Rutgers University
WINLAB
Cameras for
Indoor
Positioning
Using Computer Vision for IPS
Background
We are using various tools and resources for our project.
OpenCV is an open source library for computer vision, used widely for applications in image processing for detection, recognition, and more. We are using this heavily in our project.
CUDA is an API model created by Nvidia which we aim to use with the Nvidia GPU processors in the ORBIT facility at WINLAB for heavy processing.
Axis M1054 Network Cameras are being used for capturing the video feed.
Tools and Resources
While we have GPS for identifying where a person is, a precise method to geolocate a person indoors does not exist.
We aim to use cameras for IPS, (indoor positioning system).
Using various techniques including machine learning algorithms and neural networks for computer vision, we plan to detect, recognize, and locate people with the use of cameras.
Goals
The ever-growing Internet of Things encompasses technologies consisting of connected devises, ranging from low-level sensors to powerful cameras. These devices are used to generate data which can then be interpreted and used in a meaningul way.
Computer vision is an exciting and emerging field for processing and analyzing large amounts of data in the form of images. Its applications are plentiful, from recognition and detection to statistical analysis.
IoT and Computer Vision
Progress
Summary
-
Experimented with CUDA using the orbit nodes
-
trained nerual networks with video feed
-
wrote algorithms for x-y location
-
researched camera intrinsics and calibration techniques
Progress
Summary
-
Set up and worked with network cameras
-
trained neural networks for tracking
-
wrote algorithms for calculating distance from camera to object
-
used camera calibration techniques to obtain intrinsics
-
implemented algorithms for fine-tuning and improving the results of the NN
Summary
-
Experimented with CUDA using the orbit nodes
-
trained nerual networks with video feed
-
wrote algorithms for x-y location
-
researched camera intrinsics and calibration techniques
Summary
-
Continued to learn more about OpenCV and its capabilities
-
tested algorithms for person detection and tracking
-
Researched neural networks and image processing to differentiate between objects and people
Summary
-
Created outline of goals and plans for what to accomplish
-
Become aquainted with the tools and software, like Ubuntu, OpenCV, and Owl
-
Researched various possible implementations, applications, and techniques we could use
Summary
-
Continued to learn more about OpenCV and its capabilities
-
tested algorithms for person detection and tracking
-
Researched neural networks and image processing to differentiate between objects and people
Who We Are
Hi. I'm a senior at TABC high school in Teaneck NJ. I love learning about all types of new computer science technolgy, from history, to harware, to software, to new OSes to algorithms. I go to a pretty normal high school so we dont have too much in the computer science department. I've taught myself (with the help of the internet and StackOverflow expecially) almost everything I know. I have worked with Python, Java, Arduino and Swift, and am currently building an app for the iOS App Store. My main interests are in AI and computer vision, which is why I started this project.
Avi Cooper
Poojit Hegde
I am a junior at the Middlesex County Academy for Science, Mathematics, and Engineering Technologies. I am passionate about learning as much as I can about science and technology, and I have worked on various software and hardware projects. In my free time, I enjoy playing basketball and following the latest sports news. I plan to purse mathematics and computer science in the future.
