Review and Comparison of High-Dynamic Range Three-Dimensional Shape Measurement Techniques

Table 2

Complexity of various HDR techniques.

Technique	Ref.	Number of required projections	Achieved accuracy	Complexity of calculation
Using multiple exposures	[13]	It uses 7 different parameters and needs a total of projections.	The RMS error reaches 0.016 mm within measurement volume size.	Complicated, because of a large number of square root operations needed to calculate the pixel-wise modulation intensity.
Adjusting projected pattern intensities	[24]	If four-step phase shifting algorithm and heterodyne phase unwrapping algorithm are used, it needs projections.	The RMS error reaches 0.012 mm within measurement volume size.	Simple, because it only involves the matrix inverse operations at the pixel level.
Using polarizing filters	[11]	If four-step phase shifting algorithm and heterodyne phase unwrapping algorithm are used, it needs projections, one of which is uniform white light.	The standard deviation is 0.045 mm within measurement volume size.	More simple.
Using color invariants	[38]	If quaternary color-coded patterns are used, it needs 4 projections.	The overall average deviation is 0.506 mm within measurement volume size.	Most simple.
Photometric stereo	[52]	It needs 30 projections in different directions.	The RMS error reaches 0.005 mm within measurement volume size.	More complicated.