Browsing by Author "Pereira, Nuno José Matos"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
- 6D Pose Estimation and Object RecognitionPublication . Pereira, Nuno José Matos; Alexandre, Luís Filipe Barbosa de Almeida6D pose estimation is a computer vision task where the objective is to estimate the 3 degrees of freedom of the object’s position (translation vector) and the other 3 degrees of freedom for the object’s orientation (rotation matrix). 6D pose estimation is a hard problem to tackle due to the possible scene cluttering, illumination variability, object truncations, and different shapes, sizes, textures, and similarities between objects. However, 6D pose estimation methods are used in multiple contexts like augmented reality, for example, where badly placed objects into the real-world can break the experience of augmented reality. Another application example is the use of augmented reality in the industry to train new and competent workers where virtual objects need to be placed in the correct positions to look like real objects or simulate their placement in the correct positions. In the context of Industry 4.0, robotic systems require adaptation to handle unconstrained pick-and-place tasks, human-robot interaction and collaboration, and autonomous robot movement. These environments and tasks are dependent on methods that perform object detection, object localization, object segmentation, and object pose estimation. To have accurate robotic manipulation, unconstrained pick-and-place, and scene understanding, accurate object detection and 6D pose estimation methods are needed. This thesis presents methods that were developed to tackle the 6D pose estimation problem as-well as the implementations of proposed pipelines in the real-world. To use the proposed pipelines in the real-world a data set needed to be capture and annotated to train and test the methods. Some controlling robot routines and interfaces were developed in order to be able to control a UR3 robot in the pipelines. The MaskedFusion method, proposed by us, achieves pose estimation accuracy below 6mm in the LineMOD dataset and an AUC score of 93.3% in the challenging YCB-Video dataset. Despite longer training time, MaskedFusion demonstrates low inference time, making it suitable for real-time applications. A study was performed about the effectiveness of employing different color spaces and improved segmentation algorithms to enhance the accuracy of 6D pose estimation methods. Moreover, the proposed MPF6D outperforms other approaches, achieving remarkable accuracy of 99.7% in the LineMOD dataset and 98.06% in the YCB-Video dataset, showcasing its potential for high-precision 6D pose estimation. Additionally, the thesis presents object grasping methods with exceptional accuracy. The first approach, comprising data capture, object detection, 6D pose estimation, grasping detection, robot planning, and motion execution, achieves a 90% success rate in non-controlled environment tests. Leveraging a diverse dataset with varying light conditions proves critical for accurate performance in real-world scenarios. Furthermore, an alternative method demonstrates accurate object grasping without relying on 6D pose estimation, offering faster execution and requiring less computational power. With a remarkable 96% accuracy and an average execution time of 5.59 seconds on a laptop without an NVIDIA GPU, this method demonstrates efficiency and practicality performing unconstrained pick-and-place tasks using a UR3 robot.
- Leveraging Machine Learning for Weed Management and Crop Enhancement: Vineyard Flora ClassificationPublication . Corceiro, Ana; Pereira, Nuno José Matos; Alibabaei, Khadijeh; Gaspar, Pedro DinisThe global population’s rapid growth necessitates a 70% increase in agricultural production, posing challenges exacerbated by weed infestation and herbicide drawbacks. To address this, machine learning (ML) models, particularly convolutional neural networks (CNNs), are employed in precision agriculture (PA) for weed detection. This study focuses on testing CNN architectures for image classification tasks using the PyTorch framework, emphasizing hyperparameter optimization. Four groups of experiments were carried out: the first one trained all the PyTorch architectures, followed by the creation of a baseline, the evaluation of a new and extended dataset in the best models, and finally, the test phase was conducted using a web application developed for this purpose. Of 80 CNN sub-architectures tested, the MaxVit, ShuffleNet, and EfficientNet models stand out, achieving a maximum accuracy of 96.0%, 99.3%, and 99.3%, respectively, for the first test phase of PyTorch classification architectures. In addition, EfficientNet_B1 and EfficientNet_B5 stood out compared to all other models. During experiment 3, with a new dataset, both models achieved a high accuracy of 95.13% and 94.83%, respectively. Furthermore, in experiment 4, both EfficientNet_B1 and EfficientNet_B5 achieved a maximum accuracy of 96.15%, the highest one. ML models can help to automate crop problem detection, promote organic farming, optimize resource use, aid precision farming, reduce waste, boost efficiency, and contribute to a greener, sustainable agricultural future.
- Methods for Detecting and Classifying Weeds, Diseases and Fruits Using AI to Improve the Sustainability of Agricultural Crops: A ReviewPublication . Corceiro, Ana; Alibabaei, Khadijeh; Assunção, Eduardo Timóteo; Gaspar, Pedro Dinis; Pereira, Nuno José MatosThe rapid growth of the world’s population has put significant pressure on agriculture to meet the increasing demand for food. In this context, agriculture faces multiple challenges, one of which is weed management. While herbicides have traditionally been used to control weed growth, their excessive and random use can lead to environmental pollution and herbicide resistance. To address these challenges, in the agricultural industry, deep learning models have become a possible tool for decision-making by using massive amounts of information collected from smart farm sensors. However, agriculture’s varied environments pose a challenge to testing and adopting new technology effectively. This study reviews recent advances in deep learning models and methods for detecting and classifying weeds to improve the sustainability of agricultural crops. The study compares performance metrics such as recall, accuracy, F1-Score, and precision, and highlights the adoption of novel techniques, such as attention mechanisms, single-stage detection models, and new lightweight models, which can enhance the model’s performance. The use of deep learning methods in weed detection and classification has shown great potential in improving crop yields and reducing adverse environmental impacts of agriculture. The reduction in herbicide use can prevent pollution of water, food, land, and the ecosystem and avoid the resistance of weeds to chemicals. This can help mitigate and adapt to climate change by minimizing agriculture’s environmental impact and improving the sustainability of the agricultural sector. In addition to discussing recent advances, this study also highlights the challenges faced in adopting new technology in agriculture and proposes novel techniques to enhance the performance of deep learning models. The study provides valuable insights into the latest advances and challenges in process systems engineering and technology for agricultural activities.
- Test-as-a-Service: Application to Security TestingPublication . Pereira, Nuno José Matos; Sousa, Simão Patricio Melo de; Fernandes, João PauloIn a world where software gradually plays a key role daily, a failure may bring unpleasant consequences for its users. An example of a serious failure was the case Apple iCloud security exploit in 2014 where several private photos of celebrities have been accessed without permission[icl14a][icl14b]. Apart from economic and commercial implications, these faults lead to loss of trust in software by users, thus leading to the consequent search for an alternative and even result in leaving the old software for a new alternative. To address these shortcomings, the software industry started to use software testing to make sure that the software contains the minimum possible failures before is deployment. Software tests are used to analyse the program, namely to search some bugs. This analysis can be done without program execution (static analysis) or during execution (dynamic analysis). Static analysis tools can be used to check for potential execution of the program that have not been prematurely aborted due to unexpected event at runtime, not ensuring that the program will display the correct result. We studied some static analysis tools, JSFlow, JSPrime and TAJS, which analyse JavaScript code. These tools have been modified so they can be integrated into the Nibiru framework. Nibiru is a modular framework that aims to help in the implementation of software testing. It uses a micro-services architecture, enabling the use of multiple programming languages in his modules and has the ability to enable the implementation of its modules on multiple machines. So far the Nibiru has three operating modules and its ready to start growing with the community, so they can contribute in the construction of new modules or make small adjustments on the existing testing software to integrate the Nibiru framework.