Skip to Main Content
This paper takes a linear least-squares approach to develop a procedure for reducing effects of environmental luminance and turbidity on underwater imaging. The first step is to divide all pixels in each row of the original image to several data sets. Each data set is fitted to a straight line by the least-squares error method. Once each row has been processed, the profile of the row estimate is saved as a new image. The next step is to fit all pixels in each column of the new image by repeating the least-squares error process. Then the profile of the column estimate will provide a better imaging quality than the original one if values of the design parameters are properly assigned. As an example of an optical triangulation system, this paper carried out a series of DOE process runs to study effects of the design parameters on quality of range finding. To make the numerical technique robust against noises, both environmental luminance and turbidity are forced into the experiments by utilizing an outer array. From results of experiments and ANOVA, it is concluded that the sampling range utilized for peak detection, the number of line segments in each row, and the overlapping ratio of adjacent lines in each row are influential on changing quality of range finding.