Real-time and Automatic Map Stitching through Aerial Images from UAV
This project will be focusing on using MATLAB software to implement SURF and ORB algorithm to do real-time and automatic map stitching through aerial images from UAV. The aim of this project is to achieve at least 1hz update rate by using Intel i7-8550U CPU and 4GB RAM.
Real-time aerial map stitching through aerial images had been done through many different methods. One of the famous methods was features based algorithm to detect features. There are three famous feature-based method such as scale-invariant feature transform (SIFT), speeded up robust features (SURF) and oriented FAST and rotated BRIEF (ORB). There are many advantages for human to generate real-time map from the viewpoints between 100-200m altitude. The high quality and fast generation of map could help to save people from fire, flood, and other disaster. A research done in 2016 with the method ORB monocular SLAM was able to achieve an average 10Hz processing rate to mosaic each frame from 1080p input video. The CPU used was Intel i7-4710 with 16GB RAM. The most recent research published in 2020 achieve a result of average 68.97Hz frame to frame processing rate for the 2.7k (2704×1521) resolution input image. The method used was parallel hashing to match the binary ORB descriptor rapidly. The CPU used was Intel i7-6700 with 32 GB RAM. In this research, the aim is to improve the features-based algorithm to achieve the real-time stitching map using Intel i7-8550U CPU and 4GB RAM with at least 1Hz update rate given the sequence of image from drone.