Search

One of the practical impediments to the adoption of long focal length lenses in close range photogrammetry is the difficulty in network exterior orientation and self-calibration that can be encountered with the collinearity equation model when the camera field of view is smaller than around 10°. This paper reports on an investigation that examined two different avenues for improving the self-calibration of long focal length cameras. The first is a re-examination of the linearisation of the collinearity equations with additional calibration parameters, and especially determination of the coefficients in the design matrix corresponding to the interior orientation elements. The second is a new approach to the calculation of object space coordinates by employing an orthogonal projection model that can be formulated as a bundle adjustment with self-calibration. Accuracy aspects of both approaches are discussed and test cases employing both zoom and macro lenses are presented. [ABSTRACT FROM AUTHOR]

Although automated object surface reconstruction via feature-based matching is commonly employed in both stereoscopic close range and topographic photogrammetry, it has rarely been used in conjunction with multi-image, convergent photogrammetric networks. A prerequisite for feature-based matching of distinct points is the application of interest operators to provide the dense arrays of candidate feature points within the images forming the network. This paper describes an evaluation of three interest operators, the Förstner, SUSAN and FAST algorithms, the aim being to assess which is optimal for feature-based matching in convergent close range photogrammetry. Following a brief review of the development of interest operators, a description of the three operators is provided, with the recently developed FAST operator being discussed in more detail. The merits of image pre-processing via the Wallis filter are also outlined, after which the performance of the interest operators is experimentally assessed within an eight-image network on the basis of accuracy of interest-point localisation, detection rate and speed. The findings of the evaluation are that, of the Förstner, SUSAN and FAST operators, the FAST, which has not been employed to a significant extent in photogrammetry to date, is optimal for feature-based matching in multi-image close range networks. [ABSTRACT FROM AUTHOR]

At the present time there is a considerable amount of development interest in the use of terrestrial laser scanner (TLS) data for 3D reconstruction and modelling in the fields of heritage recording and plant documentation. TLS technology produces point clouds of very high density in a very short time. When surveying an object of complex geometry, multiple scans are typically required to generate an occlusion-free 3D model. The principal task associated with building 3D models from laser scanner data in such cases is to transform the scan data from the sensor's local coordinate system into a uniform Cartesian reference datum. This is necessary in order to join the multiple point clouds. This paper presents a registration procedure for TLS data whereby the exterior orientations of the TLS stations are determined via photogrammetry. The images from a digital camera mounted on top of the laser scanner are used first to relatively orient the network of images and then transfer this orientation to the TLS stations to provide exterior orientation and thus a one-step registration of the point clouds from each scanner position. Test results obtained with the proposed approach are presented to highlight its efficiency and accuracy. [ABSTRACT FROM AUTHOR]

A generic network design in close range photogrammetry is one where optimal multi-ray intersection geometry is obtained with as few camera stations as practicable. Hyper redundancy is a concept whereby, once the generic network is in place, many additional images are recorded, with the beneficial impact upon object point precision being equivalent to the presence of multiple exposures at each camera position within the generic network. The effective number of images per station within a hyper redundant network might well be in the range of 10 to 20 or more. As is apparent when it is considered that a hyper redundant network may comprise hundreds of images, the concept is only applicable in practice to fully automatic vision metrology systems, where it proves to be a very effective means of enhancing measurement accuracy at the cost of minimal additional work in the image recording phase. This paper briefly reviews the network design and accuracy aspects of hyper redundancy and illustrates the technique by way of the photogrammetric measurement of surface deformation of a radio telescope of 26 m diameter. This project required an object point measurement accuracy of σ = 0·065 mm, or 1/400 000 of the diameter of the reflector. [ABSTRACT FROM AUTHOR]