A method for seamless mosaic of large format equivalent central projection images

Abstract

The invention discloses a large-format equivalent center projection image seamless splicing method comprising the following steps: a. acquiring a rotating image based on a rotating camera system; b, correcting the rotating image onto a selected reference plane according to a horizontal angle reading provided by a total station during rotating photography to complete rotation image projection transformation; c, according to high-precision same-name image points obtained by matching within the an overlapping range of the rotating image, using a rotating image precise relative orientation model that takes into account camera pose shift to obtain precise relative orientation elements of the rotated image; d, correcting the rotating image onto a reference plane and realizing large-format equivalent center projection image seamless splicing according to the precise relative orientation elements obtained by the relative orientation model; and e, performing splicing accuracy evaluation, to bemore specific, splicing accuracy can be estimated by the unit weight mean error of same-name feature points on the rotating image after relative orientation. The method effectively increases the imagecoverage range of a single camera station and solves the problem that the angle of field of view of a digital camera is too small.

Description

technical field [0001] The invention relates to the field of large-format image acquisition in digital close-range photogrammetry and computer vision, and specifically relates to a large-format equivalent central projection image seamless splicing method. Background technique [0002] At present, digital cameras have become the mainstream technology for digital photogrammetry and computer vision data acquisition. Practical digital cameras need to have both high resolution and large field of view. However, some performances of digital cameras are far inferior to traditional silver-salt film imaging. The CCD or CMOS photosensitive sensor area is affected by materials and processes, and the image-bearing surface is much smaller than the film of silver-salt analog cameras with silver halide as the photosensitive material, resulting in The field of view of the digital camera is small at the same focal length, and increasing the ground coverage will help improve the accuracy of th...

AI Technical Summary

Problems solved by technology

Linear array push-broom imaging is not a projection center, and it is difficult to directly apply the photogrammetry analysis model; the combined camera achieves large-format image acquisition through image stitching, but its cost is high, the equipment is large, and the ground is not flexible; multi-baseline photogrammetry is to realize large-scale Another way to obtain large-format images, but the requirements for image overlap are high, and it is difficult to measure in the field, and it is difficult to meet the requirements of fast and low-cost operations

Images