International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
Comparative Assessment of Object-based and Pixel-based Approaches for Crop Cover Classification
Get PDFIntroduction/Importance of Study: Accurate crop identification and classification are crucial for effective agro-based planning and ensuring food availability. Reliable classification helps optimize agricultural productivity and resource management.
Novelty Statement: This study innovatively compares pixel-based and object-based approaches for machine learning-oriented classification methods to develop crop-type maps in Rahim Yar Khan, Pakistan.
Material and Method: Utilizing the Google Earth Engine (GEE) cloud computing platform, pre-processing steps were applied to Synthetic Aperture Radar Sentinel-1 and Sentinel-2 data. Integration of Sentinel-1 (VV, VH) and Sentinel-2 satellite bands enabled the computation of various indices and the production of composite images for subsequent analysis. The primary objective was to evaluate the effectiveness of these approaches in classifying major crops: cotton, rice, and sugarcane. Time-specific images were employed to leverage crop seasonality; for instance, an August composite image was prioritized for cotton, while September composites were used for rice and sugarcane classification. The study utilized two object-based segmentation approaches: Simple Non-Iterative Clustering (SNIC) on the GEE platform and Object-Based Image Analysis (OBIA) using Multi-Resolution Segmentation in E-Cognition software. The Random Forest (RF) machine learning algorithm was applied to both pixel-based and object-based approaches. Field sample data, including cotton, rice, sugarcane, orchards, and other crops, were used for classification, validation, and accuracy assessment. A comparative analysis was conducted to evaluate the performance of pixel-based and object-based methods.
Result and Discussion: The RF algorithm applied to pixel-based approaches using Sentinel-1 and Sentinel-2 imagery bands with composite indices demonstrated superior results. The pixel-based RF classification achieved 98% accuracy with a kappa coefficient of 92%. In comparison, RF applied to SNIC in GEE achieved 96% accuracy with a kappa coefficient of 95%, while OBIA in E-Cognition attained an accuracy of 89%.
Concluding Remarks: The study concludes that tuning the segmentation parameters in both E-Cognition and SNIC algorithms can enhance the accuracy of object-based classification
Urban Flooding and Climate Change Vulnerability-A Case Study of North Karachi
Get PDF
rban flooding in Karachi has been exacerbated by insufficient climate resilience measures, inadequate urban planning, and underdeveloped drainage systems. These deficiencies have led to widespread flooding in residential and commercial areas, causing significant damage to infrastructure and amenities. North Karachi, a densely populated suburb, is particularly vulnerable due to its proximity to major water bodies such as the Lyari River (LR) and Gujjar Nala (GN). The area\u27s elevation ranges from 5 to 96 meters, creating a natural slope towards the southeast, making areas in this direction, including UCs 3, 5, 6, and 8, highly prone to flooding. This vulnerability is further influenced by the geographical layout, with the Lyari River and the surrounding Pub Ranges affecting rainfall runoff patterns. To evaluate flood vulnerability in North Karachi Town, various analyses were performed. Elevation data, sourced from Google Earth Pro, was converted from vector to raster format using ArcMap’s interpolation tool. This analysis revealed a slope from the northwest to the southeast, influenced by the Pub Ranges to the west and the Lyari River to the east. UCs 4, 7, and 9 are especially at risk due to their lower elevations and proximity to the Lyari River. The Normalized Difference Vegetation Index (NDWI) was employed to assess vegetation stress. Pre-monsoon NDWI values ranged from -0.26898 to -0.04352, indicating severe water stress. Post-monsoon values ranged from -0.2021 to 0.04597, with the maximum value of 0.04597 corresponding to humid and flooded conditions. The study highlights the crucial need to maintain clear waterways to manage flood risks effectively. Authorities should focus on ensuring that drains and river channels are free of debris and encroachments to mitigate future flooding
Predictive Maintenance in Industrial Internet of Things: Current Status
Get PDFIntroduction/Importance of Study: Predictive Maintenance (PdM) is a key challenge within the Industrial Internet of Things (IIoT). It aims to enhance system operations by minimizing equipment failures, leading to smoother operations and increased productivity. By anticipating maintenance needs before failures occur, PdM ensures more reliable and efficient industrial processes. Novelty Statement: This study examines maintenance techniques and datasets that leverage AI and ML for predictive maintenance in the context of industrial IoT. The primary goal is to enhance productivity, identify faults before failures occur, and minimize downtime. By utilizing advanced algorithms, the study aims to improve the efficiency and reliability of industrial systems. Material and Method: A systematic literature review of state-of-the-art predictive maintenance in the context of industrial IoT, incorporating machine learning (ML) and artificial intelligence (AI) methods, is conducted. This review is based on research articles retrieved from the Dimensions.ai database, covering publications from 2018 to 2024.Result and Discussion: This comprehensive analysis offers valuable insights for advancing Predictive Maintenance (PdM) strategies in the Industrial Internet of Things (IIoT), ultimately contributing to more efficient manufacturing processes. The study highlights leading publication venues and top keywords in this research area, providing a clear picture of emerging trends. It also explores the prognosis of PdM within the manufacturing industry. Additionally, the review discusses relevant models, methods, input variables, and datasets in the PdM and IIoT domain, with a particular focus on machine learning (ML) and artificial intelligence (AI) techniques. Among the most widely used techniques for PdM in IIoT are deep learning, artificial neural networks, and random forest.Concluding Remarks: Subsequently, the study highlights various challenges, offering future research directions aimed at refining predictive maintenance techniques
Event-Based Vision for Robust SLAM: An Evaluation Using Hyper E2VID Event Reconstruction Algorithm
Get PDFThis paper investigates the limitations of traditional visual sensors in challenging environments by integrating event-based cameras with visual SLAM (Simultaneous Localization and Mapping). The work presents a novel comparison between a visual-only SLAM implementation using the state-of-the-art HyperE2VID reconstruction method and conventional frame-based SLAM. Traditional cameras struggle in low dynamic range and motion blur scenarios, limitations that are addressed by event-based cameras, which offer high temporal resolution and robustness in such conditions. The study employs the HyperE2VID algorithm to reconstruct event frames from event data, which are then processed through the SLAM pipeline and compared with conventional frames. Performance metrics, including Absolute Pose Error (APE) and feature tracking performance, were evaluated by contrasting visual SLAM implementations on reconstructed images against those from traditional cameras across three event camera dataset sequences: Dynamic-6DoF, Poster-6DoF, and Slider depth sequence. Experimental results demonstrate that event-based cameras yield higher-quality reconstructions, significantly outperforming conventional cameras, especially in scenarios marked by motion blur and low dynamic range. Among the tested sequences, the Poster-6DoF sequence exhibited the best performance due to its information-rich scenes, while the Slider depth sequence faced challenges related to drag and scaling, as it lacked rotational motion. Although the APE values for the Slider depth sequence were the lowest, it did experience trajectory drift. In contrast, the Poster-6DoF sequence displayed superior overall performance, with reconstructions closely aligning with those produced by conventional camera-based SLAM. The Dynamic-6DoF sequence showed the poorest performance, marked by high absolute pose error and trajectory drift. Overall, these findings highlight the substantial improvements that event-based cameras can bring to SLAM systems operating in challenging environments characterized by motion blur and low dynamic ranges
Explicit State Model Checking Effects on Learning-Based Testing
Get PDFExploring the impact of integrating an explicit state model checker into the learning-based testing (LBT) framework presents an intriguing challenge. Traditionally, LBT has leveraged symbolic model checkers such as NuSMV and SAL, which use Binary Decision Diagrams (BDDs) to analyze multiple states concurrently. In contrast, explicit state model checkers evaluate one state at a time, a key distinction that suggests potential advantages for explicit state checking in the context of LBT. Thus, it is valuable to investigate how integrating an explicit state model checking algorithm might influence the performance of LBT. Model checkers explore the state space to verify conformance with user-defined correctness requirements, typically represented as Linear Temporal Logic (LTL) formulas. If a property violation is detected, it is presented as a counterexample. NuSMV and SAL employ different algorithms for generating and displaying counterexamples. This paper specifically examines the effect of SPIN-generated counterexamples on the LBT process. Evaluation metrics include Total LBT Iterations, First Bug Reporting Time (in milliseconds), Counterexample Length, Precision, and Efficiency, among others. Total Model Checking Time (in milliseconds) captures the cumulative time spent verifying the model over all iterations. SPIN consistently requires the least time for all specifications compared to other model checkers. As a result, experiments demonstrate that SPIN is more efficient when integrated with LBT, leading to faster convergence of the LBT hypothesis to the target System Under Test (SUT) in comparison to NuSMV and SAL
Synergizing Digital Twin Technology for Advanced Depression Categorization in Social-Media through Data Mining Analysis
Get PDFThe progression from negative emotions to depression is a significant concern, marked by persistent sadness and an inability to cope with challenging circumstances. Regrettably, it can lead to the extreme step of suicide. According to the World Health Organization (WHO), 4.4% of the global population currently grapples with depression. Shockingly, 700,000 individuals worldwide took their own lives in 2023, and this tragic number continues to escalate. Our objective is to detect signs of depression in individuals through their social media posts, SMS, or comments. We collected nearly 10,000 pieces of information from Twitter comments, Facebook posts, and remarks. Employing data mining and machine learning algorithms has proven instrumental in swiftly discerning individuals\u27 emotional states. To predict depression versus non-depression, we employed six classifiers, with support vector machines (SVMs) demonstrating the highest accuracy. A comparison between SVM and Naïve Bayes revealed that Naïve Bayes yielded superior results in our study
Alex Net-Based Speech Emotion Recognition Using 3D Mel-Spectrograms
Get PDFSpeech Emotion Recognition (SER) is considered a challenging task in the domain of Human-Computer Interaction (HCI) due to the complex nature of audio signals. To overcome this challenge, we devised a novel method to fine-tune Convolutional Neural Networks (CNNs) for accurate recognition of speech emotion. This research utilized the spectrogram representation of audio signals as input to train a modified Alex Net model capable of processing signals of varying lengths. The IEMOCAP dataset was utilized to identify multiple emotional states such as happy, sad, angry, and neutral from the speech. The audio signal was preprocessed to extract a 3D spectrogram that represents time, frequencies, and color amplitudes as key features. The output of the modified Alex Net model is a 256-dimensional vector. The model achieved adequate accuracy, highlighting the effectiveness of CNNs and 3D Mel-Spectrograms in achieving precise and efficient speech emotion recognition, thus paving the way for significant advancements in this domain
Overview of Immersive Data Visualization: Enhancing Insights and Engagement Through Virtual Reality
Get PDFIn recent years, the explosion of data has been immense, especially in terms of volume and velocity which poses a new challenge in the visualization of data and extracting patterns from it efficiently. Visualization is one the most critical aspects of data analysis as it also helps in the selection of an appropriate model for machine learning. However, this changes when we are dealing with complex data or hyper-dimensional datasets. 2D visualization of this complex or hyper-dimensional dataset can be hard to visualize owing to the inherent loss of information due to spatial constraints which consequently hinders the extraction of meaningful patterns for the development of machine learning models. In recent years, there has been substantial advancement in immersive technologies like Virtual Reality, Augmented Reality, Mixed Reality and adoption in various sectors especially in gaming, entertainment, and training. However, when it comes to data analysis and data visualization, immersive technology is at an emerging stage but has promising potential. This review research paper, through a series of application domains, aims to uncover this promising potential of virtual reality by shedding light on its capabilities and its limitations in representing complex and hyper-dimensional data to uncover new insights, pattern recognition, and decision-making processes
Honey Adulteration Detection through Hyperspectral Imaging and Machine Learning
Get PDFIntroduction/Importance of Study:
The purity and authenticity of honey are paramount for ensuring consumer trust and maintaining the integrity of the honey industry. There is a pressing need for advanced and efficient detection methods to increase the prevalence of honey adulteration.
Novelty statement:
Our research provides a solution to the challenge of predicting the change in adulterated honey properties through hyperspectral imaging and advanced machine learning algorithms, filling a critical gap in existing methodologies.
Material and Method:
A publicly available dataset with spectral features, extracted through hyperspectral imaging, across different classes of honey and adulteration levels has been examined and various machine learning models were developed to identify honey adulteration concentration and type of honey. The dataset was balanced and a five-fold cross-validation technique was used to train the machine learning models.
Result and Discussion:
Random forest was found to perform better in three identified scenarios i.e. (a) type of honey (b) adulteration level (c) both (a, b); with a maximum average accuracy of 99.69% performing better than the one reported in the literature (95%). For both single-output and multiple-output ML models, the trend in feature importance was observed. The single model identifying the class of honey utilized low and mid-frequency spectra while the multi-model used mid-frequency spectrum only.
Concluding Remarks:
The proposed approach aims to provide an accurate and cost-effective solution to address the challenges associated with honey adulteration, contributing to the enhancement of honey quality assessment and consumer confidence
Meta-Space: Pioneering Education in the Metaverse
Get PDFIn the evolving landscape of learning methodologies, technology has emerged as a catalyst, transforming the educational experience. This study delves into the realm of Virtual Reality (VR) and Augmented Reality (AR), collectively referred to as the "Metaverse," as a pivotal tool in education. By conducting systematic literature reviews, we investigate the potential, effectiveness, and associated pros and cons of employing the Metaverse for learning. Our findings affirm that the Metaverse proves to be a highly effective learning platform, enhancing engagement through lifelike avatars and bridging the gap between the real and virtual worlds. While this innovative approach facilitates visualizing materials and fosters interactive and interesting learning environments, challenges such as the cost of requisite devices remain. Despite limitations, the advantages of integrating the Metaverse into education are evident, necessitating ongoing development to amplify benefits and address existing constraints. This research contributes valuable insights to the ongoing discourse on leveraging Metaverse technologies for enriching educational practices