses on cultivating persimmon cv. 'Rojo Brillante', comprising two articles (I and II). In the first of these articles, the potential of Vis-NIR spectroscopy (430-1040 nm) was studied to predict macronutrients and micronutrients using PLS regression models. This study applied differential treatments for N nutrients (0 %, 33 %, 50 % and 100 %) and K2O (0 %, 50 % and 100 %) of crop demand. The results showed that it is possible to accurately predict macronutrients such as phosphorus (P), calcium (Ca) and magnesium (Mg), with a coefficient of determination in the prediction (R2P) of 0.78 to 0.63. Boron (B) and manganese (Mn) obtained the best micronutrient prediction coefficients, with R2P of 0.79 and 0.69, respectively. The second article evaluated hyperspectral imaging (HSI) in the range (500-980 nm) for nutrient concentration estimation. The results showed the prediction of macronutrients such as nitrogen (N), P, potassium (K), Ca and Mg with R2P from 0.80 to 0.62 and, for micronutrients, only in B, an acceptable value for the estimation was obtained (R2p 0.69). In addition, using the projection influence variable reduction (VIP) method, a reliable prediction was obtained for N (R2P 0.76) and B (R2P 0.61) nutrients. In the second chapter, another emblematic crop of economic importance in the Valencian Community, citrus, was studied. Estimation tools were developed for citrus cv. 'Clementina de Nules' and the results were published in two articles (III and IV). Thus, in the third article, the capacity of spectrometric techniques to determine the concentration of nutrients in a complete vegetative cycle was studied. The results showed prediction with an R2P of 0.70 to 0.65 for P, K Ca and B. Using the weight regression coefficient (BW), a subset of more influential bands was determined for P, K and B nutrients. The results showed that the bands of greatest importance for these nutrients are located in the Vis region (430-750 nm), which is associated with photosynthetic pigment uptake. Finally, in the fourth article, the potential of HSI to discriminate young leaves from leaves of previous vegetative cycles has been studied, which would improve the diagnosis given that the reference tables in this crop are made on leaves of spring sprouting. Subsequently, the prediction of nutrient concentration of young leaves was carried out using 49 spectral bands, obtaining better results for the nutrients P, K, Ca, iron (Fe) and Mn with R2P from 0.69 to 0.60. In addition, these nutrients were predicted by minimizing the number of bands to ten, with the BW and an R2P of 0.67 to 0.57. ¿, Maylin Acosta thanks IFARHU-SENACYT for the Professional Excellence Scholarships, contract No. 270-2021-020. Sandra Munera thanks the Juan de la Cierva-Formación contract (FJC2021-047786-I) co-funded by MCIN/AEI/10.13039/501100011033 and EU NextGenerationEU/PRTR. This work is co-funded by MICIN-AEI through project TED2021-130117B-C31, GVA-IVIA through projects 52203 and 52204, and the European Regional Development Fund (ERDF) of the Generalitat Valenciana 2021–2027.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Firmness classification in &apos, Rojo Brillante&apos, persimmon fruit after cold storage using hyperspectral imaging and deep learning with variational autoencoders

  • Castillo-Gironés, Salvador
  • González, Ángel
  • Martínez-Sober, Marcelino
  • Cubero-García, Sergio
  • Vilhena, Nariane Q.
  • Blanco-Álvarez, Iván
  • Gomez-Sanchis, Juan
  • Munera, S|||0000-0003-3064-1186
[EN] Firmness is one of the most important quality parameters of 'Rojo Brillante' persimmon fruit, and it can be significantly affected during cold storage. Industry requirements demand 40 N at harvest for storage and 20 N for commercialisation. This study evaluated the use of hyperspectral imaging to determine the firmness of persimmon stored under industrial conditions. To achieve this, fruit samples were stored for up to three months under various industrial conditions to produce a wide range of firmness levels. Hyperspectral images of each sample fruit were captured within the 420-1010 nm range, and the firmness was measured by a texturometer. An innovative methodology was applied, combining a convolutional variational autoencoder (VAE) for feature extraction from firmness curves and images and clustering analysis to identify firmness levels. The clustering analysis identified three distinct firmness categories (hard, medium and soft) corresponding to the requirements of the industry. Using six wavelengths, a support vector machine (SVM) classification model accurately assigned 81.3 % of fruit samples to the corresponding categories, thus paving the way for the utilisation of in-line detection using multispectral images. This approach demonstrates the potential of combining hyperspectral imaging and advanced machine learning techniques for non-destructive firmness prediction in 'Rojo Brillante' persimmon fruit quality assessment., This work was partially funded through the project AEI PID2023-150192OR-C31 and C33 with the support of FEDER funds, project AEI TED2021-130117B-C31 and C33, by the project GVA-PROMETEO CIPROM/2021/014 and by the project PID2021-127975OR-C21. Sal-vador Castillo thanks INIA for the FPI-INIA grant number PRE2020-094491, partially supported by European Union FSE funds. Nariane Q. Vilhena thanks the Ministerio de Ciencia e Innovacion of Spain for grant PRE2018-085833. Sandra Munera expresses her thanks for the post-doctoral contract Juan de la Cierva-Formacion (FJC2021-047786-I) co-funded by MCIN/AEI/10.13039/501100011033 and European Union NextGeneration EU/PRTR.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Exploring 3D Reconstruction Methods for Assessing Biometric Parameters in Avocado Trees

  • Rodriguez-Ortega, Alejandro|||0000-0003-4781-6368
  • Rey, Beatriz|||0000-0001-9213-1443
  • Albert Gil, Francisco Eugenio|||0000-0001-9452-614X
  • Aleixos Borrás, María Nuria|||0000-0001-6051-3375
  • Martín, Rodrigo
  • Cubero, Sergio
  • Quiñones, Ana
  • Blasco, Jose
[EN] Within the framework of the current Common Agricultural Policy, resource reduction is pivotal for achieving the goal outlined in the Farm to Fork Strategy by the European Commission. Adequate fertilisation plans are necessary to fulfil these objectives, adapting to plant needs throughout their growth cycle. To establish these needs, quantifying vegetative growth, based on biometric parameters, between key phenological stages is crucial. These parameters can be derived from 3D point clouds acquired via technological systems such as LiDAR or photogrammetry, enabling precise measurements. This work aimed to develop automated and non-destructive methods to assess these biometric parameters in avocado trees at different growth stages from point clouds obtained using two 3D techniques: a LiDAR sensor and photogrammetry. Measurements were conducted in an experimental avocado orchard in Moncada (Valencia), Spain, across three phenological stages from June to November 2023. Automatic measurements were taken using sensors integrated into a scouting robot, and traditional statistical methods were employed to compare manually obtained data with scanned data. Preliminary results demonstrated the progression of biometric parameters measured by LiDAR and photogrammetry across different developmental stages, showing a significant correlation between manual measures and scanned data. Both technologies performed similarly, particularly in measuring tree volume from point clouds. These findings highlight the potential of LiDAR and photogrammetry in accurately assessing biometric parameters in agricultural settings, supporting sustainable farming practices., This work is part of the projects ED2021-130117B-C31 Smart autonomous electrical robot for a digital and sustainable agriculture in the Valencian Community, co-funded by MCIN/AEI/10.13039/501100011033 and UE NextGenerationEU/PRTR (TED2021-130117BC31-C32), and GVA-IVIA 52203 and 52204 co-funded by ERDF of the CV 2021-2027
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Raman spectroscopy for multi-label identification of common apple pesticide mixtures using CNNs and gradient-weighted class activation mapping

  • Castillo-Gironés, Salvador
  • Arnould, Quentin
  • Gomez-Sanchis, Juan
  • Blasco, Jose
  • Pigeon, Oliver
  • Baeten, Vicent
  • Fernandez-Pierna, Juan Antonio
[EN] Fruits on the market must be free of harmful pesticides, but residues often persist, posing a risk to public health. Detecting these residues efficiently is crucial. However, traditional methods are slow and lack real-time capability. Besides, the detection challenge becomes even more difficult when pesticides are used in mixtures, and Raman spectroscopy offers a promising real-time solution for identifying pesticide residues. This study investigates the detection of commonly used pesticide active compounds on apples in complex mixtures via Raman spectroscopy and a multilabel classification approach, addressing the challenge of identifying active compounds within complex mixtures. Models were also tested on independent, entirely different mixtures to assess their limitations and robustness. Unique spectral fingerprints were established for pure active compounds, which enabled differentiation in mixture analyses. Dimensionality reduction techniques (t-SNE and PCA) effectively distinguished pure compounds from mixtures. Three machine learning models: Partial Least Squares (PLS), Support Vector Machine (SVM), and a 1-D Convolutional Neural Network (1-D CNN) were trained to identify active compounds in mixtures. All models classified key compounds with high F1 scores, with Captan detected with F1 scores higher than 99 %, but some compounds, such as Folpet or Mancozeb, showed poor predictions. The 1-D CNN achieved the best performance with the lowest Hamming Loss, and Grad-CAM analysis confirmed it identified relevant spectral regions. This work demonstrates the potential of machine learning-enhanced Raman spectroscopy for effective pesticide monitoring, regulatory compliance, and food safety., This work is funded by GVA-IVIA and FEDER funds by project 52204, project PID2023-150192OR-C31, C32 and C33, project AEI TED2021-130117B-C31 and C33, by the project GVA-PROMETEO CIPROM/2021/014 and by the project PID2021-127975OR-C21. Salvador Castillo thanks INIA for the FPI-INIA grant number PRE2020-094491, partially supported by European Union FSE funds. We thank the Colruyt group for financial support during the project entitled 'Detection of synthetic chemical pesticides by Raman spectroscopy'. We also thank the Protection, Control Products and Residues Unit of the Walloon Agricultural Research Centre (CRA-W) for supplying the samples used in this work.
ReDivia Repositorio Digital del Instituto Valenciano de Investigaciones Agrarias
Publicación Artículo científico (JournalArticle)

Non-Destructive Appraisal of Macro- and Micronutrients in Persimmon Leaves Using Vis/NIR Hyperspectral Imaging

  • Acosta, Maylin
  • Rodríguez-Carretero, Isabel
  • Blasco, José
  • De-Paz, José M.
  • Quinones, Ana
Visible and near-infrared (Vis/NIR) hyperspectral imaging (HSI) was used for rapid and non-destructive determination of macro- and micronutrient contents in persimmon leaves. Hyperspectral images of 687 leaves were acquired in the 500–980 nm range over 6 months, covering a complete vegetative cycle. The average reflectance spectrum of each leaf was extracted, and foliar ionomic analysis was used as a reference method to determine the actual concentration of the nutrients in the leaves. Analyses were performed via emission spectrometry (ICP-OES) for macro- and micronutrients after microwave digestion and using the Kjeldahl method to quantify nitrogen. Partial least square regression (PLS-R) was used to predict the nutrient concentration based on spectral data from the leaf using actual values of each element as predictor variables. Several methods were used to pre-process the spectra, including Savitzky–Golay (SG) smoothing, standard normal variate (SNV) and first (1D) and second derivatives (2D). Seventy-five percent of the samples were used to calibrate and validate the model by cross-validation, whereas the remaining twenty-five % were used as an independent test set. The best performance of the models for the test set achieved an R2 = 0.80 for nitrogen. Results were also satisfactory for phosphorous, calcium, magnesium and boron, with determination coefficient R2 values of 0.63, 0.66, 0.58 and 0.69, respectively. For the other nutrients, lower prediction rates were attained (R2 = 0.48 for potassium, R2 = 0.38 for iron, R2 = 0.24 for copper, R2 = 0.23 for zinc and R2 = 0.22 for manganese). The variable importance in projection (VIP) was used to extract the most influential bands for the best-predicted nutrients, which were N, K and B.