Parameter de-correlation and model-identification in hybrid-style terrestrial laser scanner self-calibration

摘要: Abstract One of the important systematic error parameters identified in terrestrial laser scanners is collimation axis error, which models non-orthogonality between two instrumental axes. The quality this parameter determined by self-calibration, as measured its estimated precision and correlation with tertiary rotation angle κ scanner exterior orientation, strongly dependent on instrument architecture. While generally very high for panoramic-type scanners, it comparably poor hybrid-style instruments. Two methods improving hybrid self-calibration are proposed herein: (1) inclusion independent observations ; (2) use a new model. Five real datasets were captured different to test each method’s efficacy. first method achieves desired outcome complete decoupling from , shown that simply transferred other model variables. second partial de-correlation acceptable levels. Importantly, does so without any adverse, secondary correlations therefore recommended future use. Finally, identification has been greatly aided previous studies graphical analyses residuals. This paper presents results showing architecture dependence technique, revealing limitations scanners.