The iterative Sfm alternating bundle adjustment method Break down the system into two phases SFM1: find pose phase SFM2 find model phase (method (a)or b)) Initialize first guess of model The first guess is a flat model perpendicular to the image and is Zinit away e. g Zinit=0.5 meters or any reasonable guess Iterative while( err is not small SFM1: find pose phase SFM2: find model phase Measurement error(Err) small or model and pose stabilized Pose estimation vo.a 16
The iterative SFM alternating bundle adjustment method • Break down the system into two phases: --SFM1: find pose phase --SFM2: find model phase (method (A) or (B)) • Initialize first guess of model – The first guess is a flat model perpendicular to the image and is Zinit away (e.g. Zinit = 0.5 meters or any reasonable guess) • Iterative while ( Err is not small ) • { – SFM1: find pose phase – SFM2: find model phase – Measurement error(Err) small or model and pose stabilized • } Pose estimation V0.a 16
Overall processing revisit Given: measurements Images of N frames Point feature tracked by klt kanade-lucas Tomasi feature tracker or SURF Speeded Up robust okiuhtfan n aMuse Features)methods Examples demo http://wwwyoutube.com/watch?v=rxpx9tjlpdo To find pose rotation R, translation t)of every frame and the model structure x Pose estimation vo.a 17
Overall processing revisit • Given: measurements – Images of N frames • Point feature tracked by KLT Kanade–Lucas– Tomasi_feature_tracker or SURF (Speeded Up Robust Features) methods • Examples, demo • http://www.youtube.com/watch?v=RXpX9TJlpd0 • To find pose (Rotation R, translation T ) of every frame, and the model structure X Pose estimation V0.a 17
Putting it altogether Repeat the explanations SFM1 (find pose phase) and SFM2 (find model phase) with implementation details Pose estimation vo.a
Putting it altogether Repeat the explanations SFM1 (find pose phase) and SFM2 (find model phase) with implementation details. Pose estimation V0.a 18
Putting it altogether Use KLT (or SIFT, Harris then correlation to obtain features in [u, v] There are t=1, 2, .. ,r image frames So there are A,1=(R,Ttl, 0 2=[R+2,.,etr=R, Ttr poses There is only one model M= /X, Y, Z I with j=1,.,N features Initialize first guess of model The first guess is a flat model perpendicularto the image and is Zinit away (e. g. Zinit =0.5 meters or any reasonable guess Iterative while err is not small k llllllll/// //for every time frame t, use all N features, run SFM1 once); so SFM1 runs /times here For(t=1;t<=t++) I Inputs: You have f(focal length), M =[X,,Z For each frame t, you have i=1,, N, image feature points and measurements [u, v]it Output: pose 8, After the above is run 0+1=(R, TI1, 0+2=(R T2,.,0(R, Tt r poses are found llllll/ SFM2: model finding/ mmmlmmll For i=1, i<=N i++ (for every feature, use all T frames, run SFM2 once: so SFM2 runs n times here (SFM2: find model phase Pose estimation vo.a Measurement error(Err) small or model and pose stabilized
Putting it altogether • Use KLT (or SIFT, Harris then correlation) to obtain features in [u,v] T • There are t=1,2,…, image frames, • So there are t=1={R,T} t=1 , t=2={R,T} t=2 , …., t=={R,T} t= poses. • There is only one model Mi=[X,Y,Z]I ,with i=1,..,N features • Initialize first guess of model – The first guess is a flat model perpendicular to the image and is Zinit away (e.g. Zinit = 0.5 meters or any reasonable guess) • Iterative while ( Err is not small ){ – /////////////// SFM1: Pose finding //////////////////////////////////// – //(for every time frame t, use all N features, run SFM1 once); so SFM1 runs times here – For (t=1; t<=; t++) – { Inputs: You have f(focal length), Mi=[X,Y,Z]i – For each frame t, you have i=1,,,N, image feature points and measurements [u,v] T i,t – Output: pose t – } – After the above is run – t=1={R,T} t=1 , t=2={R,T} t=2 , …., t=={R,T} t= poses are found – ////////////////////// SFM2: model finding ////////////////////// – (For i=1,i<=N;i++) (for every feature, use all frames, run SFM2 once; so SFM2 runs N times here) • {SFM2: find model phase} – Measurement error(Err) small or model and pose stabilized} Pose estimation V0.a 19
Recall: SFM1: Find Pose phase //for every time frame t, use all N features, run SFM1 once); So SFM1 runs times here For(t=1;t<=t++) t Inputs: You have f(focal length), M=Xy,Z For each frame t, you have i=l,, N, image feature points and measurements [u,vlit Output: pose 0t After the above is run t=1-t;t=1,Ut=2-t1t=2,…Ut= SI=,Ty t=D poses are found Pose estimation vo.a
Recall: SFM1:Find Pose phase //(for every time frame t, use all N features, run SFM1 once); so SFM1 runs times here For (t=1; t<=; t++) { Inputs: You have f(focal length), Mi=[X,Y,Z]i For each frame t, you have i=1,,,N, image feature points and measurements [u,v] T i,t} Output: pose t } After the above is run t=1={R,T} t=1 , t=2={R,T} t=2 , …., t=={R,T} t= poses are found Pose estimation V0.a 20