Comparative assessments of peak detection algorithms for underwater laser ranging

Abstract

The cross-section of a laser beam is almost impossible of a perfect circular shape; without doubting most of the cross-sections of laser beams are nearly in ellipse-like form. Therefore, by using an ellipse-like laser beam cross-section for underwater range finding, we are really interested in understanding the quality of range finding with different peak detection algorithms. According to our previous results, we found that the peak detection algorithm of illumination center is good for estimating the location of the laser spot in the image. However, background noise is still a serious problem while selecting a square of pixels for centroid calculation. In this paper, we modified the illumination center to a different algorithm which introduces the technique of the principal component analysis (PCA) to the centroid calculation. This algorithm restricts the selection of pixels to a region within an ellipse for centroid calculation. According to the results of range finding, we found that the ranging quality achieved by using the algorithm of the modified illumination center is much better than that obtained by using the algorithm of the illumination center. Therefore, the algorithm of the modified illumination center is proved to be effective to reduce the effect of background noise on range finding. Also, this result demonstrates that the way of selecting effective pixels for centroid calculation plays an important role.