Digital image correlation through a rigid borescope Phillip L. Reu1* 1Sandia National Laboratories, PO Box 5800, Albuquerque, NM 87185 ABSTRACT There occasionally occur situations in field measurements where direct optical access to the area of interest is not possible. In these cases the borescope is the standard method of imaging. Furthermore, if shape, displacement, or strain are desired in these hidden locations, it would be advantageous to be able to do digital image correlation (DIC) through the borescope. This paper will present the added complexities and errors associated with imaging through a borescope for DIC. Discussion of non-radial distortions and their effects on the measurements, along with a possible correction scheme will be discussed. Keywords: digital image correlation, distortion correction 1. INTRODUCTION Digital image correlation (DIC) has become a standard tool to measure 2D and 3D shape, motion and deformation. To conduct DIC in a hidden area a borescope may be a convenient method of imaging. There are two types of borescopes: fiber-optic scopes which use a lens and intervening fiber to relay the scene to the camera, or a rigid scope which uses a series of relay lenses to transfer the scene to a camera. Both types could be used for DIC; however the fiber-optic solution has severe image resolution limitations. Because of this, it was decided to use a rigid borescope system to attempt DIC in a hidden cavity. At the moment this limits the application to 2D DIC, however, arrangements could be imagined where 3D DIC could be conducted. Because of the complex optical path, distortions will be larger than a typical camera lens and likely not radial in nature. This paper discusses a proof-of-concept experiment conducted to determine the optical distortions and resulting errors in a DIC analysis. Furthermore non-radial distortion correction methods, which were originally developed for use in stereo-microscopes [1], will be used to minimize the distortions. 2. BORESCOPE DISTORTION TESTING 2.1. Experimental setup to determine lens distortions of a rigid borescope A rigid borescope, model D8094K-101, was obtained from Lennox Instruments, a commercial manufacturer of both rigid and flexible borescopes. The borescope can be extended to various lengths using extension tubes. The optional prism objective was used on the end to create a viewing direction perpendicular to the borescope shaft. Dirt on the internal borescope lenses turned out to be problematic, as dirt spots look like a stationary speckle in the DIC and corrupt the results. After carefully cleaning the objectives, they were setup on an optical table looking at a small speckle pattern affixed to an xy-stage. The imaging section of the borescope is shown in Figure 1. The shorter borescope with only the eyepiece and imaging section was approximately 1 meter in length; the full borescope with two extensions was 2.8 meters in length. The camera attachment for the rigid borescope was not appropriate for the Point Grey 5 Megapixel camera and caused severe vignetting on the 2/3-inch detector. The standard camera sold with the system is a much smaller and lower resolution detector and would not yield the desired spatial resolution for our application. To overcome the vignetting, a 75-mm lens was used to relay the image from the eyepiece to the camera (see Figure 1). With care and alignment, the vignetting was able to be minimized. A preliminary study showed that the effect of this additional lens was minimal on the distortions. The same 75-mm lens was used along with the same speckle pattern area to test the distortion removal algorithms and for comparison with the borescope results. * plreu@sandia.gov T. Proulx (ed.), Optical Measurements, Modeling, and Metrology, Volume 5, Conference Proceedings of the Society for Experimental Mechanics Series 9999999, DOI 10.1007/978-1-4614-0228-2_19, © The Society for Experimental Mechanics, Inc. 2011 141
RkJQdWJsaXNoZXIy MTMzNzEzMQ==