Evaluation and testing of the Zeiss SCAI roll film scanner
暂无分享,去创建一个
In the summer of 1996 the Swiss Federal Office of Topography (L+T) performed in cooperation with ETH benchmark tests at different companies in order to evaluate three photogrammetric film scanners. After the tests a Zeiss SCAI scanner was acquired, installed, and tested in spring 1997 as to whether it fulfils the contract specifications. In this paper the results of these two tests performed with the SCAI, and in particular geometric and radiometric investigations, are presented. Good quality test patterns and accurate processing methods for the performance evaluation have been employed. The geometric tests include geometric errors, misregistration between colour channels, geometric repeatability, and determination of the geometric resolution. The radiometric tests include investigations on noise, linearity, dynamic range, spectral variation of noise, and artifacts. After a brief description of the scanner, details on the above investigations, including test patterns, processing methods, results and analysis are presented. Regarding the geometric accuracy the RMS was 2.0-2.3 μm and the mean maximum absolute error 6.0-8.1 μm. The errors are bounded, i.e. the maximum absolute error is 2.7-3.6 RMS. Systematic co-registration errors in scan direction between the colour channels of up to ca. 2.5 μm have been observed. The short term repeatability was very high. The radiometric noise level is 1-1.6 and 0.9-1.3 grey values for 7 and 14 μm scan pixel size respectively. The dynamic range is 1.6-1.9D and 1.75-2.05D for 7 and 14 μm respectively with a good linear response up to this value. Quite some artifacts and electronic noise up to 5 grey values were observed. There were no significant differences between R, G, B channels with respect to geometry but the B channel exhibits more radiometric noise and less dynamic range than the R and G. The effective y-pixel size can be smaller than the nominal one by 10% (exposure times > 4.5 ms) and up to 50% (exposure times < 4.5 ms). This reduction can be negatively influenced by inappropriate selection of exposure time and scan speed by the user. The software is generally positive but improvements, especially regarding automatic density determination and automatic setting of the scan parameters including scan speed and exposure time, are needed. Latter improvements have been partly implemented in the new software version PSC 3.0 which was not available to us at the time of these investigations.
[1] A. Gruen. ADAPTIVE LEAST SQUARES CORRELATION: A POWERFUL IMAGE MATCHING TECHNIQUE , 1985 .
[2] Emmanuel P. Baltsavias,et al. Multiphoto geometrically constrained matching , 1991 .