Abstract
The manual monitoring of refilling stations in industrial environments can lead to inefficiencies and errors, which can impact the overall performance of the production line. In this paper, we present an unsupervised detection pipeline for identifying refilling stations in industrial environments. The proposed pipeline uses a combination of image processing, pattern recognition, and deep learning techniques to detect refilling stations in visual data. We evaluate our method on a set of industrial images, and the findings demonstrate that the pipeline is reliable at detecting refilling stations. Furthermore, the proposed pipeline can automate the monitoring of refilling stations, eliminating the need for manual monitoring and thus improving industrial operations’ efficiency and responsiveness. This method is a versatile solution that can be applied to different industrial contexts without the need for labeled data or prior knowledge about the location of refilling stations.

Abstract
In industrial environments, the measurement and monitoring of filling levels (FL) in refilling stations (RS) are critical for quality control processes. Traditional methods used for this purpose, such as manual inspection and sensor-based techniques, have proven to be costly and time-consuming. As an alternative, this paper proposes a novel approach that leverages computer vision (CV) and advanced image processing techniques. This approach provides a more efficient and accurate method for monitoring filling levels in refilling stations, thereby reducing operational costs. The system operates through a comprehensive five-stage pipeline, including pre-processing, perspective transformation, thresholding and edge detection, post-processing and filling level calculation. The performance evaluation of this approach demonstrated promising results in accurately determining filling levels in most scenarios. However, we also identified challenges such as overlapping columns and occlusions in the camera’s field of view that require further improvements. By addressing these challenges, our research aims to develop a streamlined and automated method for filling level measurement in refilling stations, thereby enhancing productivity in industrial environments. Ultimately, this proposed approach holds potential to significantly improve the efficiency of refilling stations across multiple sectors. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Abstract
Artificial intelligence (AI) has gained significant evolution in recent years that, if properly harnessed, may meet or exceed expectations in a wide range of application fields. However, because Machine Learning (ML) models have a black-box structure, end users frequently seek explanations for the predictions made by these learning models. Through tools, approaches, and algorithms, Explainable Artificial Intelligence (XAI) gives descriptions of black-box models to better understand the models’ behaviour and underlying decision-making mechanisms. The AI development in companies enables them to participate in Industry 4.0. The need to inform users of transparent algorithms has given rise to the research field of XAI. This paper provides a brief overview and introduction to the subject of XAI while highlighting why this topic is generating more and more attention in many sectors, such as industry.

Abstract
Industry 4.0 was publicly introduced in Germany in 2011 and is known as the fourth industrial revolution, whose goal is to improve manufacturing processes and increase the competitiveness of the manufacturing industry. Industry 4.0 uses technological concepts such as Cyber-Physical Systems, Internet of Things and Cloud Computing to create services, reduce costs and increase productivity in industry. This paper aims to explore the use of context-aware applications in Industry 4.0 in order to assist workers in decision making and thus improve the performance of factory production lines. This literature review is part of the project “Continental AA’s Factory of the Future” (Continental FoF) and will integrate a context-aware system in Industry 4.0 of the mentioned company, which is a manufacturer of radio frequency devices for the automotive industry. This systematic literature review identifies, from the researched solutions, the concept of context and context-awareness, the main technologies used in context-aware systems, how context management is performed, as well as the most used integration and communication protocols. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Abstract
Public lighting plays a very important role for society's safety and quality of life. The identification of faults in public lighting is essential for the maintenance and prevention of safety. Traditionally, this task depends on human action, through checking during the day, representing expenditure and waste of energy. Automatic detection with deep learning is an innovative solution that can be explored for locating and identifying of this kind of problem. In this study, we present a first approach, composed of several steps, intending to obtain the segmentation of public lighting, using Seville (Spain) as case study. A dataset called NLight was created from a nighttime image taken by the JL1-3B satellite, and four U-Net and FPN architectures were trained with different backbones to segment part of the NLight. The U-Net with InceptionResNetv2 proved to be the model with the best performance, obtained 761 of 815, correct locations (93.4%). This model was used to predict the segmentation of the remaining dataset. This study provides the location of lamps so that we can identify patterns and possible lighting failures in the future.

Abstract
Selective thinning is a crucial operation to reduce forest ignitable material, to control the eucalyptus species and maximise its profitability. The selection and removal of less vigorous stems allows the remaining stems to grow healthier and without competition for water, sunlight and nutrients. This operation is traditionally performed by a human operator and is time-intensive. This work simplifies selective thinning by removing the stem selection part from the human operator's side using a computer vision algorithm. For this, two distinct datasets of eucalyptus stems (with and without foliage) were built and manually annotated, and three Deep Learning object detectors (YOLOv5, YOLOv7 and YOLOv8) were tested on real context images to perform instance segmentation. YOLOv8 was the best at this task, achieving an Average Precision of 74% and 66% on non-leafy and leafy test datasets, respectively. A computer vision algorithm for automatic stem selection was developed based on the YOLOv8 segmentation output. The algorithm managed to get a Precision above 97% and a 81% Recall. The findings of this work can have a positive impact in future developments for automatising selective thinning in forested contexts.