Aerial Identification of Fruit Maturity in Amazonian Palms via Plant-Canopy Modeling
Abstract
:1. Introduction
Literature Review
2. Materials and Methods
- Module B1 is responsible for acquiring the multispectral images, which include five bands (R, G, B, RE, and NIR) as well as the RGB image.
- In Module B2, a palm identification model is utilized to segment and extract the region of interest (RoI) from each band.
- Module B3 conducts a temporal variability analysis and modeling of each feature to understand their response and correlation with the maturity stages of the fruit. The resulting data are structured and appropriately labeled.
- Module B4 is responsible for training, validating, and testing the ML models for identifying fruit maturity stages through correlations.
- In Module B5, the segmentation steps are performed using the ML model, feature extraction is carried out on the RoI, and fruit maturity is estimated by correlating it with the reflectance of the canopy.
2.1. Protocol for Image Acquisition
- General conditions: First, palms were selected by their genus and species and the location of each georeference. Then, palms were physically marked with reflective tape to differentiate between genders. Next, imagery was captured around each palm at a distance of less than three meters, with the UAV controlled manually following a polygon waypoint path. A dataset was created from the acquired images, along with the palm, fruit and weather conditions. To ensure consistency in the physiological state between the fruits and the canopy, days with similar illumination and solar radiation were selected for the UAV mission. A weekly flight was conducted for each variety, with local time between 9–11 a.m. and 3–5 p.m.
- Canopy level: The minimum flight altitude was determined at 60 m. Lower altitudes could result in partial or total exclusion of the palm due to geolocation precision and climatic factors.
- Fruit level: In order to capture images of the fruit clusters on each palm, manual flights were conducted, ensuring that sufficient space was maintained around each palm to maneuver the UAV and avoid collisions. Images should be captured of each cluster from multiple angles, distances, and heights, with a focus on capturing images as closely as possible. Additionally, images of the fruits and inflorescence were captured. The canopy around the palm and at the top should also be captured at various heights and distances.
- Segmentation (RoI): The region of interest (RoI) was extracted by applying a Mask R-CNN-based algorithm for object identification developed by the authors in previous work reported in [36]. Since the object detector of the original algorithms works with RGB images, a sub-process was designed for the extraction of the RoI to all spectral bands.
2.2. Segmentation and Extraction of RoI
2.3. Feature Extraction
2.3.1. Vegetation Indices
2.3.2. Graphs
2.3.3. Convolutional Neural Networks (CNNs)
2.4. Data Modeling and Classification
3. Results
3.1. Data Collection
3.2. Feature Extraction and Data Modeling
3.3. Estimation of Maturity State through Correlations
4. Discussion
5. Conclusions
Author Contributions
Funding
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Maciel, E.A.; Martins, F.R. Rarity patterns and the conservation status of tree species in South American savannas. Flora Morphol. Distrib. Funct. Ecol. Plants 2021, 285, 151942. [Google Scholar] [CrossRef]
- Navarro-Cruz, A.R.; Lazcano-Hernández, M.; Vera-López, O.; Kammar-García, A.; Segura-Badilla, O.; Aguilar-Alonso, P.; Pérez-Fernández, M.S. Mauritia flexuosa L. f. In Fruits of the Brazilian Cerrado; Springer: Cham, Switzerland, 2021; pp. 79–98. [Google Scholar] [CrossRef]
- Velarde V., M.J.; Moraes R., M. Densidad de individuos adultos y producci ón de frutos del asaí (Euterpe precatoria, Arecaceae) en Riberalta, Bolivia. Ecol. Boliv. 2008, 43, 99–110. [Google Scholar]
- Quintero-Angel, M.; Martínez-Girón, J.; Orjuela-Salazar, S. Agroindustrial valorization of the pulp and peel, seed, flour, and oil of moriche (Mauritia flexuosa) from the Bita River, Colombia: A potential source of essential fatty acids. Biomass Convers. Biorefinery 2022, 1, 1–9. [Google Scholar] [CrossRef] [PubMed]
- SINCHI. Fichas Palmas Amazónicas con Potencial Seje, Moriche y Asaí; SINCHI: Bogota, Colombia, 2018; ISBN 978-958-5427-05-1. Available online: https://www.sinchi.org.co/seje-moriche-y-asai-palmas-amazonicas-con-potencial (accessed on 6 June 2023).
- Ponce, M.E. Patrones de caída de frutos en Mauritia Flexuosa L.F. y fauna involucrada en los procesos de remoción de semillas. Acta Botán. Venez. 2002, 25, 119–142. [Google Scholar]
- van der Hoek, Y.; Solas, S.Á.; Peñuela, M.C. The palm Mauritia flexuosa, a keystone plant resource on multiple fronts. Biodivers. Conserv. 2019, 28, 539–551. [Google Scholar] [CrossRef]
- Hernández, M.S.; Castro, S.Y.; Giraldo, B.; Barrera, J. Seje, Moriche, Asaí: Palmas Amazónicas con Potencial, 1st ed.; Diana Patricia Mora Rodríguez: Bogotá, Colombia, 2018; p. 123. [Google Scholar]
- Cárdenas López, D.; Arias, G.J.C. Manual de Identificación, Selección y Evaluación de Oferta de Productos Forestales no Maderables; SINCHI: Bogota, Colombia, 2007; p. 32. ISBN 978-958-8317-22-9. Available online: https://sinchi.org.co/manual-de-identificacion-seleccion-y-evaluacion-de-oferta-de-productos-forestales-no-maderables (accessed on 6 June 2023).
- Ponce-Corona, E.; Guadalupe Sánchez, M.; Fajardo-Delgado, D.; Acevedo-Juárez, B.; De-La-Torre, M.; Avila-George, H.; Castro, W. A systematic review of the literature focused on the use of unmanned aerial vehicles during the vegetation detection process. RISTI-Rev. Iber. Sist. Tecnol. Inf. 2020, 2020, 82–101. [Google Scholar] [CrossRef]
- Coelho Eugenio, F.; Badin, T.L.; Fernandes, P.; Mallmann, C.L.; Schons, C.; Schuh, M.S.; Soares Pereira, R.; Fantinel, R.A.; Pereira da Silva, S.D. Remotely Piloted Aircraft Systems (RPAS) and machine learning: A review in the context of forest science. Int. J. Remote Sens. 2021, 42, 8207–8235. [Google Scholar] [CrossRef]
- Tian, H.; Wang, T.; Liu, Y.; Qiao, X.; Li, Y. Computer vision technology in agricultural automation—A review. Inf. Process. Agric. 2020, 7, 1–19. [Google Scholar] [CrossRef]
- Azcón-Bieto, J.; Talón, M. Fundamentos de fisiología vegetal. arXiv 2003, arXiv:1011.1669v3. [Google Scholar]
- Montero, I.M.; Barrera, J.A.; Giraldo, B.; Lucena, A. Fichas Tecnicas de Especies de uso Forestal y Agroforestal de la Amazonia Colombiana; Instituto Amazónico de Investigaciones Científicas SINCHI: Bogotá, Colombia, 2016; Available online: https://sinchi.org.co/index.php/fichas-tecnicas-de-especies-de-uso-forestal-y-agroforestal-de-la-amazonia-colombiana1 (accessed on 6 June 2023).
- Wünsche, J.N.; Greer, D.H.; Laing, W.A.; Palmer, J.W. Physiological and biochemical leaf and tree responses to crop load in apple. Tree Physiol. 2005, 25, 1253–1263. [Google Scholar] [CrossRef]
- Tekalign, T.; Hammes, P.S. Growth and productivity of potato as influenced by cultivar and reproductive growth: I. Stomatal conductance, rate of transpiration, net photosynthesis, and dry matter production and allocation. Sci. Hortic. 2005, 105, 13–27. [Google Scholar] [CrossRef]
- dos Santos, M.R.; Martinez, M.A.; Donato, S.L.; Coelho, E.F. ‘Tommy Atkins’ mango yield and photosynthesis under water deficit in semiarid region of Bahia. Rev. Bras. Eng. Agric. Ambient. 2014, 18, 899–907. [Google Scholar] [CrossRef] [Green Version]
- Larramendi, L.A.R.; Hernández, F.G.; Castro, H.G.; Flores, M.F.; Castañeda, J.C.G.; Ruiz, R.P. Anatomía foliar relacionada con la ruta fotosintética en árboles de café (Coffea arabica L., var. Caturra Rojo) expuestos a diferentes niveles de radiación solar en la Sierra Maestra, Granma, Cuba. Acta Agron. 2016, 65, 248. [Google Scholar] [CrossRef]
- González-Pérez, J.S.; Becerril-Román, A.; Quevedo-Nolasco, A.; Velasco-Cruz, C.; Jaén-Contreras, J. Área Foliar y Eficiencia en el uso de la Radiación del Ciruelo Japonés. Agro Product. 2018, 11, 19–26. [Google Scholar] [CrossRef]
- Alcantara Cortes, J.S.; Acero Godoy, J.; Alcántara Cortés, J.D.; Sánchez Mora, R.M. Universidad Francisco de Paula Santander. Av Gran Colombia # 12 E – 96 Barrio Colsag. Cúcuta, Colombia. Enfermera. Magister en Enfermería. Nova 2019, 17, 109–129. [Google Scholar] [CrossRef] [Green Version]
- Barrera, J.; Carrillo, M.; Pena, L.; Cardona, J.; Orjuela, N.; Garcia, L.; Castro, S.; Hernandez, M.S.; Martinez, O.; Giraldo, B. Growth and development of moriche fruits (Mauritia flexuosa) in Guaviare, Colombia. In Proceedings of the XXX International Horticultural Congress IHC2018: VII International Symposium on Tropical and Subtropical Fruits, Avocado, II International Symposium on Jackfruit and Other Moraceae and II International Symposium on Date Palm, Istanbul, Turkey, 12–16 August 2018; International Society for Horticultural Science: Leuven, Belgium, 2020; Volume 1299, pp. 141–146. [Google Scholar] [CrossRef]
- Liu, R.; Shang, R.; Liu, Y.; Lu, X. Global evaluation of gap-filling approaches for seasonal NDVI with considering vegetation growth trajectory, protection of key point, noise resistance and curve stability. Remote Sens. Environ. 2017, 189, 164–179. [Google Scholar] [CrossRef] [Green Version]
- Cen, H.; Wan, L.; Zhu, J.; Li, Y.; Li, X.; Zhu, Y.; Weng, H.; Wu, W.; Yin, W.; Xu, C.; et al. Dynamic monitoring of biomass of rice under different nitrogen treatments using a lightweight UAV with dual image-frame snapshot cameras. Plant Methods 2019, 15, 32. [Google Scholar] [CrossRef] [PubMed]
- Zhang, X.; Zhang, F.; Qi, Y.; Deng, L.; Wang, X.; Yang, S. New research methods for vegetation information extraction based on visible light remote sensing images from an unmanned aerial vehicle (UAV). Int. J. Appl. Earth Obs. Geoinf. 2019, 78, 215–226. [Google Scholar] [CrossRef]
- Ma, X.; Feng, J.; Guan, H.; Liu, G. Prediction of chlorophyll content in different light areas of apple tree canopies based on the color characteristics of 3d reconstruction. Remote Sens. 2018, 10, 429. [Google Scholar] [CrossRef] [Green Version]
- Wan, L.; Li, Y.; Cen, H.; Zhu, J.; Yin, W.; Wu, W.; Zhu, H.; Sun, D.; Zhou, W.; He, Y. Combining UAV-based vegetation indices and image classification to estimate flower number in oilseed rape. Remote Sens. 2018, 10, 1484. [Google Scholar] [CrossRef] [Green Version]
- Rouse, J.W.; Riter, S. Erts experiments compiled. IEEE Trans. Geosci. Electron. 1973, 11, 3–76. [Google Scholar] [CrossRef]
- Smith, A.M.; Bourgeois, G.; Teillet, P.M.; Freemantle, J.; Nadeau, C. A comparison of NDVI and MTVI2 for estimating LAI using CHRIS imagery: A case study in wheat. Can. J. Remote Sens. 2008, 34, 539–548. [Google Scholar] [CrossRef]
- Jimenez-Sierra, D.A.; Benítez-Restrepo, H.D.; Vargas-Cardona, H.D.; Chanussot, J. Graph-based data fusion applied to: Change detection and biomass estimation in rice crops. Remote Sens. 2020, 12, 2683. [Google Scholar] [CrossRef]
- Elarab, M.; Ticlavilca, A.M.; Torres-Rua, A.F.; Maslova, I.; McKee, M. Estimating chlorophyll with thermal and broadband multispectral high resolution imagery from an unmanned aerial system using relevance vector machines for precision agriculture. Int. J. Appl. Earth Obs. Geoinf. 2015, 43, 32–42. [Google Scholar] [CrossRef] [Green Version]
- Gitelson, A.A.; Viña, A.; Ciganda, V.; Rundquist, D.C.; Arkebauer, T.J. Remote estimation of canopy chlorophyll content in crops. Geophys. Res. Lett. 2005, 32, L08403. [Google Scholar] [CrossRef] [Green Version]
- Qu, Y.; Clausen, A.; Jørgensen, B.N. Application of Deep Neural Network on Net Photosynthesis Modeling. In Proceedings of the IEEE International Conference on Industrial Informatics (INDIN), Palma de Mallorca, Spain, 21–23 July 2021; Institute of Electrical and Electronics Engineers Inc.: Piscataway, NJ, USA, 2021; Volume 2021. [Google Scholar] [CrossRef]
- Barbedo, J.G.; Tibola, C.S.; Fernandes, J.M. Detecting Fusarium head blight in wheat kernels using hyperspectral imaging. Biosyst. Eng. 2015, 131, 65–76. [Google Scholar] [CrossRef]
- Vélez, S.; Ariza-Sentìs, M.; Velente, J. Mapping the spatial variability of botrytis bunch rot risk in vineyards using UAV multispectral imagery. Eur. J. Agron. 2020, 142, 126691. [Google Scholar] [CrossRef]
- Páscoa, R.N.; Lopo, M.; Teixeira dos Santos, C.A.; Graça, A.R.; Lopes, J.A. Exploratory study on vineyards soil mapping by visible/near-infrared spectroscopy of grapevine leaves. Comput. Electron. Agric. 2016, 127, 15–25. [Google Scholar] [CrossRef]
- Marin, W.; Mondragon, I.F.; Colorado, J.D. Aerial Identification of Amazonian Palms in High-Density Forest Using Deep Learning. Forests 2022, 13, 655. [Google Scholar] [CrossRef]
- Devia, C.A.; Rojas, J.P.; Petro, E.; Martinez, C.; Mondragon, I.F.; Patino, D.; Rebolledo, M.C.; Colorado, J. High-Throughput Biomass Estimation in Rice Crops Using UAV Multispectral Imagery. J. Intell. Robot. Syst. Theory Appl. 2019, 96, 573–589. [Google Scholar] [CrossRef]
- Xu, W.; Lan, Y.; Li, Y.; Luo, Y.; He, Z. Classification method of cultivated land based on uav visible light remote sensing. Int. J. Agric. Biol. Eng. 2019, 12, 103–109. [Google Scholar] [CrossRef] [Green Version]
- García-cervigón, D.; José, J. Estudio de Índices de Vegetación a Partir de Imágenes Aéreas Tomadas Desde UAS/RPAS y Aplicaciones de estos a la Agricultura de Precisión. Master’s Thesis, Technical University of Madrid, Madrid, Spain, 2015; p. 78. [Google Scholar]
- Du, M.; Noguchi, N. Monitoring of wheat growth status and mapping of wheat yield’s within-field spatial variations using color images acquired from UAV-camera System. Remote Sens. 2017, 9, 289. [Google Scholar] [CrossRef] [Green Version]
- Kanke, Y.; Tubaña, B.; Dalen, M.; Harrell, D. Evaluation of red and red-edge reflectance-based vegetation indices for rice biomass and grain yield prediction models in paddy fields. Precis. Agric. 2016, 17, 507–530. [Google Scholar] [CrossRef]
- Naito, H.; Ogawa, S.; Valencia, M.O.; Mohri, H.; Urano, Y.; Hosoi, F.; Shimizu, Y.; Chavez, A.L.; Ishitani, M.; Selvaraj, M.G.; et al. Estimating rice yield related traits and quantitative trait loci analysis under different nitrogen treatments using a simple tower-based field phenotyping system with modified single-lens reflex cameras. ISPRS J. Photogramm. Remote Sens. 2017, 125, 50–62. [Google Scholar] [CrossRef]
- Prabhakara, K.; Dean Hively, W.; McCarty, G.W. Evaluating the relationship between biomass, percent groundcover and remote sensing indices across six winter cover crop fields in Maryland, United States. Int. J. Appl. Earth Obs. Geoinf. 2015, 39, 88–102. [Google Scholar] [CrossRef] [Green Version]
- Gnyp, M.L.; Miao, Y.; Yuan, F.; Ustin, S.L.; Yu, K.; Yao, Y.; Huang, S.; Bareth, G. Hyperspectral canopy sensing of paddy rice aboveground biomass at different growth stages. Field Crops Res. 2014, 155, 42–55. [Google Scholar] [CrossRef]
- Morris, O.J.; Constantinides, A.G. Graph theory for image analysis: An approach based on the shortest spanning tree. IEE Proc. F Commun. Radar Signal Process. 1986, 133, 146–152. [Google Scholar] [CrossRef]
- Cousty, J.; Najman, L.; Dias, F.; Serra, J. Morphological filtering on graphs. Comput. Vis. Image Underst. 2013, 117, 370–385. [Google Scholar] [CrossRef] [Green Version]
- Farkas, I.J.; Derényi, I.; Barabási, A.L.; Vicsek, T. Spectra of “real-world” graphs: Beyond the semicircle law. Phys. Rev. E-Stat. Phys. Plasmas Fluids Relat. Interdiscip. Top. 2001, 64, 12. [Google Scholar] [CrossRef] [Green Version]
- Fowlkes, C.; Belongie, S.; Chung, F.; Malik, J. Spectral Grouping Using the Nyström Method. IEEE Trans. Pattern Anal. Mach. Intell. 2004, 26, 214–225. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Naranjo-Torres, J.; Mora, M.; Hernández-García, R.; Barrientos, R.J.; Fredes, C.; Valenzuela, A. A review of convolutional neural network applied to fruit image processing. Appl. Sci. 2020, 10, 3443. [Google Scholar] [CrossRef]
VI | Description | Mathematical Expression |
---|---|---|
GRVI | Red Vegetation Index [37] | |
RGVI | Green Vegetation Index [25] | |
MGRVI | Modified Red Vegetation Index [37] | |
NGRVI | Reciprocal transformation based on MGRVI normalization [24] | |
GRBVI | Normalized Green-Red Difference Index [38] | |
VARI | Visible Atmospherically Resistant Index [37] | |
RG | Red Green Ratio [39] | |
GR | Red Green Ratio Index [23] | |
VDVI | Visible-Band Difference Vegetation Index [38] | |
EXG | Excess green [38] | |
CIVE | Color Index of Vegetation [26] | |
NGBDI | Normalized Green-Blue Difference Index [40] | |
NRBDI | Normalized Red-Blue Difference Index [25] | |
RGBVI | Red, Blue-Green Vegetation Index [25] | |
RBGVI | Red, Blue-Green Vegetation Index [25] | |
NIRG | Green model [31] | |
NIRRE | Red border model [31] | |
NDVI | Normalized Difference Vegetation Index [27] | |
RVI | Ratio Vegetation Index [41] | |
DVI | Difference Vegetation Index [42] | |
GNDVI | Green NDVI [43] | |
CTVI | Corrected Transformed Vegetation Index [42] | |
SAVI | Soil-Adjusted Vegetation Index [43] | |
MSAVI | Modified SAVI [44] | |
NBVI | Green NDVI |
Plot | Flights | UAV Waypoints | Images |
---|---|---|---|
Plot 1 | 30 | 18 | 540 |
Plot 2 | 30 | 51 | 1530 |
Characteristic | Coefficient | Ratio |
---|---|---|
NRBDI | 0.482360 | Moderate |
RBGVI | 0.465781 | Moderate |
RG | 0.284647 | Low |
NIRG | 0.268070 | Low |
GNDVI | 0.248591 | Low |
NB | 0.224336 | Low |
RGVI | 0.178566 | Low |
Grafo | 0.170969 | Low |
RGBVI | 0.155478 | Low |
RVI | 0.117775 | Low |
MGRVI | 0.110696 | Low |
NGRVI | 0.110696 | Low |
GRVI | −0.178566 | Low negative |
VARI | −0.274951 | Low negative |
Dataset 1 | Dataset 2 | Dataset 3 | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Class | FP | TP | Total | Class | FP | TP | Total | Class | FP | TP | Total |
VC | 26 | 24 | 50 | VC | 34 | 16 | 50 | VC | 31 | 19 | 50 |
MC | 21 | 29 | 50 | MC | 12 | 38 | 50 | MC | 21 | 29 | 50 |
Total | 47 | 53 | 100 | Total | 46 | 54 | 100 | Total | 52 | 48 | 100 |
Accuracy | 47% | 53% | 100% | Accuracy | 46% | 54% | 100% | Accuracy | 52% | 48% | 100% |
Model | Accuracy | F1 Score | Time |
---|---|---|---|
RidgeClassifier | 0.7 | 0.7 | 0.02 |
LinearSVC | 0.7 | 0.7 | 0.09 |
CalibratedClassifierCV | 0.7 | 0.7 | 0.33 |
LogisticRegression | 0.7 | 0.7 | 0.04 |
RidgeClassifierCV | 0.69 | 0.69 | 0.02 |
LinearDiscriminantAnalysis | 0.68 | 0.68 | 0.02 |
NuSVC | 0.66 | 0.66 | 0.12 |
Perceptron | 0.66 | 0.66 | 0.02 |
QuadraticDiscriminantAnalysis | 0.66 | 0.66 | 0.02 |
SVC | 0.66 | 0.66 | 0.07 |
AdaBoostClassifier | 0.64 | 0.64 | 0.19 |
ExtraTreesClassifier | 0.64 | 0.64 | 0.21 |
LGBMClassifier | 0.64 | 0.64 | 0.11 |
RandomForestClassifier | 0.64 | 0.64 | 0.34 |
XGBClassifier | 0.63 | 0.63 | 0.11 |
SGDClassifier | 0.63 | 0.63 | 0.02 |
PassiveAggressiveClassifier | 0.61 | 0.58 | 0.01 |
KNeighborsClassifier | 0.62 | 0.62 | 0.04 |
BaggingClassifier | 0.61 | 0.61 | 0.12 |
LabelSpreading | 0.60 | 0.60 | 0.10 |
LabelPropagation | 0.60 | 0.60 | 0.08 |
ExtraTreeClassifier | 0.54 | 0.54 | 0.01 |
DecisionTreeClassifier | 0.54 | 0.54 | 0.03 |
GaussianNB | 0.55 | 0.52 | 0.01 |
NearestCentroid | 0.54 | 0.54 | 0.01 |
BernoulliNB | 0.54 | 0.53 | 0.01 |
DummyClassifier | 0.46 | 0.46 | 0.01 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Marin, W.; Mondragon, I.F.; Colorado, J.D. Aerial Identification of Fruit Maturity in Amazonian Palms via Plant-Canopy Modeling. Remote Sens. 2023, 15, 3752. https://doi.org/10.3390/rs15153752
Marin W, Mondragon IF, Colorado JD. Aerial Identification of Fruit Maturity in Amazonian Palms via Plant-Canopy Modeling. Remote Sensing. 2023; 15(15):3752. https://doi.org/10.3390/rs15153752
Chicago/Turabian StyleMarin, Willintong, Ivan F. Mondragon, and Julian D. Colorado. 2023. "Aerial Identification of Fruit Maturity in Amazonian Palms via Plant-Canopy Modeling" Remote Sensing 15, no. 15: 3752. https://doi.org/10.3390/rs15153752