Abstract

If a lidar system emits long rectangular sensing pulses, the lidar return signal F(t) at the moment t after the pulse emission will have low resolution and will be described by the equation where c is the speed of light, Φ(z) is the maximum-resolved (δ-pulse) lidar profile, and t is the rectangular-pulse duration; the height of the rectangular-pulse shape is assumed to be equal to unity. In order to improve the lidar resolution, we have developed earlier^1,2 a simple recurrence algorithm to invert Eq. 1 with respect to Φ(z), namely where Φ(z) is supposed to be known in some cτ/2-long initial spatial interval and F′(t) is the first derivative of F(t). Be sides simplicity, another feature of this algorithm is that it is not so sensitive to noise as compared to some other deconvolution algorithms we have developed.² In this work we estimate the error caused by rectangular-pulse approximation of various sensing laser pulses in order to outline the conditions under which this approximation is acceptable, and, consequently, useful because of its advantages.

© 1994 IEEE