2024-25

Permanent URI for this collection

http://192.168.75.5:4000/handle/123456789/18453

Browse

Now showing 1 - 20 of 77

Adaptable Arms: innovations in flexible robotic manipulation
(NHCE, 2025) SURESH R 1NH22A1410 VARUN MS 1NH22A1411 SANGAMESH POLICE PATIL 1NH22ME411 SUDEEP 1NH21ME073
The problem definition of flexible robotic arm movement in the agriculture field centres around the need for automation solutions that can handle the diverse, unpredictable, and often delicate tasks involved in modern farming. Agriculture presents unique challenges that traditional rigid robotic arms struggle to address due to the variability of the environment, crop types, terrain conditions, and the need for precision in handling perishable goods. Flexible robotic arms offer a promising solution to these challenges, but their design and deployment come with several complex problems that need to be solved.
ADAPTIVE LEARNING USING BRAIN COMPUTER INTERFACE
(NHCE, 2025) DIKSHITHA D 1NH21A1027 MRUNAL RK 1NH21A1060 ARPITHA VINOD 1NH21CS032 C SUJITHA 1NH21CS055
This project aims to develop an advanced adaptive learning platform that leverages Brain-Computer Interface (BCI) technology and artificial intelligence to deliver a highly personalized and immersive educational experience. By utilizing EEG datasets and a BCI simulator, the platform monitors and interprets cognitive states such as focus, attention, fatigue, and engagement. This data is used to dynamically adjust the learning process, ensuring an optimal learning environment for each user. The platform enables users to upload their own study materials, which are then processed and transformed using Al-powered tools. These tools include text summarization for simplifying complex content, video and audio generation for visual and auditory learners, and the creation of interactive flashcards and flowcharts to cater to diverse learning styles. The system adapts the format of the content based on individual preferences, maximizing comprehension and retention. To keep users engaged and prevent cognitive fatigue, the platform integrates interactive games that serve as mental refreshers during study sessions. Additionally, user progress is tracked through quizzes that assess their understanding and mastery of the current topic. Based on quiz performance, the platform determines whether the user is ready to advance to the next topic or requires reinforcement of the current material through alternative approaches.
ADVANCED CALORIE AND PROTEIN COUNTER USING CHATTERBOX
(NHCE, 2025) KAMLESH SEERVI ; SHREYA KUMAR ; LIKITH SIGH D ; SHIVRAJ MUKKANNI
AI ASSISTED ECONSULT
(NHCE, 2025) ABHISHEK 1NH21A1002 VINAYAK RANJANAGI 1NH21AI120 UMAR FAROOQ MULLA 1NH21CS251 PARVEZ ALAM 1NH21CS174
The digital transformation era is significantly impacting various sectors, with the healthcare industry being at the forefront of this evolution. As technology continues to advance, its integration into healthcare is becoming increasingly essential for enhancing patient care and streamlining medical services. From artificial intelligence (AI) to machine learning (ML), and from real-time data analytics to telemedicine, these technologies are reshaping the way healthcare is delivered. This project focuses on developing a comprehensive telemedicine portal that serves as a vital link between patients, doctors, and hospitals. The primary objective of this portal is to bridge existing gaps in healthcare delivery, ensuring that quality medical services are accessible to everyone, regardless of geographical or logistical constraints. In many parts of the world, access to healthcare remains a challenge due to factors such as distance, limited medical infrastructure, and socioeconomic disparities. The telemedicine portal aims to address these challenges by providing a platform that connects patients with healthcare providers, thereby democratizing access to medical services. This initiative is particularly relevant in rural and underserved areas where healthcare facilities are often scarce. By leveraging advanced technological solutions, the portal ensures that patients can consult with doctors remotely, receive diagnoses, and access medical advice without the need for physical travel. This not only saves time and resources but also reduces the burden on already overwhelmed healthcare systems.
AI for combating cyber bullying
(NHCE, 2025) CHANDAN R 1NH21AI019 GIRIDHARAN H 1ΝΗ21AI032 CHRIS JORDAN 1NH21CS062 SANJANA JELLA 1NHH21CS299
Cyberbullying poses a significant threat to online safety, particularly impacting younger users who are more susceptible to toxic digital interactions. This research presents an innovative Al-based solution designed to automatically detect harmful behavior and toxic comments in online environments. By leveraging advanced deep learning models, specifically Bidirectional Long Short-Term Memory (Bi-LSTM) networks combined with Natural Language Processing (NLP) techniques, the system is trained on extensive datasets of user comments to identify both subtle and overt signs of cyberbullying. The approach incorporates effective preprocessing methods such as TextVectorization, along with caching, shuffling, and batching techniques to optimize performance. The Al system's ability to process and classify harmful content in real-time makes it a powerful tool for monitoring online communities, promoting safer interactions, and significantly reducing incidents of online harassment. Additionally, its adaptability across diverse platforms ensures scalability and flexibility in combating evolving cyberbullying tactics. By equipping digital platforms with this technology, the solution fosters healthier online environments, empowers moderators, and acts as a critical resource for organizations dedicated to enhancing user safety and well-being.
AI-DRIVEN REAL TIME TRANSLATION SYSTEM FOR INDIAN LANGUAGES
(NHCE, 2025) KALYAN RAM PALADUGU 1NH21AI042 ARITRA ACHARYA 1NH21CS030 VAMSI KRISHNA 1NH21AI061 SUBHRANEEL MUKHOPADHYAY 1NH21CS240
AI-DRIVEN TRAFFIC MANAGEMENT SYSTEM
(NHCE, 2025) RITOSUVRA RAY 1NH21CS197 SOUBHIK DAS 1NH21CS238 AMAN JAIN 1ΝΗ21A1010 LOCHAN SINGH D 1NH21A1050
The AI-Driven Traffic Management System leverages cutting-edge technologies, including computer vision, and machine learning, to revolutionize urban traffic management. Designed to address critical challenges like traffic congestion, road safety, and environmental sustainability, these systems utilize real-time data and advanced analytics to make informed decisions. By seamlessly integrating technology into urban infrastructure, AI-driven systems create smarter, more efficient cities that meet the growing demands of urbanization.
AI-DRIVEN VIRTUAL LAB GENERATOR
(NHCE, 2025) ABHISHEK K 1NH21A1003 MOHAMMED FUDAIL 1NH21A1058 SHARUN RAJ K 1NH21CS216 PAPPU HARSHITH 1NH21CS289
The Virtual Lab Generator is an innovative platform designed to redefine digital education by integrating ChatGPT and Unity for creating dynamic, customizable virtual laboratories. By leveraging a web-based interface, the platform enables educators and students to generate tailored lab simulations through simple natural language prompts. ChatGPT analyzes these inputs, determines experiment configurations, and collaborates with Unity's WebGL engine to render interactive, physics-based simulations that are both realistic and engaging. This project employs a modular architecture comprising a user-friendly web frontend, an Al-driven backend for interpreting user inputs, a structured repository of JSON based lab templates, and Unity for generating detailed virtual experiments. Users can describe experiments ranging from basic science demonstrations to advanced technical procedures, with the system dynamically generating labs featuring real-time interactivity and customizable parameters. This addresses common challenges in traditional labs, including high costs, resource limitations, and safety concerns, making scientific exploration accessible to a broader audience. The Virtual Lab Generator is particularly impactful in educational and research settings, offering personalized experimentation opportunities for learners and scalable solutions for institutions. The modular framework ensures the platform can evolve, with planned expansions including additional lab templates, integration with learning management systems, and support for immersive technologies like augmented reality. Such advancements aim to enrich interactivity, foster engagement, and broaden the scope of scientific experimentation.
AI-INTEGRATED SMART HELMET
(NHCE, 2025) ADARSH BABU-1NH21AI005 NOEL SIBI-1NH20AI144 ILHAAN IBRAHIM RM-1NH21CS100 JAYAADITHYA J-INH21CS108
Road safety remains a critical concern, especially for motorcyclists who are more vulnerable to accidents. My project, "Al Integrated Smart Helmets," addresses this issue by combining advanced artificial intelligence with innovative technology to enhance rider safety and convenience. This helmet is equipped with state-of-the-art sensors and an Al-driven system offering features such as crash detection, emergency notifications, accident prediction, vehicle diagnostics, and voice- controlled commands. In the event of an accident, the helmet's sensors detect the impact and automatically send alerts to emergency services and pre-configured contacts with the accident's location, ensuring timely medical assistance. The Al-powered accident prediction feature analyzes riding patterns, weather conditions, and road data to warn users about potential risks, while vehicle diagnostics allow riders to monitor their bike's condition-such as fuel levels, engine health, and tire pressure-through a connected mobile app, reducing the likelihood of breakdowns. Additionally, the helmet incorporates voice-controled commands for hands-free navigation, music control, and call management, enhancing convenience while ensuring riders stay focused on the road. The connected mobile app serves as a hub, offering real-time insights and notifications while seamlessly integrating with the helmet via Bluetooth for smooth operation. Developed during my undergraduate engineering studies, this project demonstrates the potential of merging Al with wearable technology to solve real-world problems. By addressing safety, convenience, and vehicle maintenance, the Al Integrated Smart Helmet empowers riders to make informed decisions and fosters a safer riding environment. This innovative solution sets a new standard for road safety, encouraging wider adoption of smart technology among motorcyclists and showcasing my commitment to leveraging Al for meaningful, real-world impact.
AI-powered personalized health management for diabetes
(NHCE, 2025) CHETHAN S REDDY 1NH21AI024 DHIKAN KUMAR REDDY F 1NH21AI026 SUMITH BILAMKАЯ 1NH21CS241 SHREYAS SHREYAS 1NH21CS225
significantly influencing quality of life and presenting challenges for healthcare systems. Effective management of diabetes is crucial as it mitigates risks of severe complications, including cardiovascular diseases, kidney damage, and nerve disorders. The "HealthWise" project sets out to revolutionize diabetes care by developing an AI-powered personalized health management platform, tailored specifically for diabetic patients. This platform leverages cutting-edge machine learning algorithms and data analytics to provide individualized nutrition and exercise plans, thereby enhancing overall health and improving glucose control. Diabetes management requires addressing complex interconnections between various health parameters, including glucose levels, dietary habits, and physical activity. Conventional approaches often fail to integrate these aspects effectively, leading to fragmented and suboptimal care. HealthWise overcomes these challenges by employing advanced predictive models that analyse real-time data collected from wearable devices and mobile applications. These models ensure continuous and comprehensive monitoring of users’ health metrics, offering actionable insights in real time. A significant innovation of HealthWise lies in its user-centric design. The platform’s intuitive interface simplifies access to personalized recommendations, progress tracking, and health monitoring. Users receive real-time alerts for critical health events and motivational feedback to encourage adherence to health plans. Features such as meal logging, glucose trend analysis, and customized fitness routines empower users to take control of their health with ease. The backend infrastructure of HealthWise is built using FastAPI for API development, integrating SQLAlchemy for efficient database operations, and Cryptography for securing sensitive data. Predictive analytics is driven by machine learning frameworks such as TensorFlow and PyTorch, ensuring high accuracy and reliability in generating health recommendations. Additionally, the platform incorporates continuous glucose
Air gesture suite
(NHCE, 2025) VINAY PRASAD K 1NH21AI119 YASHAS D 1NH21AI123 JYOTHSNA K 1NH21CS115 PRATIBHA PRAKASH MACHAKANUR 1NH21CS185
The way people interact with computers has dramatically changed in the past few years with the advent of cutting-edge technology such as computer vision, machine learning, and speech recognition. One such cutting-edge system makes it possible to engage with virtual tools by doing away with the necessity for physical contact to run a computer. The three main components of this system are an air-based virtual keyboard, a gesture-driven calculator, a virtual drawing canvas, and a gesture-controlled virtual mouse with voice recognition. Together, these components provide a user-friendly and effective interface that makes human-computer interaction easier and more accessible. Voice commands, gestures, and hand movements can all be interpreted by the system. The computer's sensory organs, a camera and a microphone, are used to record these inputs. The system interprets these inputs and uses computer vision, machine learning, and speech recognition technologies to do jobs digitally. With this creative method, users can engage with computers without requiring conventional input devices like a keyboard, mouse, or stylus. Every part of the system has a unique function: With a gesture-controlled virtual mouse, users can move the pointer and carry out mouse actions with their hands. The Air-Based Virtual Keyboard replaces a physical keyboard by converting finger positions and airborne motions into keystrokes. Gesture-Driven Virtual Calculator: This feature calculates mathematical values using particular hand motions. Virtual Drawing Canvas: This enables creative expression without requiring a stylus or touch input by converting hand gestures into digital artwork.
Artificial intelligence-driven magnetic levitation automated train
(NHCE, 2025) MOHAMMED SAMEER-1NH22A1407 GOWTHAMI Η Ν-1ΝΗ22A1403 MADAN B K-1NH22ME406 MOHAMMED FOUZAN SIDDIQUI-1NH21ME044
Magnetie Levitation (Maglevtechnikegy greveins a hecalitemgh it Непреки нудат by wing magnetic fields solith and propel vehicles, eliminating the need for traditional wheel and reducing friction. This movative concept allows for higher spends, smoother rides, and greater eficiens, making it a promising technology for finare tramport solutions. The purpose of this project is to develop a simplified model that demonstrates the principles of Magics technology using readily available cremponents. In this project, a small-scale Maglev train is constructed, with a set of spare-shaped magtets serving as the track and a basic box functioning as the train. The system utilizes the magnetic mputsices between like poles of the magnetis to levitate the min, suspending & above the track without direct comact. To facilitate movement along the track, tan 6V iny motors are used to propel the train, with power supplied through a 2-channel 5V relay module The core of the project is an Arduino Uso (CH340) microcontroller, which coordinates the operation of the train. The Arduino controls the motors and the reby module, coturing the train moves smouthly along the track. A sensor is employed to monitor the position of the zin, enabling atomatic aljustments to maintain stable levitation. By adjusting the speed of the motors, the system ensures that the main stays at the optimal height above the insck and mover at a controlled pace. This model illustrates how magnetic levitation works at a fasic level ofleting insight into the technology that could potentially revolutionize transportation. The alay module, which interfaces with the motors, allows for precise control of the trin's movement, ensuring that accelerates and decelerates sesonthly. The use of sensors alds on addinozal layer af stability muring the levitation is maintained during the movement. Throughs this project, we aim to showcase the fundamental principles of magnetic levitation and its applications in real-world matsportation. Although this model is scaled down, effectively demonstrates the key elements-levitation, movement, and control-that are integral to Maglev systems. The projet not only serves as an educational set in also suntributes to the invader explontism of Maglev technology, peiting soward a fiture where more efficient and fласт патриана syvems are possible
Assistive vision: action detection and alert system for the visually impaired using deep learning and IoT
(NHCE, 2025) ABIKP ΙΝΗ21Α1004 ALWYN A INH21EC015 NEELUS INH21A1065 HITHESH GUFTHA V INHZ1EC967
For visually impaired people around the world, navigating in complex situations independently and safely continues to be a major difficulty. Over 2.2 billion people worldwide suffer from some kind of vision impairment, according to the World Health Organization, and millions of them rely significantly on assistive technology to go around. In addition to limiting one's freedom of movement, visual impairments have a significant impact on one's self-esteem, independence, and general quality of life. Although they have inherent limitations, traditional aids such as guiding dogs and white canes have long been the main equipment for mobility support. A white cane, for example, can provide tactile feedback to identify barriers in close vicinity, but it is unable to recognize risks that are far away or dynamic, such as moving cars, uneven terrain, or quickly shifting environments. In the same way, guiding dogs, Despite being priceless friends, they are expensive and not always available because of the substantial resources and training needed. The drawbacks of these conventional aids underscore the pressing need for creative, technologically advanced solutions that might provide improved safety, mobility, and situational awareness. Unprecedented chances to solve these issues have been made possible by recent developments in the domains of artificial intelligence (AI), notably deep learning, and the Internet of Things (IoT). This project presents the "Intelligent Blind Stick," a state-ofthe- art assistive technology solution that combines cutting-edge deep learning algorithms with Internet of Things devices to empower visually impaired people. The main objective of the system is to provide users with emergency support, predictive alarms, and real-time feedback so they may move through their environment more confidently and independently.
AUGMENTED REALITY BASED REAL TIME LANGUAGE TRANSLATION SYSTEM
(NHCE, 2025) V RAHUL 1NH21CS255 MOHAMMED AYMAN 1N21CS159 MOHAMMED OWAIS ARIF 1NH21AI059 M M VARUN CHENGAPPA 1NH21A1051
One new area that has a lot of potential for removing language barriers is real-time language translation. Optical character recognition (OCR), natural language processing (NLP), and live video processing are all combined into one seamless application in this project's real-time language translation system. The program offers an interactive, user-friendly platform for text detection and translation in live video streams and was created with Flask, Paddle OCR, Google Translator API, OpenCV, and PostgreSQL. In real-world situations, such as signboards, documents, or any textual content recorded by a camera, the technology seeks to help people comprehend foreign writing. Developing a real-time language translation system that can identify text in live video streams, translate it into a target language, and superimpose the translated text on the video is the aim of this project. User authentication, text detection, OCR, translation, live video processing, and database connectivity are all included in the modular system. Every part functions in unison to deliver precise and effective translations in real time, guaranteeing user interaction and usability. The architecture of the system is designed to manage several tasks. First, secure application access is guaranteed by the user authentication module, which is implemented with Flask and PostgreSQL. In addition to choosing their preferred source and destination languages, users can register and log in. Next, PaddleOCR uses bounding boxes to identify text sections in live video frames. Next, the Google Translator API receives the identified text and begins translating it. Pillow is used to display the translated text over the video feed, while OpenCV makes sure that live video processing runs smoothly.
Auto-Eval
(NHCE, 2025) VINITH RAJU ML 1NH21CS146 G POOJA SRI 1NH21CS081 AATREY KIRAN 1NH21A1001 ADETTH RAJU 1NH21A1007
The significant rise in student enrolments, especially in online and distance learning environments, highlights the urgent need for fast, efficient, and unbiased grading systems. Traditional assessment methods, often long, error-prone, and subject to human bias, hinder educators from delivering timely and meaningful feedback essential for student understanding and performance. As noted by C. Halkiopoulos et al. [1], these limitations emphasize the importance of leveraging technology to enhance the grading process. This project aims to revolutionize assessment by introducing an AI-driven grading system that integrates advanced technologies to automate evaluations while enhancing the overall learning experience. By reducing the burden on educators, this system allows for rapid, accurate grading, enabling teachers to focus on more meaningful interactions with students. Beyond efficiency, the system prioritizes personalization, offering targeted feedback and recommendations based on each student's unique needs and progress. With its ability to identify areas of improvement, the AI system transforms assessments into tools for continuous learning, fostering engagement and improved educational outcomes. Through intuitive interfaces for students and teachers, the project envisions a future where grading systems not only streamline processes but also create opportunities for deeper, more personalized learning experiences. This innovative approach alis with the evolving demands of modern education, ensuring students receive actionable feedback while educators enjoy reduced workloads and enhanced teaching capabilities.
Automated alerting system is for medical emergencies
(NHCE, 2025) N MAHITH KUMAR 1NH21CS170 SYED VAHID 1NH21CS244 M BHANU PRASAD 1NH21A1055 G YASHWANTH REDDY 1NH21A1130
Timely response to medical emergencies is critical for improving patient outcomes in healthcare settings. Effective alerting systems are pivotal in ensuring healthcare providers receive timely notifications to initiate appropriate interventions. This paper provides an overview of methodologies and technologies utilized in alerting systems designed for medical emergencies. It delves into essential components such as sensor integration, communication protocols, and decision-support algorithms that are foundational to these systems. The primary focus lies in optimizing alert delivery to reduce response times and enhance patient care quality. Furthermore, the paper examines prevalent challenges including system reliability, false alarms, and interoperability issues. It also explores current research trends and technological advancements aimed at mitigating these challenges. By critically evaluating existing frameworks and proposing potential enhancements, this paper aims to contribute to the continuous improvement and reliability of emergency alerting systems in healthcare.
Automated response system for earthquake using deep learning techniques
(NHCE, 2025) JERISH JESUDAS 1NH21AI040 DIYA PONNARIAN 1NH21EC051 HYNDAVI .B 1NH21EC030
Earthquakes are one of the most destructive natural disasters, capable of causing significant loss of life, damage to infrastructure, and long-term economic repercussions. The unpredictable nature of earthquakes makes them a critical concern for disaster management, as traditional methods of detection and response often fail to provide timely warnings. Current earthquake monitoring systems depend on seismic networks, which detect the occurrence of tremors after they have happened, providing little opportunity for preventative measures. This study proposes the development of an Automated Response System for Earthquakes Using Deep Learning Techniques, with an emphasis on using advanced machine learning methods such as Long Short-Term Memory (LSTM) networks to address this problem. The primary objective of this system is to detect seismic events in real-time and to predict their impact, enabling a quick response that can save lives and mitigate damage. Unlike traditional earthquake monitoring techniques, which primarily rely on detecting seismic waves after the event has occurred, deep learning models like LSTM are capable of analyzing seismic data in real-time to predict earthquakes with a degree of accuracy and lead time that was previously not possible. LSTM, a type of recurrent neural network (RNN), is especially well-suited for time-series forecasting tasks, such as predicting earthquake events, due to its ability to learn from long-term dependencies in sequential data. By analyzing the continuous stream of seismic data, LSTM models can be trained to recognize patterns that indicate imminent earthquakes, allowing for early detection.
Brain tumor detection and segmentation using u-net algorithm
(NHCE, 2025) GUTTA GEETHIKA 1NH21A1038 RAGHAVID KUMAR 1NH21A1075 GUTTA GEETHIKA 1NH21A1038 RAGHAVID KUMAR 1NH21A1075
Medical image segmentation plays a crucial role in the healthcare domain, enabling accurate diagnosis and effective treatment planning. This project focuses on the segmentation of brain MRI images using a hybrid approach that combines the deep learning-based U-Net architecture with traditional image processing techniques. The objective is to enhance segmentation accuracy, particularly in dealing with challenges like noise, irregular boundaries, and varying image quality in medical data. The U-Net model, specifically designed for medical image segmentation, serves as the foundation of this system due to its ability to capture detailed spatial and contextual features. Complementing the U-Net’s capabilities, traditional image processing methods such as thresholding, watershed segmentation, and morphological operations are applied to refine segmentation results. This combination ensures precise delineation of brain structures, overcoming the limitations of individual methods. To prepare the dataset for optimal performance, preprocessing steps including resizing, normalization, and data augmentation are implemented. These steps improve the robustness of the system, enabling it to generalize well across diverse datasets. The system's performance is evaluated using standard metrics like the Dice coefficient, Jaccard index, and F1-score, which provide a comprehensive understanding of segmentation accuracy. Visualization techniques are also employed to make the results interpretable and suitable for clinical applications.
Codeleed: integrated e-learning platform for coding
(NHCE, 2025) AZIM M MULLA 1NH22A1402 KISHOR MB 1NH22A1405 BHARATHM 1NH22CS403 M NITHIN 1NH22CS409
Traditional coding education often fails to effectively integrate theoretical knowledge with practical applications, creating significant barriers for learners striving to gain proficiency. Many learners struggle to bridge the gap between abstract concepts and real-world coding challenges due to the fragmented nature of existing educational resources. To address these limitations, we present an innovative e-learning platform designed to provide a holistic and guided learning experience. The platform combines structured instructional content with interactive coding exercises to seamlessly blend theoretical understanding with hands-on practice. A robust technology stack underpins the platform, featuring React.js for an engaging and dynamic user interface, Spring Boot for a scalable and efficient backend, and seamless integration with external tools like the YouTube API. These elements ensure the delivery of high-quality instructional videos paired with a diverse repository of coding exercises, spanning foundational programming principles to advanced problem-solving techniques. Key features include real-time feedback mechanisms and contextual guidance that empower learners to identify, understand, and correct errors as they arise. This iterative approach not only builds confidence but also fosters a deeper conceptual understanding of coding principles. Personalized learning pathways further enhance the user experience, dynamically adapting to individual proficiency levels and learning paces. This ensures that both beginners and advanced learners receive tailored support and challenges that align with their skill sets. Additionally, the platform emphasizes collaboration and peer learning through real-time code sharing and group-based exercises. These features promote teamwork, an essential skill in professional coding environments, and enhance engagement by encouraging active participation and shared problem-solving. The platform's design prioritizes accessibility and inclusivity, offering a responsive interface compatible with various devices. By eliminating technical barriers and providing structured, high-quality resources, the platform democratizes access to coding education, making it viable for learners from diverse backgrounds, regardless of their starting point. Beyond its immediate impact on learners, the platform aims to address the broader challenges of a rapidly evolving technological landscape. By equipping individuals with the skills and confidence to tackle real-world coding tasks, it contributes to bridging the industry skills gap, fostering innovation, and preparing the workforce for future demands. Through its unique combination of adaptive learning pathways,
Crop recommendation and crop disease detection using artificial intelligence (ai) and internet of things (iot)
(NHCE, 2025) KRUTHIKA H S 1NH22AI406 NIHALL 1NH22AI408 MAHALAKSHMI YM 1NH22EC406 JAMPULA SREYA 1NH21EC072
designed to empower farmers with data-driven insights for sustainable agriculture. This system integrates loT, machine learning (ML), and deep learning (DL) technologies to monitor field conditions, recommend suitable crops, and identify crop diseases accurately. Real-time field data is collected through loT sensors, which measure essential parameters such as soil moisture, temperature, pH, humidity, and sunlight exposure. The system also incorporates nutrient data (N, P, K) from laboratory reports to enhance accuracy. External weather data from APls and historical crop performance metrics are integrated into the dataset, enabling location-specific predictions and insights. The data transmission framework uses loT sensors to send field data to a central gateway. This gateway consolidates sensor inputs and securely transmits them to a cloud or edge server for storage and processing. Preprocessing steps such as data cleaning, normalization, and feature engineering ensure high-quality data for analysis. Normalization makes data such as pH and nutrient levels comparable, while feature engineering derives additional insights like soil moisture trends, temperature fluctuations, and drought probabilities. These refined datasets support robust model training and analysis. The ML model leverages soil and environmental data to create a crop suitability profile, recommending optimal crops for planting. Decision trees, random forests, and neural networks trained on labeled datasets provide accurate predictions. A separate deep learning model, built using convolutional neural networks (CNNs), is designed to detect and classify crop diseases based on leaf images. The PlantVillage dataset from Kaggle serves as the training data for this model, while the HDFS library ensures efficient storage and retrieval of large datasets during training and inference. A Flask-based backend system seamlessly integrates the ML and DL models, managing data flow between loT sensors, cloud storage, and the user interface. This backend processes user requests, executes model predictions, and delivers actionable insights. The client side interface, developed using HTML and CSS, presents soil recommendations, and disease detection results in an accessible asnodil udsaetra-f, riencdrolyp fpolarmntaint.g Faanrdm ecrrso pc ahne avltihew mraenaal-gtiemmee nitn. siTghhet s intocl usmioank e ofi nfloabrm-reedp ordteecdi siNo,n sP , rKeg avradluinegs apepnrophfraiolneacsc.e hs C totoh mea bgirpnicreeudcl itsuwiroaitnlh odlofe Tcc isrdoioaptn a- rmeacanokdmi nwgm.e eBanyt dhlaeetvri eorinnasgs iignbhgyt sA,i nlt-chdoerri vpseoynrs attetionmog l sed,n feasturamrileeesdr s a gnhauiotnrl iisetthnicet faTtaoonbh odiiadls i d tyaisdn certtcceeouusg rsrroia tapycttethe. i damdl liesezcyneisg sticeeormsno p in d syeauimgeplrpdoic,on ursrlttet,ru dairtu teec.a sed Bvtyrha eenps ocroeupvsro cidtesei unnwsgtti aaaaisln toraafeg blceilao,e b malefnba,d irnm riepnianrgleg- vt lioempTnr eta a cndmtdiicso eenasAi stl oetae rnciodnhu gnt ebosnrlyosesuagterkieemsss

Browse

Browsing 2024-25 by Title

Results Per Page

Sort Options