Browsing by Author "Costa, Lino"
Now showing 1 - 10 of 15
Results Per Page
Sort Options
- An Artificial Intelligence-Based Method to Identify the Stage of Maturation in Olive Oil MillsPublication . Mendes, João; Lima, José; Costa, Lino; Rodrigues, Nuno; Leitão, Paulo; Pereira, Ana I.Identifying the maturation stage is an added value for olive oil producers and consumers, whether this is done to predict the best harvest time, give us more information about the olive oil, or even adapt techniques and extraction parameters in the olive oil mill. In this way, the proposed work presents a new method to identify and count the number of olives that enter the mill as well as their stage of maturation. It is based on artificial intelligence (AI) and deep learning algorithms, using the two most recent versions of YOLO, YOLOv7 and YOLOv8. The obtained results demonstrate the possibility of using this type of application in a real environment, managing to obtain a mAP of approximately 79% with YOLOv8 in the five maturation stages, with a processing rate of approximately 16 FPS increasing this with YOLOv7 to 36.5 FPS reaching a 66% mAP.
- Artificial intelligence to identify olive-tree cultivarsPublication . Mendes, João; Lima, José; Costa, Lino; Pereira, Ana I.The exponential advance in artificial intelligence techniques makes it possible to apply them to previously thought to be impossible sectors. In this work, a different approach is presented to identify the different varieties of olive trees present in the olive groves of Portugal. Using its leaves and deep learning algorithms necessary for its classification, the proposed system can perform a reliable, low-cost, and real-time identification of the olive trees.
- Build orientation optimization of car hoodvent with additive manufacturingPublication . Matos, Marina A.; Rocha, Ana Maria A.C.; Costa, Lino; Pereira, Ana I.Additive manufacturing is a widely used process consisting in the building of a three-dimensional (3D) object from a model projected on a computer, adding the material layer-by-layer. This technology allows the printing of complex shape objects and is being increasingly adopted by the aircraft industry, medical implants, jewelry, footwear, automotive, fashion products, among others. The build orientation optimization of 3D models has a great influence on costs and surface quality when printing three-dimensional objects. In this work, three build orientation optimization problems are studied: single objective problem, bi-objective problem and many-objective problem. To this end, three quality measures are applied: the support area, the build time and the surface roughness, for the Car Hoodvent model. First, a single-objective optimization problem is presented and solved by the genetic algorithm, obtaining optimal solutions for each objective function. Then, the study of the bi-objective optimization problem is carried out for each pair of two objectives and some representative trade-off solutions are identified. Finally, the study of the many objective optimization problem, considering the three measures optimized simultaneously, is presented with some more optimal solutions found. The bi-objective and many-objective problems are solved by a multi-objective genetic algorithm. For a better analysis and comparison of the solutions found, the Pareto fronts are used, enabling a better visualization of the solutions between the objectives. This study aims to assist the decision-maker in choosing the best part print orientation angles according to his/her preferences. The optimal solutions found confirmed the effectiveness of the proposed approach.
- Deep learning networks for olive cultivar identification: a comprehensive analysis of convolutional neural networksPublication . Mendes, João; Lima, José; Costa, Lino; Rodrigues, Nuno; Pereira, Ana I.Deep learning networks, more specifically convolutional neural networks, have shown a notable distinction when it comes to computer vision problems. Their versatility spans various domains, where they are applied for tasks such as classification and regression, contingent primarily on the availability of a representative dataset. This work explores the feasibility of employing this approach in the domain of agriculture, particularly within the context of olive growing. The objective is to enhance and facilitate cultivar identification techniques by using images of olive tree leaves. To achieve this, a comparative analysis involving ten distinct convolutional networks (VGG16, VGG19, ResNet50, ResNet152-V2, Inception V3, Inception ResNetV2, XCeption, MobileNet, MobileNetV2, EfficientNetB7) was conducted, all initiated with transfer learning as a common starting point. Also, the impact of adjusting network hyperparameters and structural elements was explored. For the training and evaluation of the networks, a dedicated dataset was created and made available, consisting of approximately 4200 images from the four most representative categories of the region. The findings of this study provide compelling evidence that the majority of the examined methods offer a robust foundation for cultivar identification, ensuring a high level of accuracy. Notably, the first nine methods consistently attain accuracy rates surpassing 95%, with the top three methods achieving an impressive 98% accuracy (ResNet50, EfficientNetB7). In practical terms, out of approximately 2016 images, 1976 were accurately classified. These results signify a substantial advancement in olive cultivar identification through computer vision techniques.
- Impact of hyper-parameter tuning on CNN accuracy in agricultural image classificationPublication . Mendes, João; Lima, José; Costa, Lino; Hendrix, Eligius M.T.; Pereira, Ana I.This study explores the impact of hyper-parameter optimization on the performance of convolutional neural networks (CNNs) for olive cultivar classification using transfer learning. Pre-trained ImageNet models such as VGG16, InceptionV3, and ResNet50 were adapted to a proprietary dataset, with VGG16 selected for detailed evaluation. Key hyper-parameters, including layer count, neurons per layer, dropout rate, learning rate, and batch size, were tuned using random search. The best configuration achieved a validation accuracy of 87.5%, significantly outperforming the control model. Sensitivity analyses with Morris and Sobol methods identified the number of layers as the most influential factor, followed by dropout and learning rates through interaction effects. These findings demonstrate the importance of tailoring CNN architecture and regularization settings to the problem domain. These results underscore the importance of tuning architectural depth and regularization mechanisms for performance optimization. As a practical guideline, models with fewer layers and intermediate dropout levels demonstrated higher robustness and generalization, offering an effective strategy for adapting CNNs to agricultural classification tasks.
- Implementation of robust multi-objective optimization in the build orientation problemPublication . Matos, Marina A.; Rocha, Ana Maria A.C.; Costa, Lino; Pereira, Ana I.Additive manufacturing (AM) is an emerging technology to create 3D objects layer-by-layer directly from a 3D CAD model. The build orientation is a critical issue in AM and its optimization will significantly reduce the building costs and improve object accuracy. This paper aims to optimize the build orientation problem of a 3D CAD model using a robust multi-objective approach, taking into account the staircase effect and the support area characteristics. Thus, themain objective is to obtain a robust Pareto optimal front, composed of solutions that are not quite sensitive to perturbations in the variables. In this manner, a set of robust solutions is presented as alternatives and the decision-maker can identify the compromise solutions and choose according to his/her preferences.
- Machine learning to identify olive-tree cultivarsPublication . Mendes, João; Lima, José; Costa, Lino; Rodrigues, Nuno; Brandão, Diego; Leitão, Paulo; Pereira, Ana I.The identification of olive-tree cultivars is a lengthy and expensive process, therefore, the proposed work presents a new strategy for identifying different cultivars of olive trees using their leaf and machine learning algorithms. In this initial case, four autochthonous cultivars of the Trás-os-Montes region in Portugal are identified (Cobrançosa, Madural, Negrinha e Verdeal). With the use of this type of algorithm, it is expected to replace the previous techniques, saving time and resources for farmers. Three different machine learning algorithms (Decision Tree, SVM, Random Forest) were also compared and the results show an overall accuracy rate of the best algorithm (Random Forest) of approximately 93%.
- A multi-objective approach to solve the build orientation problem in additive manufacturingPublication . Matos, Marina A.; Rocha, Ana Maria A.C.; Costa, Lino; Pereira, Ana I.Additive manufacturing (AM) has been increasingly used in the creation of three-dimensional objects, layer-by-layer, from three-dimensional (3D) computer-aided design (CAD) models. The problem of determining the 3D model printing orientation can lead to reduced amount of supporting material, build time, costs associated with the deposited material, labor costs, and other factors. This problem has been formulated and studied as a single-objective optimization problem. More recently, due to the existence and relevance of considering multiple criteria, multi-objective approaches have been developed. In this paper, a multi-objective optimization approach is proposed to solve the part build orientation problem taking into account the support area characteristics and the build time. Therefore, the weighted Tchebycheff scalarization method embedded in the Electromagnetism-like Algorithm will be used to solve the part build orientation bi-objective problem of four 3D CAD models. The preliminary results seem promising when analyzing the Pareto fronts obtained for the 3D CAD models considered. Concluding, the multi-objective approach effectively solved the build orientation problem in AM, finding several compromise solutions.
- A multi-objective approach to the optimization of home care visits schedulingPublication . Alves, Filipe; Costa, Lino; Rocha, Ana Maria A.C.; Pereira, Ana I.; Leitão, PauloDue to the increasing of life expectancy in the developed countries, the demand for home health care services is growing dramatically. Usually, home services are planned manually and lead to various optimization problems in their activities. In this sense, health units are confronted with appropriate scheduling which may contain multiple, often conflicting, objectives such as minimizing the costs related to the traveling distance while minimizing the traveling time. In order to analyze and discuss different trade-offs between these objectives, it is proposed a multi-objective approach to home health care scheduling in which the problem is solved using the Tchebycheff method and a Genetic algorithm. Different alternative solutions are presented to the decision maker that taking into account his/her preferences chooses the appropriate solution. A problem with real data from a home health care service is solved. The results highlight the importance of a multi-objective approach to optimize and support decision making in home health care services. Moreover, this approach provides efficient and good solutions in a reasonable time.
- Multi-objective optimization in the build orientation of a 3D CAD modelPublication . Matos, Marina A.; Rocha, Ana Maria A.C.; Costa, Lino; Pereira, Ana I.Over the years, rapid prototyping technologies have grown and have been implemented in many 3D model production companies. A variety of different additive manufacturing (AM) techniques are used in rapid prototyping. AM refers to a process by which digital 3D design data is used to build up a component in layers by depositing material. Several high-quality parts are being created in various engineering materials, including metal, ceramics, polymers and their combinations in the form of composites, hybrids, or functionally classified materials. The orientation of 3D models is very important since it can have a great influence on the surface quality characteristics, such as process planning, post-processing, processing time and cost. Thus, the identification of the optimal build orientation for a part is one of the main issues in AM. The quality measures to optimize the build orientation problem may include the minimization of the surface roughness, build time, need of supports, maximize of the part stability in building process or part accuracy, among others. In this paper, a multi-objective approach was applied to a computer-aided design model using MATLAB® multi-objective genetic algorithm, aiming to optimize the support area, the staircase effect and the build time. Preliminary results show the effectiveness of the proposed approach. © 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG.