IMU-Aided Deblurring with Smartphone

All code is stored in the src/ folder, and images & collected data are in the data/ folder.

Code

• parse_video.py: parses video into separate frames
• ahrs_plot.py: used for plotting IMU data
• calib.py: used for computing intrinsic matrix of the camera
• cp_hw6.py: helper file for calibration, provided for A6
• deconvolve.py: performs steps required for aided blind deconvolution
• frame_processing.py: parses frame data within an exposure into an IMUFrame object
• vec3.py: helper file for defining 3D vector object (not used)

Data

• parse_test/video_frames/: parsed frames of the exposure
• parse_test/edge_epochs.txt: log of edge timestamp data (start/end of an exposure)
• parse_test/frame_timestamps.txt: log of timestamp data
• parse_test/gyro_accel.csv: csv file containing IMU data logged at incremental timesteps
• parse_test/movie_metadata.csv: csv file containing camera data logged at incremental timesteps
• parse_test/movie.mp4: captured exposure used for testing the algorithm
• calib.npz: npz file containing intrinsic camera matrix and distortion values

Misc

• app-fdroid-release.apk: Android logger app
• imu_deblurring.pdf: original paper this implementation is based on
• logger_paper.pdf: paper providing details of the Android logger app
• report.pdf: our report pertaining to this project