Abstract
Comparative return irradiance estimation in a two channel heterodyne lidar receiver should be facilitated by processing finite data samples obtained in a period short compared with that of atmospheric fluctuations affecting beam propagation. Here the performance is compared of two ratio or log-ratio estimators with the assumptions of square law detection, uncorrelated speckle and arbitrary background noise level and sample size. In both cases storage of the sample logarithm is usually advantageous regarding short-term bias and variance especially where the sample size is small and the ratio near unity. Implications for control of overall measurement bias and optimized processing speed are briefly discussed.
