OneDrive
BaiduNetdisk(1234)
which had downsampled and divided into left video and right video.
1.Prepare
Run environment: MatlabR2019b or newer, python3.x (opencv-python)
You may use the external libraries
CVX: http://cvxr.com/cvx/
vlfeat: http://www.vlfeat.org/
then downsample you videos. For example, the downsample frequency is 2 hz, you video fps is 30 frames per second, that is to say you should select 2 frames per 30 frames in a second, just selecting the 1st frame and 16th frame in the first second is OK, and 31st, 46th, 61st frame successively.
2.SingleFlight
This code is for a single UAV or video bathymetry.Program entry main.m. It contains
1. DownSample by python;
2. Estimating camera motion;
3. Calculate camera extrinsics parameters;
4. Orthomorphic transformation;
For specific operations and parameters, see the corresponding readme and comments between codes.
3.MultipleFlight
This code is for two UAVs or videos bathymetry, which uses Video Stitching algorithm. Program entry main.m. It contains
1. Tracking feature points;
2. Background identification;
3. Video Stitching;
4. Estimating camera motion;
5. Calculate camera extrinsics parameters;
6. Orthomorphic transformation;
But first of all, you have to down-sample the two videos and place the pictures to ./dataset/left/ and ./dataset/right/.
*Note: Please calibrate your camera's intrinsics parameters in advance. And place it to ./Stitching/neededData/.
For specific operations and parameters, see the corresponding readme and comments between codes.
Some source code are also provided by Su Tan,
the original copy can be downloaded from this webpage https://github.com/SuTanTank/VideoStitchingViaShakinessRemoving
4.Bathymetry method
Including timeCor algorithm.
When you finished SingleFlight or MultipleFlight mission, save the filter image, orthoimage, camera extrinsics files, then set parameters bathyParams.m file, (cite a case timeCor), run mainMakeTimeStack.m -> mainV2.m.