Adaptive Wildlife Management System for Crop Protection Leveraging Edge Computing and Deep Learning
Authors: Gopikrishnan M, S. Anushya, B. Sreeharika, C. Navya Sree
Abstract
Adaptive wildlife management is essential to take the action between humans and wildlife, particularly in agricultural areas where crop damage caused by wild animals leads to significant economic losses. The advancements in edge computing and deep learning systems can be developed to monitor, detect, and deter wildlife intrusions. In our study we focus on designing the system using components, like Raspberry Pi as the primary processing unit, a Pi Camera for visual data collection, a PIR sensor for motion detection, an ultrasonic sensor for distance measurement, a GSM module for sending alerts, a buzzer for deterrence, and an LCD for displaying real-time status updates. The Pi Camera captures images of the animals, which we are an- alyzed using deep learning algorithm called convolutional neural networks (CNNs) for training on a diverse datsets of wild animals to distinguish between different species, reducing false alarms and improving decision-making. Edge computing minimizes response time making the system practical for deployment in remote areas. This will combine the real-time processing and accurate detection. Our model achieves the highest accuracy at 96.7% also focus on a vast dataset for a better result.
Potassium Permanganate Infusion: A Novel Approach To Activated Charcoal Modification
Authors: Iqbal N Shaikh, Faazil Shaikh
Abstract
To develop a modified potassium permanganate-infused activated material with improved efficiency in ethylene absorption for extended freshness of perishable goods. Ethylene absorption properties of activated charcoal were improved by strong bases and potassium permanganate (KMnO4) at concentrations of 1M, 2M and 3M. Activated charcoal was soaked in a 5 w/v% solution of KMnO4 for overnight and then filtered and the powder is collected and dried. The effects of ethylene absorbers with charcoal-KMnO4 and without as control on bananas during storage at room temperature were investigated.
SafeSurf: Enhancing Web Security Through Phishing Detection
Authors: Meena Kumari, Anuj Kumar Singh, Alok Singh, Harshit Pratap Singh
Abstract
Phishing attacks pose a severe cybersecurity concern because they exploit users' trust to get sensitive information. Conventional detection measures, such as rule-based and blacklist-based systems, usually fail to keep up with the most recent phishing techniques, particularly zero-day attacks. This review paper looks at the use of deep learning (DL) and machine learning (ML) models for detecting phishing websites, including ensemble methods (e.g., XGBoost, Adaboost) and decision tree-based approaches.The literature review identifies concerns such as computing requirements, data imbalance, and scalability issues, as well as each model's benefits, drawbacks, and overall performance. To address these issues, the study proposes a hybrid technique that includes real-time detection-optimized XGBoost for classification, Swarm Intelligence for feature optimization, and NLP-based feature extraction.To improve flexibility and scalability, future research will focus on cross-platform compatibility, reinforcement learning, and adversarial training. This extensive examination provides guidelines for developing dependable phishing detection systems that find a balance between precision, effectiveness, and real-time application.
SafeSurf: Enhancing Web Security Through Phishing Detection
Authors: Meena Kumari, Anuj Kumar Singh, Alok Singh, Harshit Pratap Singh
Abstract
Phishing attacks pose a severe cybersecurity concern because they exploit users' trust to get sensitive information. Conventional detection measures, such as rule-based and blacklist-based systems, usually fail to keep up with the most recent phishing techniques, particularly zero-day attacks. This review paper looks at the use of deep learning (DL) and machine learning (ML) models for detecting phishing websites, including ensemble methods (e.g., XGBoost, Adaboost) and decision tree-based approaches.The literature review identifies concerns such as computing requirements, data imbalance, and scalability issues, as well as each model's benefits, drawbacks, and overall performance. To address these issues, the study proposes a hybrid technique that includes real-time detection-optimized XGBoost for classification, Swarm Intelligence for feature optimization, and NLP-based feature extraction.To improve flexibility and scalability, future research will focus on cross-platform compatibility, reinforcement learning, and adversarial training. This extensive examination provides guidelines for developing dependable phishing detection systems that find a balance between precision, effectiveness, and real-time application.
Sustainable Efficiency and Scalability with Cloud-Based Business Management Systems (CBMS)
Authors: Mrs. Moushami Panda, Dr. Smruti Rekha Sahoo, Dr. Dulu Patnaik
Abstract
Cloud-Based Business Management Systems (CBMS) have transformed organizational operations by offering scalable, efficient, and sustainable solutions. This study examines how CBMS enhance efficiency, scalability, and sustainability across industries by streamlining workflows, reducing costs, and optimizing resource use. Key benefits include real-time data access, seamless collaboration, and integration with AI and machine learning, boosting productivity while minimizing physical infrastructure and environmental impact. CBMS also support business scalability, enabling growth without substantial capital investment. The research addresses challenges such as data security, compliance, and user adoption, providing strategies for overcoming these barriers. Case studies highlight CBMS’s transformative impact on operational efficiency and long-term sustainability. Ultimately, CBMS play a vital role in balancing technological advancement, business growth, and environmental responsibility. This study serves as a guide for business leaders, policymakers, and IT professionals looking to adopt cloud-based solutions in today’s evolving market.
AreoVac Forensic Solution: A Cleaning Technology That Aids in Collecting Forensic Data
Authors: Meena Kumari, Zainab Ilma, Harsh Sengar, Priyanshu Yadav, Dr. Ravi Prakash Verma
Abstract
AeroVac ForensiTech Cleaning Technology is a high-tech vaccum cleaner designed to revolutionize forensic research. Unlike the conventional cleaning equipment, it seamlessly integrates data collection and cloud-based data transfer capabilities. Examples of trace evidence gathered by the hoover as it carries out its main cleaning are dust, particulates, and micro-debris-all highly important for possible forensic analysis. Data collected is moved to the cloud in a safe and secure manner so that a centralized repository for analysis is achieved. Such data can be retrieved, analysed and traced in case of an accident or illegal behaviour, providing valuable information for forensic purposes. The AeroVac system is a cleaning tool and an indispensable forensic aid as effective cleaning technology is integrated with cutting-edge data analytics, ensuring better situational awareness and real-time evidence preservation. This project, there being a growing need for intelligent and networked devices in forensic applications, addresses this through a proactive approach to evidence gathering and management in a range of settings.