- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- 3D Surveying and Cultural Heritage
- Robotic Path Planning Algorithms
- Indoor and Outdoor Localization Technologies
- Advanced Image and Video Retrieval Techniques
- Remote Sensing and LiDAR Applications
- Advanced Neural Network Applications
- Open Education and E-Learning
- Advanced Clustering Algorithms Research
- Advanced Sensor and Control Systems
- Data Management and Algorithms
- Robot Manipulation and Learning
- Underwater Vehicles and Communication Systems
- Infrastructure Maintenance and Monitoring
- Optical measurement and interference techniques
- Currency Recognition and Detection
- Antenna Design and Analysis
- Radiative Heat Transfer Studies
- Advanced MIMO Systems Optimization
- Music and Audio Processing
- Generative Adversarial Networks and Image Synthesis
- Embedded Systems Design Techniques
- Online Learning and Analytics
- Educational Games and Gamification
Tata Consultancy Services (India)
2012-2024
Jadavpur University
2022
Indian Institute of Technology Guwahati
2016
Australian National University
2003
Obstacle detection is an essential task for the autonomous navigation by robots. The becomes more complex in a dynamic and cluttered environment. In this context, RGB-D camera sensor one of most common devices that provides quick reasonable estimation environment form RGB depth images. This work proposes efficient obstacle tracking method using images to facilitate detection. To achieve early obstacles stable their states, as previous methods, we applied u-depth map Unlike existing present...
Visual SLAM shows significant progress in recent years due to high attention from vision community but still, challenges remain for low-textured environments. Feature based visual SLAMs do not produce reliable camera and structure estimates insufficient features a environment. Moreover, existing partial reconstruction when the number of 3D-2D correspondences is incremental estimation using bundle adjustment. This paper presents Edge SLAM, feature monocular which mitigates above mentioned...
The generation of robust global maps an unknown cluttered environment through a collaborative robotic framework is challenging. We present SLAM framework, CORB2I-SLAM, in which each participating robot carries camera (monocular/stereo/RGB-D) and inertial sensor to run odometry. A centralized server stores all the executes processor-intensive tasks, e.g., loop closing, map merging, optimization. proposed uses well-established Visual-Inertial Odometry (VIO), can be adapted use Visual (VO) when...
In this paper we propose a computationally inexpensive framework for dense 3D reconstruction on smart device platforms leveraging the feed from motion sensors. contrast to other methods, solely rely sensors compute pair wise epipolar relationships. particular, camera positions are obtained only through noisy mobile sensor data which is further optimized globally using iterative reweighted least squares. Rotations also Our method of obtaining reduce processing time entire pipeline. We use...
Robotic platforms are becoming increasingly important and have their utility cost completely justified during missions which require substituting humans. In this paper, we present our ongoing work in developing a multi-sensor robotic platform intended for deploying an indoor environment hazardous situations. The prime objective of such portable robots will be to conduct surveillance provide perception its surroundings the human-in-loop ascertain uncertain environment. Therefore, needs...
Rate of vehicular horn is an important parameter to estimate the traffic congestion a street in urban area developing nations. In this paper we propose participatory sensing based approach for condition monitoring using detection employing inbuilt sensors smart phones. Feature extraction performed on audio captured users' The features are then sent backend server classification and decision making. For feature use Modified Mel Frequency Cepstral Coefficient method, which modifies...
The maturity in autonomous navigation of Un-manned Aerial Vehicles (UAV s) provides the possibility to deploy UAV s for different kinds inspection jobs. Aircraft is a well-known periodic process aviation history, which long, costly, and subjective. Manual usually takes several hours, where multiple sensors on aircraft's outer surface, dents, lightning strikes, paint, etc. are mainly checked. main advantage UAV-based minimize turnaround time, reduces cost. Deployment collaborative UAVs must...
Rapid or real time dense 3D reconstruction of indoor outdoor model using mobile phone camera has become an active research topic because the importance models in different application like remote monitoring system and visualization. In this paper we propose as well a architecture which captures images hand held Smart generate accurate connected server parallel computation. The method performs near generation from streaming data. is capable adaptive data transmission based on available...
Autonomous Micro Aerial Vehicles (MAVs) gained tremendous attention in recent years. flight indoor requires a dense depth map for navigable space detection which is the fundamental component autonomous navigation. In this paper, we address problem of reconstructing while drone hovering (small camera motion) scenes using already estimated cameras and sparse point cloud obtained from vSLAM. We start by segmenting scene based on sudden variation 3D points introduce patch-based local plane...
The wireless communications channel is commonly modelled as imposing a Rayleigh distributed envelope onto any incumbent signals. Antenna arrays have long been recognised enabling exploitation of signal spatial diversity in such an environment based upon the assumption uncorrelated signals at different antennas. However, it precisely relative positions antennas which determine extent correlation. We consider effects on receiver output SNR performance, maximum ratio combining and equal gain...
K mean clustering is a very popular algorithm for numerical data. .It due to its simplicity of understanding and linear algorithmic complexity measure.But it has the serious limitation only data.Therefore several researchers tried improve k cluster not but also categorical dataset.In this work an effort have been made put forward proposed FCV which modified version traditional k-mean able objects having mixed type attributes i.e. categorical.For data fuzzy set similarity used differences...
Despite significant advances in recent years, the problem of image stitching still lacks a robust solution. Most feature based algorithms perform alignment on either homography-based transformation or content-preserving warping. Pairwise approach miserably fails to handle parallax whereas warping does not preserve structural property images. In this paper, we propose nonlinear optimization find out global homographies using pairwise homography estimates and point correspondences. We further...