CameraCalibrations.jl

Camera Calibration algorithms in Julia using JuliaImages ecosystem

Idea is to be able to do something like below by end of the year for now:

• determining a camera's intrinsic parameters from a series of images
• common calibration algorithms e.g. zhang, tsai, dlt etc
• work with different calibration targets e.g. chessboard
• 3d reconstructions as shown below
• checkeboard detection algorithm which can provide the necessary inputs to a camera calibration routine
• projective geometry implementations, camera models, though CoordinateTransformations.jl, CameraModels.jl(by yakir), CameraCalibrations.jl(by yakir), ImageProjectiveGeometry.jl(by peter kovessi) need to be checked to avoid repeated methods

Current Update on checkboard detection:

using Images
using CameraCalibrations

img = Gray.(load("assets/test.jpg"))
res = process_image(img)
draw_rect(img, res,  Gray(1))
save("results.png", img)

Results:

Input image	Output Image

It would be great to pull off something like below by end of year:

References:

• Camera Calibration - common methods for camera calibration discussed
• Farneback method - algorithm for detecting crosses in grid which could be a checkerboard
• Above images were taken from here
• Input image credit