leigwar lightning arland
Raul Herrera
Lightroom gives you one of the widest feature sets available in panorama stitching software, but even with the best technology you still need the proper technique. Whether your goal is outdoor photography or capturing a cityscape, make sure you have the right tools and know-how for the job.
Modern digital camer as do a lot of automatic white-balancing, aperture adjustment, focus adjustment, and shutter speed changes when you take a photo. While this can be a helpful feature, if you snap a bunch of photos to stitch together you can end up with an in consistent image due to differing exposures in each shot.
Tilt-shift photography is when you use camera movements that change the orientation or position of the lens with respect to the film or image sensor on the camera. Tilt-shift lenses allow you to change the orientation, perspective, and position of the lens without moving the camera back itself. These lenses allow you to rotate the lens plane relative to the image plane (tilt) and help you move the lens parallel to the image plane (shift).
In many outdoor and panoramic settings (when you shoot trees or buildings, for example), normal lens positioning might cause the lines of an object to have a distorted perspective. Tilt-shift allows you to set a camera in position to create natural perspective and then move the lens in such a way as to capture scale and height.
While you can make a lot of spectacular corrections with your photo editing software, the best practice is always going to be to take along a tripod. For panorama shots, especially, the stabilization you get from mounting a camera on a tripod will help keep your images level. That, in turn, will make the photos easier to stitch together.
Continue to hone your panorama photo skills and learn to merge photos in Lightroom. With the right shots and the right technology, you can master stitching together the photos that will create beautiful vistas.
I'm new to AF Photo, so this may be my ignorance. I'm trying to make a 360 cylindrical panorama (ie 360 horizontal x 120 vertical) from 12 portriat photographs taken with an 18mm (35 equivalent) lens. There is plenty of overlap and AF Photo appears to make a good job of stitching most of the pictures. Unfortulanely a stitching error is causing a gap which is preventing the full 360 wrap. If you look in the screenshot I have marked corresponding points on both sides of the gap. The pictures are taken with a tripod, level and the sun was overcast so no problems with moving shadows. It looks like AF it not giving enough freedom to the exact focal length to allow a proper wrapping. Full set of files here:
=0
Common wisdom suggests a 30% overlap between partial shots. If one is stitching a MF negative with (say) a cropped sensor camera, that might be as many as 6 shots or more. Do you go top, from left to right, and then bottom from left to right? Or what is the most effective way? Or, can one shoot only the upper (or left) half and stitch that, and then the lower (or right) half and stitch that, and then stitch both halves?
For stitching (which I do rarely) I preprocess shots in DxO PhotoLab, which takes into account focal length and distance settings of the lens model used for a shot, if the lens is supported. Corrections include vignetting, distortions, chromatic aberrations, differences of sharpness in the frame and more, all without leaving a RAW workflow. That last capability is quite valuable.
Thanks so much Nate and Digitizer. A lot of very helpful suggestions here. One of the most important takeaways for me is that less is more. That is, trying to squeeze every last datum from a negative by stitching many photos is probably counter-productive. Systematically improve process to a realistic outcome. Thanks!
I preprocess shots in DxO PhotoLab, which takes into account focal length and distance settings of the lens model used for a shot, if the lens is supported. Corrections include vignetting, distortions, chromatic aberrations, differences of sharpness in the frame and more, all without leaving a RAW workflow.
DxO lens/camera modules contain all data on specific camera and lens combinations. If you check out dxomark.com, you can see that lenses perform to different levels, depending on aperture, distance, zoom settings and on the camera which it is attached to. Combinations are measured and correction data is provided and applied automatically in DxO PhotoLab based on recorded metadata. Because of the sheer number of lens/camera combinations, not every possible combination can be measured though. Supported gear can be checked on the respective pages.
There are other products or possibilities to correct lens distortions of course. I simply happen to have PhotoLab (had it for years) and use it as needed, the new noise reduction tool is the best I know of, and provided lens/camera modules save me from having to do it myself. No matter if a lens has barrel, pincushion or moustache style distortions etc., all those defects simply go away.
Image stitching or photo stitching is the process of combining multiple photographic images with overlapping fields of view to produce a segmented panorama or high-resolution image. Commonly performed through the use of computer software, most approaches to image stitching require nearly exact overlaps between images and identical exposures to produce seamless results,[1][2] although some stitching algorithms actually benefit from differently exposed images by doing high-dynamic-range imaging in regions of overlap.[3][4] Some digital cameras can stitch their photos internally.
In order to estimate image alignment, algorithms are needed to determine the appropriate mathematical model relating pixel coordinates in one image to pixel coordinates in another. Algorithms that combine direct pixel-to-pixel comparisons with gradient descent (and other optimization techniques) can be used to estimate these parameters.
Distinctive features can be found in each image and then efficiently matched to rapidly establish correspondences between pairs of images. When multiple images exist in a panorama, techniques have been developed to compute a globally consistent set of alignments and to efficiently discover which images overlap one another.
A final compositing surface onto which to warp or projectively transform and place all of the aligned images is needed, as are algorithms to seamlessly blend the overlapping images, even in the presence of parallax, lens distortion, scene motion, and exposure differences.
Since the illumination in two views cannot be guaranteed to be identical, stitching two images could create a visible seam. Other reasons for seams could be the background changing between two images for the same continuous foreground. Other major issues to deal with are the presence of parallax, lens distortion, scene motion, and exposure differences. In a non-ideal real-life case, the intensity varies across the whole scene, and so does the contrast and intensity across frames. Additionally, the aspect ratio of a panorama image needs to be taken into account to create a visually pleasing composite.
Feature detection is necessary to automatically find correspondences between images. Robust correspondences are required in order to estimate the necessary transformation to align an image with the image it is being composited on. Corners, blobs, Harris corners, and differences of Gaussians of Harris corners are good features since they are repeatable and distinct.
One of the first operators for interest point detection was developed by Hans P. Moravec in 1977 for his research involving the automatic navigation of a robot through a clustered environment. Moravec also defined the concept of "points of interest" in an image and concluded these interest points could be used to find matching regions in different images. The Moravec operator is considered to be a corner detector because it defines interest points as points where there are large intensity variations in all directions. This often is the case at corners. However, Moravec was not specifically interested in finding corners, just distinct regions in an image that could be used to register consecutive image frames.
Harris and Stephens improved upon Moravec's corner detector by considering the differential of the corner score with respect to direction directly. They needed it as a processing step to build interpretations of a robot's environment based on image sequences. Like Moravec, they needed a method to match corresponding points in consecutive image frames, but were interested in tracking both corners and edges between frames.
SIFT and SURF are recent key-point or interest point detector algorithms but a point to note is that SURF is patented and its commercial usage restricted. Once a feature has been detected, a descriptor method like SIFT descriptor can be applied to later match them.
Image registration involves matching features[7] in a set of images or using direct alignment methods to search for image alignments that minimize the sum of absolute differences between overlapping pixels.[8] When using direct alignment methods one might first calibrate one's images to get better results. Additionally, users may input a rough model of the panorama to help the feature matching stage, so that e.g. only neighboring images are searched for matching features. Since there are smaller group of features for matching, the result of the search is more accurate and execution of the comparison is faster.
To estimate a robust model from the data, a common method used is known as RANSAC. The name RANSAC is an abbreviation for "RANdom SAmple Consensus". It is an iterative method for robust parameter estimation to fit mathematical models from sets of observed data points which may contain outliers. The algorithm is non-deterministic in the sense that it produces a reasonable result only with a certain probability, with this probability increasing as more iterations are performed. It being a probabilistic method means that different results will be obtained for every time the algorithm is run.
c01484d022