MAN page from PLD autopano-sift-C-2.5.1-1.i686.rpm

AUTOPANO

Section: User Manual (1)
Updated: JANUAR 2005
Index

NAME

autopano - Generate panorama project from SIFT keypoint files.

SYNOPSIS

autopano[options]output.ptokeyfile1[.gz]keyfile2[.gz][keyfile3[.gz] [..]]

DESCRIPTION

Generate panorama project from keypoint data. First, read in two or more SIFTkeypoint files, then correlate the keypoint data and build a list of matches.This list undergoes some filtering and the best matches are used to createcontrol point information. The control point information are writting to aPTO compatible panorama project file. For example, programs such ashugincan read it.

Prior to any further use of the PTO, you ABSOLUTELY HAVE TO ADJUST THE FOCALLENGTH of all the images. This can be done in hugin, in the "Camera and Lens"tab.

OPTIONS

--ransac <on|off|1|0>: Switch RANSAC postfiltration on or off. Default is on. There are only a fewreasons to switch this off: if the keypoint density and matches are verysparse, RANSAC may filter too much. Or, if the lens geometry used is unusual(fish eye, very wide angle, micropanorama, ...) and does not resemble arectilinear projection. That is, if you get really bad results with RANSAC on,disable it.
--maxmatches <count>: Set the maximum number of control point pairs you want to generate per imagepair. The default is 16, so for each image pair, there are no more than 16control points generated. If there are more control points to choose from, anarea maximization metric is applied to keep the matches that cover most of theimage area. You can disable this by setting count to zero. Then all matchesare kept.
--disable-areafilter: At the final stage of creating control points, there is a list of matches forevery image pair that overlaps. All this matches are thought to be correct andhave been filtered using the RANSAC algorithm. However, often there are morematches available than the user wants to keep (see the "--maxmatches" option).In previous versions we applied a simple metric to pick out the matches thathave a good matching score. Now, we have an area maximization algorithm, thatmaximizes the area covered by the matches. In general, this leads to bettercoverage of the image overlap area. However, if you want the old behaviour,that only considers the SIFT matching score of every match, enable thisoption.
--integer-coordinates: Use natural number coordinates in the PTO file for the found matches. Defaultis to use sub-pixel float coordinates to increase precision. You may want totry this option in case you use another frontend than hugin and you experienceproblems with the PTO files created by autopano-sift.
--absolute-pathnames <on|off|1|0>: Store the absolute pathnames of the image files in the PTO file. Normally, incase all images are in the same directory and the output PTO file is alsosaved in this directory, only the image filenames are used. Otherwise theabsolute pathnames are used. Using this option you can enforce the behaviour.
output.pto: The filename of the PTO project file to generate.
keyfile[.gz]: The SIFT XML keypoint data file, as generated with thegeneratekeys(1)program. If the filename ends in ".gz", transparentgzip(1)decompression is used.

ALIGN OPTIONS

--align

Enable fully automatic pre-aligning algorithm. This results in yaw, pitch androtation coordinates being assigned to the images in the resulting PTO file.This option is not perfect yet, but in most cases the result is far betterthan without using it.

There are a number of conditions on the input images thatcan be used with this algorithm. First, the images should all be of the samedimension, scale and have a simple (rectangular) geometry with roughlyeuclidean distances. Second, the order of the input images is considered sothat the first images build an ordered row on the horizon. That is, the first,second, third, etc. images are strictly aligned left-to-right or right-to-leftand all lie roughly on the horizon. They must also all be of the same rotationangle, which must be one of three rough cases: bottom-is-bottom,bottom-is-left, bottom-is-right. Bottom-is-top is forbidden. In case thebottom is either left or right, we estimate its position based on averagekeypoint density (also experimental).

In case the horizontal first row is not properly detected, try to increase thedownscale resolution when creating the keypoint data. This will produce morekeypoints which makes it easier to build the horizont-row. In case it stilldoes not work, you should not enable this option. Please report bugs orsuccesses with this option.

--bottom-is-left

--bottom-is-right

Only usable if --align option is enabled. If your input images are 90 or -90degree tilted for the first row of horizontal images, you can force theorientation by telling the program where the bottom (floor) of the images islocated: on the left or right side. If this option is not used, the programwill attempt to automatically figure the orientation based on average keypointdensity in the left and right half of the image.

--generate-horizon <count>

Generate a horizon from the first row of aligned horizontal images. This canonly be used if the --align option is enabled. Then, up to count number ofhorizon control lines are written into the resulting PTO file at the middle ofthe first row images. The lines are optimally spaced and you should use valuessuch as 2, 6 or 14 to get optimal results (the sum of power of two, startingwith 2, 2 + 4, 2 + 4 + 8).

REFINEMENT OPTIONS

--refine: Enable the refinement step. The refining is done as last step before writingthe PTO output file. For every matching control point pair a small patch inthe original images is extracted at the original resolution. The imagepatches are matched against each other with the highest possible quality ofmatching, usually yielding dozens of keypoints. As this matches are derivedfrom the original resolution image, their location is more precise and theyare used to refine the original keypoint using one of the two methods below.Note that enabling the refinement step makes the total process longer, henceit is not enabled by default.
--refine-by-middle
--refine-by-mean: Two methods are available to choose the best point from the high resolutionmatches. Refine by middle searches the match closest to the originalkeypoints position and keeps only this match. Refine by mean builds thegeometric center of all keypoints found in the patch and uses thiscoordinates.
--keep-unrefinable <on|off|1|0>: In case a good match cannot be refined because it is located to near to theboundary of the image, the original match is kept by default. To disable thisbehaviour, you can use this option, which throws away matches that cannot berefined. However, you might want to increase the number of matches to keepper image pair using the "--maxmatches" option then.

COMPONENT IDENTIFICATION

The program complains about non-connected components in case there is one ormore images or image groups which have no relationship to the rest of theimages. This means there is no way to jump from an image in one group toanother group by just following control point pairs between any two images.This component identification is crucial for further optimization: if thereis one or more non-connected components, global optimization based on controlpoint pairs will be impossible.

To fix this problem, first identify the reason why there are no keypointmatches between the images in the different components. A common reason isthat the images in one component are very diffuse and only have very fewcontrol points (such as images of the sky or water). If this is the case, youcan try to increase the number of keypoints found in each image by increasingthe downscaleResolution parameter of thegeneratekeys(1)program. Or you can add control point pairs between the images of theindividual components manually, using software such as hugin. Another reasonfor different components could be that the images do not belong together toone panorama.

COMMON PROBLEMS

PTO does not optimize/render in hugin. Please check that you adjusted thecamera lens or focal length parameter for each image. We intentionally set itto a value that will disallow any operation from within hugin as to force theuser to set this parameter. Without knowing this parameter, any work would beinvalid, thats why we force such strict behaviour, sorry.

BUGS

No bugs known, if you find any, please send a bug report to me. I will try tofix it.

AUTHOR

Sebastian Nowozin <nowozin at cs dot tu dash berlin dot de>

This document was created byman2html,using the manual pages.