Olive-Fruit Mass and Size Estimation Using Image Analysis and Feature Modeling
<p>Example of image captured of the A1 set.</p> "> Figure 2
<p>Step by step illustration of the image analysis algorithm on a sub-image of the study set: (<b>a</b>) inverted value channel; (<b>b</b>) saturation channel; (<b>c</b>) combined channel of saturation and value; (<b>d</b>) background estimation; (<b>e</b>) high contrasted image; (<b>f</b>) segmented image; (<b>g</b>) final postprocessed segmentation.</p> "> Figure 3
<p>Result of the segmentation of the image of the A1 set, originally shown in <a href="#sensors-18-02930-f001" class="html-fig">Figure 1</a>.</p> "> Figure 4
<p>Examples of the different categories of pixels established to evaluate the segmentation results: <span class="html-italic">true positives</span> (TP, in blue); <span class="html-italic">false positives</span> (FP, in red); <span class="html-italic">false negatives</span> (FN, in green).</p> "> Figure 5
<p>Correlation study performed for the Arbequina variety, considering the three different sizing features of the fruits the experiment is focused on: The major axis (<b>a</b>), minor axis (<b>b</b>) and mass (<b>c</b>).</p> "> Figure 6
<p>Correlation study performed for the Picual variety, considering the three different sizing features of the fruits the experiment is focused on: The major axis (<b>a</b>), minor axis (<b>b</b>) and mass (<b>c</b>).</p> "> Figure 7
<p>Correlation study of variety-independent model trained on the instances from both Arbequina and Picual varieties, considering the three different targeted features: The major axis (<b>a</b>), minor axis (<b>b</b>) and mass (<b>c</b>).</p> ">
Abstract
:1. Introduction
2. Materials and Methods
2.1. Reference Data and Image Acquisition
- a KERN PCB 3500-2 precision balance (KERN & Sohn GmbH, Balingen, Germany).
- a 0.01 mm-resolution 0.02 mm-accuracy Electronic Digital Vernier Caliper.
2.2. Image Analysis and Segmentation
2.2.1. Preprocessing
2.2.2. Image Segmentation
2.2.3. Postprocessing
2.3. Estimation Model Training
3. Results and Discussion
3.1. Evaluation of the Image Analysis Algorithm
- TP: Those foreground/olive pixels in the segmented image (white pixels) matching with their analogue ones in the corresponding ground-truth image (they keep being white pixels).
- FP: Those foreground/olive pixels in the segmented image (white pixels) that were labeled as background (black pixels) in the corresponding ground-truth image.
- FN: Those background pixels in the segmented image (white pixels) that were labeled as foreground/olive (white pixels) in the corresponding ground-truth image.
3.2. Results of the Image Analysis Algorithm
3.3. Evaluation of the Estimation Models
- Root-Mean-Square Error:
- Relative Root-Mean-Square Error expressed as percentage
- Relative Mean Error expressed as percentage
3.4. Results of the Estimation of Olive Features
4. Conclusions
Author Contributions
Funding
Conflicts of Interest
References
- Vossen, P. Olive Oil: History, Production, and Characteristics of the World’s Classic Oils. Hortic. Sci. 2007, 42, 1093–1100. [Google Scholar]
- Su, C.; Sun, J.; Zhu, W.; Peng, L. History, Distribution, and Potential of the Olive Industry in China: A Review. Sustainability 2018, 10, 1426. [Google Scholar] [CrossRef]
- International Olive Oil Council (IOC). Updates Series of World Statistics on Production, Imports, Exports and Consumption. 2018. Available online: http://www.internationaloliveoil.org/estaticos/view/131-world-olive-oil-figures (accessed on 18 June 2018).
- United States Department of Agriculture, Foreign Agricultural Service. Oilseeds: World Markets and Trade. Available online: https://apps.fas.usda.gov/psdonline/circulars/oilseeds.pdf (accessed on 19 June 2018).
- Russo, C.; Cappelletti, G.M.; Nicoletti, G.M.; Di Noia, A.E.; Michalopoulos, G. Comparison of European olive production systems. Sustainability 2016, 8, 825. [Google Scholar] [CrossRef]
- Connor, D.J.; Gómez-del-Campo, M.; Rousseaux, M.C.; Searles, P.S. Structure, management and productivity of hedgerow olive orchards: A review. Sci. Hort. 2014, 169, 71–93. [Google Scholar] [CrossRef]
- Moreda, G.P.; Ortiz-Cañavate, J.; García-Ramos, F.J.; Ruiz-Altisent, M. Non-destructive technologies for fruit and vegetable size determination—A review. J. Food Eng. 2009, 92, 119–136. [Google Scholar] [CrossRef] [Green Version]
- Sánchez Gómez, A.H.; García García, P.; Rejano Navarro, L. Elaboration of table olives. Grasas Aceites 2006, 57, 86–94. [Google Scholar] [CrossRef]
- Kailis, S.; Harris, D. Producing Table Olives; Landlinks Press: Collingwood, Australia, 2007; pp. 173–174, 244–246. ISBN 978-0-643092-03-7. [Google Scholar]
- Aquino, A.; Millan, B.; Diago, M.P.; Tardaguila, J. Automated early yield prediction in vineyards from on-the-go image acquisition. Comput. Electron. Agric. 2018, 144, 26–36. [Google Scholar] [CrossRef]
- Mery, D.; Pedreschi, F. Segmentation of colour food images using a robust algorithm. J. Food Eng. 2005, 66, 353–360. [Google Scholar] [CrossRef]
- Cervantes, J.; Taltempa, J.; García-Lamont, F.; Castilla, J.S.R.; Rendon, A.Y.; Jalili, L.D. Comparative Analysis of the Techniques Used in a Recognition System of Plant Leaves. RIAI—Rev. Iberoam. Autom. Inform. Ind. 2017, 14, 104–114. [Google Scholar] [CrossRef]
- Zhang, Y.; Wu, L. Classification of fruits using computer vision and a multiclass support vector machine. Sensors 2012, 12, 12489–12505. [Google Scholar] [CrossRef] [PubMed]
- Sa’ad, F.S.A.; Ibrahim, M.F.; Shakaff, A.Y.; Zakaria, A.; Abdullah, M.Z. Shape and weight grading of mangoes using visible imaging. Comput. Electron. Agric. 2015, 115, 51–56. [Google Scholar] [CrossRef]
- Mizushima, A.; Lu, R. An image segmentation method for apple sorting and grading using support vector machine and Otsu’s method. Comput. Electron. Agric. 2013, 94, 29–37. [Google Scholar] [CrossRef]
- Omid, M.; Khojastehnazhand, M.; Tabatabaeefar, A. Estimating volume and mass of citrus fruits by image processing technique. J. Food Eng. 2010, 100, 315–321. [Google Scholar] [CrossRef]
- Diaz, R.; Faus, G.; Blasco, M.; Blasco, J.; Molto, E. The application of a fast algorithm for the classification of olives by machine vision. Food Res. Int. 2000, 33, 305–309. [Google Scholar] [CrossRef]
- Beyaz, A.; Taha, M.; Duygu, İ. Identification of some Spanish olive cultivars using image processing techniques. Sci. Hortic. 2017, 225, 286–292. [Google Scholar] [CrossRef]
- Piras, F.; Grillo, O.; Venora, G.; Lovicu, G.; Campus, M.; Bacchetta, G. Effectiveness of a computer vision technique in the characterization of wild and farmed olives. Comput. Electron. Agric. 2016, 122, 86–93. [Google Scholar] [CrossRef]
- Gatica, G.; Best, S.; Ceroni, J.; Lefranc, G. Olive Fruits Recognition Using Neural Networks. Procedia Comput. Sci. 2013, 17, 412–419. [Google Scholar] [CrossRef]
- Sonka, M.; Hlavac, V.; Boyle, R. Image Processing, Analysis, and Machine Vision, 4th ed.; Cengage Learning: Belmont, CA, USA, 2014; pp. 36–38. ISBN 978-1-133-59369-0. [Google Scholar]
- Connolly, C.; Fliess, T.A. Study of efficiency and accuracy in the transformation from RGB to CIELAB color space. IEEE Trans. Image Process. 1997, 6, 1046–1047. [Google Scholar] [CrossRef] [PubMed]
- Soille, P. Morphological Image Analysis—Principles and Applications, 2nd ed.; Springer: Berlin, Germany, 2013; pp. 63–97. ISBN 978-3-642-07696-1. [Google Scholar]
- Otsu, N. A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cyb. 1979, 9, 62–66. [Google Scholar] [CrossRef]
- Tukey, J.W. Comparing individual means in the analysis of variance. Biometrics 1949, 99–114. [Google Scholar] [CrossRef]
Set/Image | RC | PC | F-Score |
Arbequina | |||
A1 | 0.9614 | 0.9372 | 0.9491 |
A2 | 0.9545 | 0.9432 | 0.9488 |
A3 | 0.9551 | 0.9535 | 0.9543 |
A4 | 0.9536 | 0.9745 | 0.9639 |
A5 | 0.9510 | 0.9510 | 0.9510 |
Overall | 0.9551 | 0.9519 | 0.9534 |
Picual | |||
P1 | 0.9464 | 0.9810 | 0.9634 |
P2 | 0.9414 | 0.9922 | 0.9661 |
P3 | 0.9380 | 0.9869 | 0.9618 |
P4 | 0.9316 | 0.9967 | 0.9631 |
Overall | 0.9393 | 0.9892 | 0.9636 |
Arbequina + Picual | |||
Overall | 0.9481 | 0.9685 | 0.9580 |
Arbequina Validation Set (N = 150) | ||||
Feature | Estimation Model | RMSE | SE (%) | (%) |
Major axis | Specific | 0.4885 (mm) | 3.46 | 0.86 |
Variety-independent | 0.5778 (mm) | 4.09 | 0.14 | |
Minor axis | Specific | 0.6007 (mm) | 4.99 | 0.09 |
Variety-independent | 0.7811 (mm) | 6.49 | 2.39 | |
Mass | Specific | 0.1220 (g) | 9.62 | 0.78 |
Variety-independent | 0.1775 (g) | 13.99 | 1.51 | |
Picual Validation Set (N = 150) | ||||
Feature | Estimation Model | RMSE | SE (%) | (%) |
Major axis | Specific | 0.4163 (mm) | 1.98 | 0.03 |
Variety-independent | 0.4770 (mm) | 2.27 | 0.60 | |
Minor axis | Specific | 0.6804 (mm) | 4.38 | 0.29 |
Variety-independent | 0.8036 (mm) | 5.17 | 1.53 | |
Mass | Specific | 0.250 (g) | 7.89 | 2.39 |
Variety-independent | 0.2439 (g) | 7.69 | 1.65 |
Specific Estimation Models (N = 150) | ||
Feature | Arbequina | Picual |
Major axis | (0.9921, 0.0344) a | (1.0005, 0.0199) a’ |
Minor axis | (1.0026, 0.0502) b | (1.004, 0.043) b |
Mass | (0.9985, 0.0946) c | (1.0275, 0.0741) c’ |
Variety-Independent Estimation Models (N = 150) | ||
Feature | Arbequina | Picual |
Major axis | (1.0011, 0.042) a | (0.9936, 0.0221) a |
Minor axis | (1.0246, 0.0611) b | (0.9847, 0.0486) b’ |
Mass | (1.0068, 0.1428) c | (0.9851, 0.0731) c |
© 2018 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 (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Ponce, J.M.; Aquino, A.; Millán, B.; Andújar, J.M. Olive-Fruit Mass and Size Estimation Using Image Analysis and Feature Modeling. Sensors 2018, 18, 2930. https://doi.org/10.3390/s18092930
Ponce JM, Aquino A, Millán B, Andújar JM. Olive-Fruit Mass and Size Estimation Using Image Analysis and Feature Modeling. Sensors. 2018; 18(9):2930. https://doi.org/10.3390/s18092930
Chicago/Turabian StylePonce, Juan Manuel, Arturo Aquino, Borja Millán, and José Manuel Andújar. 2018. "Olive-Fruit Mass and Size Estimation Using Image Analysis and Feature Modeling" Sensors 18, no. 9: 2930. https://doi.org/10.3390/s18092930