A. Fusiello
This toolkit is a collection of Matlab function implementing several Computer Vision algorithms.
autocal - Autocalibration from fundamental matrices bundleadj - Bundle adjustment (Euclidean) calibLVH - Camera calibration from infnity homographies calibSMZ - Camera calibration from object-image homographies calib_radial - calibrate radial distortion with resection camera - Camera exterior orientation given COP,LOOK and UP cop_from_bearings - Camera locations from bearings dlt - Direct Linear Transform eight_points - 8-points algorithm for F/E essential_lin - Essential matrix with 8-points algorithm eucl_upgrade - Upgrade projective reconstruction given K exterior_iter - Exterior orientation with anisotropic Procrustes analisys exterior_lin - Exterior orientation with Fiore's algorithm exterior_nonlin - Non-linear refinement of exterior orientation fdh - compute focal distance from homography fund - Build fundamental matrix from camera matrices fund_lin - Fundamental matrix with 8-points algorithm fund_nonlin - Non-linear refinement of fundamental matrix fund_rob - Robust fundamental matrix gpa - Generalized Procrustes analysis hom_lin - Homography with DLT algorithm hom_nonlin - Non-linear refinement of H hom_rob - Robust homography fit hom_synch - Homography syncronization htx - Apply homogeneous transform icp - Iterative Closet Point algorithm imhs - Harris-Stephens corner strength imrectify - Apply rectification to images imstereo_ncc - Stereo block-matching with NCC imstereo_sch - Stereo block-matching with SCH (Census) imstereo_ssd - Stereo block-matching with normalized SSD imwarp - Image warp irdx - Undo radial distortion krt - Internal and external parameters from P lsq_nonlin - Levenberg-Marquardt method for non-linear LS opa - Orthogonal Procrustes analysis photostereo - Photometric stereo prec - Sturm-Triggs' projective reconstruction precond - Normalize the coordinates of kD points radial - Regress radial distortion parameters rdx - Apply radial distortion rectifyF - Epipolar rectification (uncalibrated) rectifyP - Epipolar rectification (calibrated) relative_lin - Relative orientation relative_nonlin - Non-linear refinement of relative orientation reproj_res - Reprojection residual and jacobian for 1 point - 1 camera reproj_res_batch - Reprojection residual for all points all cameras resect_lin - Camera matrix (resection) resect_nonlin - Non-linear refinement of resection robust_fit - robust regression rotation_synch - Rotation syncronization sampson_fund - Sampson residual for F (and its Jacobian) sampson_hom - Sampson residual for H (and its Jacobian) shape_from_silhouettes - Shape from silhouettes simpleGN - Gauss-Newton method for non-linear LS space_carve - Space carving algorithm translation_synch - Translation synchronization triang_disp - Triangulation from disparity triang_lin - Triangulation for one point in multiple images triang_lin_batch - Triangulation for n points in multiple images triang_nonlin - Non-linear refinement of triangulation
The MATLAB Computer Vision Toolkit does not have any other external dependency on toolboxes or packages, and runs indifferently on Octave and Matlab. It has been tested on Octave 4.2.1 and Matlab R2017a.
These are toolkits that are based on functions contained in the Computer Vision Toolkit:
The functions in this toolkit implement the methods described in the book:
The book is written in Italian, although the code documentation is in English.