3jBSrSSKrSSKJr SSKJrJrJr SSKrSSKJ r SSK J r SSK J r JrJr SSKJrJrJrJrJrJrJr S /r"S S \ R25rg) z!Module containing RANSAC modules.N)partial)CallableOptionalTuple)nn)KORNIA_CHECK_SHAPE)find_essentialfind_fundamentalsampson_epipolar_distance)find_homography_dltfind_homography_dlt_iteratedfind_homography_lines_dlt"find_homography_lines_dlt_iterated#line_segment_transfer_error_one_wayoneway_transfer_errorsample_is_valid_for_homographyRANSACc^\rSrSrSrS%S\S\S\S\S\S \S \S \S \ \S S4U4Sjjjr S&S\S\S\S\S\ \ RS \ R4 Sjjr\S\S\S\S\S \4 Sj5rS\ RS\ RS \ R4SjrS\ RS\ RS\ RS\S \\ R\ R\\44 SjrS\ RS\ RS \\ R\ R44SjrS\ RS \ R4SjrS\ RS\ RS\ RS \ R4S jrS&S\ RS\ RS!\ \ RS S4S"jjrS&S\ RS\ RS!\ \ RS \\ R\ R44S#jjrS$rU=r$)'r*aModule for robust geometry estimation with RANSAC. https://en.wikipedia.org/wiki/Random_sample_consensus. Args: model_type: type of model to estimate: "homography", "fundamental", "fundamental_7pt", "homography_from_linesegments". inliers_threshold: threshold for the correspondence to be an inlier. batch_size: number of generated samples at once. max_iterations: maximum batches to generate. Actual number of models to try is ``batch_size * max_iterations``. confidence: desired confidence of the result, used for the early stopping. max_local_iterations: number of local optimization (polishing) iterations. N model_typeinl_th batch_sizemax_iter confidence max_lo_iters score_typeprosac_samplingseedreturnc (>[T U]5 /SQUlSS/UlXplX lX@lX0lXlXPl X`l XlXl Xl U U U US:Xa0[Ul[Ul["UlSUlSUlgUS:Xa0[*Ul[,Ul[.UlSUlSUlgUS:Xa0[0Ul[2UlSUl[2UlSUlgUS :Xa8[0Ul[5[2S S 9UlS Ul[2UlSUlgUS :Xa0[0Ul[6UlSUl[2UlSUlg[9USUR35e)a(Initialize the RANSAC estimator. Args: model_type: type of model to estimate: "homography", "fundamental", "fundamental_7pt", "essential", "homography_from_linesegments". inl_th: threshold for the correspondence to be an inlier. Internally is squared. batch_size: number of generated samples at once. max_iter: maximum batches to generate. Actual number of models to try is ``batch_size * max_iter``. confidence: desired confidence of the result, used for the early stopping. max_lo_iters: number of local optimization (polishing) iterations. score_type: scoring method to use: "ransac" or "msac". prosac_sampling: whether to use PROSAC sampling instead of random sampling. seed: optional random seed for reproducible results. If None, uses global random state. ) homography fundamentalfundamental_7pthomography_from_linesegments essentialmsacransacr!r$r"r#7POINT)methodr%z is unknown. Try one of N)super__init__supported_modelssupported_scoresrrrrrrrrrrerror_fnr minimal_solverr polisher_solverminimal_sample_sizepolisher_sample_sizerrrr r rr NotImplementedError) selfrrrrrrrrr __class__s P/home/wildlama/miniconda3/lib/python3.13/site-packages/kornia/geometry/ransac.pyr/RANSAC.__init__8sy6 ! "( 2$  $$$($.     %1DM"5D #?D '(D $()D % 9 9?DM";D #ED '(D $()D % = (5DM"2D '(D $#3D ()D % , ,5DM")*:8"LD '(D $#3D ()D % ; &5DM"0D '(D $#3D ()D %% 4LTMbMbLc&de e sample_sizepop_size iterationdevicec0Uc[R"S5nURbH[R"US9nUR URU-5 [R "X2XVS9nO[R "X2US9nUR USS9upU $)aMinimal sampler, but unlike traditional RANSAC we sample in batches. Yields the benefit of the parallel processing, esp. on GPU. Args: sample_size: number of samples to draw from the population. pop_size: size of the population to sample from. batch_size: number of sample sets to generate. iteration: current iteration number (used for PROSAC sampling). device: device to place the samples on. Returns: Tensor of sampled indices with shape :math:`(batch_size, sample_size)`. cpu)r@)r@ generator)kdim)torchr@r Generator manual_seedrandtopk) r8r=r>rr?r@rCrJ_outs r:sample RANSAC.samples$ >\\%(F 99 v6I  ! !$))i"7 8::j6WD::j6BD[a0 r<n_inlnum_tcconfc (SnX::agX:XagX::agUS:agUS::agSn[U5HnXPU- X- - -nM [[R"SU- 5[ U*[R"[ USU- 555- 5$)aTUpdate max_iter to stop iterations earlier https://en.wikipedia.org/wiki/Random_sample_consensus. Args: n_inl: number of inliers. num_tc: total number of correspondences. sample_size: size of minimal sample. conf: desired confidence level. Returns: Maximum number of samples needed to achieve the desired confidence. g& .>rDg?)rangeintmathlogminmax)rPrQr=rReps prob_inlieris r:max_samples_by_confRANSAC.max_samples_by_confs   ?   3; 3; {#A AI&*5 5K$488C$J'#sdDHHScKFW=X4Y*ZZ[[r<kp1kp2c URSSup4URX[R"X4URUR S95nU$)zEstimate models from minimal samples. Args: kp1: source keypoints with shape :math:`(batch_size, sample_size, 2)`. kp2: target keypoints with shape :math:`(batch_size, sample_size, 2)`. Returns: Estimated models tensor. Ndtyper@)shaper3rGonesrer@)r8r`rarr=Hs r:estimate_model_from_minsample$RANSAC.estimate_model_from_minsamplesH#&))BQ-   %**ZTWT]T]fifpfp*q rr<modelsc &[UR5S:XaUSn[UR5S:XaUSnURSnURS:Xa8URUR USSS5UR USSS5U5nO5URUR USS5UR USS5U5nXd:*nUR SS9nUR S:Xa.URSURS US 9R SS9- n O+UR S :XaUn O[S UR 35eU R5n XR5n XR5n X:R5n XznXX4$) aEVerify models by computing inliers and selecting the best model. Args: kp1: source keypoints. kp2: target keypoints. models: candidate models to verify. inl_th: inlier threshold. Returns: Tuple containing: - Best model - Inlier mask for the best model - Score of the best model rcNrr$rDrFr&rT)rYrZr'zUnsupported score type: ) lenrfrr2expandsumrclamp ValueErrorargmaxitemclone)r8r`rarkrrerrorsinl_mask score_ransacscorebest_model_idxbest_model_score num_inliers model_best inliers_bests r:verify RANSAC.verifys$ syy>Q d)C syy>Q d)C\\!_ ??< <]]3::j"a#CSZZPZ\^`acdEegmnF]]3::j"a#@#**ZY[]^B_aghF#|||* ??f $LLOfllsl&G&K&KPQ&K&RRE __ ( E77HIJ J 0557"2779 +113 / )9FFr<cNURS:Xa[X5nXX#4$X4$)zRemove degenerate samples based on model-specific constraints. Args: kp1: source keypoints. kp2: target keypoints. Returns: Tuple of filtered keypoints (kp1, kp2). r!)rr)r8r`ramasks r:remove_bad_samplesRANSAC.remove_bad_sampless0 ??l *1#;D9ci' 'xr<cj[R"U5RSS9RSS9nX$)zRemove degenerate models based on simple heuristics. Args: models: candidate models to filter. Returns: Filtered models tensor. rmrn)rGisfiniteall)r8rkrs r:remove_bad_modelsRANSAC.remove_bad_modelss6~~f%))b)155"5=|r<inliersc XSnX#SnURS5nURXE[R"SXdRUR S95nU$)zPolish the model using inliers through local optimization. Args: kp1: source keypoints. kp2: target keypoints. inliers: boolean mask indicating inlier correspondences. Returns: Polished model tensor. NrDrd)sizer4rGrgrer@)r8r`rarkp1_inlkp2_inlnum_inlmodels r: polish_modelRANSAC.polish_model$s],t$,t$,,q/$$ ejjG==QXQ_Q_`  r<weightscrURS;a[USS/5 [USS/5 URSURS:XaURSUR:a2[ SURSURSUR35eURS:Xa[U/S Q5 [U/S Q5 URSURS:XaURSUR:a2[ SURS URSUR35eg g ) aValidate input tensors for shape and size requirements. Args: kp1: source keypoints. kp2: target keypoints. weights: optional correspondence weights (not used currently). Raises: ValueError: if input shapes are invalid or insufficient correspondences. )r!r"N2rzMkp1 and kp2 should be equal shape at least [z+, 2], got z, r$)rrrz., 2, 2], got N)rrrfr5rs)r8r`rars r:validate_inputsRANSAC.validate_inputs9s8 ??; ; sS#J / sS#J /IIaLCIIaL0ciilTE]E]6] 1122]^a^g^g]hhjknktktjuw ??< < sO 4 sO 4IIaLCIIaL0ciilTE]E]6] 1122`adajaj`kl {$7^ =r<c `URXU5 [UR5n[U5n[R "SSUR URS9n[R "USUR[RS9n[UR5GHnURURXPRXR5n Xn X)n URX5up[U 5S:XaMaURX5n URS;aU R!SSS5n UR#U 5n U b[U 5S:XaMUR%XXR&S-5upnnX:dMUUR(:dM[UR*5HnUR-XU5nUb[U5S:XaM)UR%XUUR&S-5unnnnUU:a7UUR(:a'UR/5Sn UR/5nUnM O U R/5nUR/5nUnUR1[3U5XPRUR45nUS-UR-U:dGM Xg4$ Xg4$) aCall main forward method to execute the RANSAC algorithm. Args: kp1: source image keypoints :math:`(N, 2)`. kp2: distance image keypoints :math:`(N, 2)`. weights: optional correspondences weights. Not used now. Returns: - Estimated model, shape of :math:`(1, 3, 3)`. - The inlier/outlier mask, shape of :math:`(1, N)`, where N is number of input correspondences. rdrD)r@rer)r%r#rmrc)rfloatr5rorGzerosrer@boolrUrrNrrrirreshaperrrr6rrrvr^rVr)r8r`rarbest_score_totalrQbest_model_totalinliers_best_totalr]idxs kp1_sampled kp2_sampledrkrr model_scorer}rLmodel_lo inliers_loscore_lonum_inliers_lo new_max_iters r:forwardRANSAC.forwardWss Sw/"'(@(@"A#h ;;q!399SZZP+0;;vq[`[e[e+ft}}%A;;t77RSU_U_`D)K)K'+'>'>{'X $K;1$77 QF"BBAq1++F3FCK1$47;{{3VU`U`bcUc7d 4EK.K4C\C\4\t001A#007CH (c(mq.@ >Bkk#T\^b^i^ikl^l>m;Az8^ ;.^tG`G`5` ( 0 3","2"2"4&. 2$);;= %,]]_"#.  $77()63K3KT__  ET__, <33[&Z 33r<)rrr2rrrr5r3rr6r4rrrr0r1) r!g@i gGz?r-r'FN)N)__name__ __module__ __qualname____firstlineno____doc__strrrVrrr/rGr@TensorrN staticmethodr^rirrrrrrr__static_attributes__ __classcell__)r9s@r:rr*s ' " %"RfRfRf Rf  Rf  RfRfRfRfsmRf RfRfjrv*-;>KNX`afamamXn :\3\\#\U\WZ\\B  ELL UZUaUa (G<<(G&+ll(GW[A4<<A4&+llA4=Eell=SA4 u||U\\) *A4A4r<)rrW functoolsrtypingrrrrGrkornia.core.checkrkornia.geometry.epipolarr r r kornia.geometry.homographyr r rrrrr__all__Modulerr<r:rsO$( ,, 0`` *n4RYYn4r<