Maximum-likelihood multiresolution laser radar range imaging

Author

Greer, Donald R. ; Fung, Irene ; Shapiro, Jeffrey H.

Author_Institution

Dept. of Electr. Eng. & Comput. Sci., MIT, Cambridge, MA, USA

Volume

6

Issue

1

fYear

1997

fDate

1/1/1997 12:00:00 AM

Firstpage

36

Lastpage

46

Abstract

Maximum-likelihood range imaging is considered for pulsed-imager operation of a coherent laser radar. The expectation-maximization (EM) algorithm is used to develop an explicit procedure for maximum-likelihood fitting of a multiresolution (wavelet) basis-at a sequence of increasingly fine resolutions-to laser radar range data. Specialization to the Haar-wavelet basis yields a procedure that is both computationally efficient and numerically robust. Basic analytical properties of the estimation algorithm and its performance are presented, along with results based on simulated and real laser radar range data. It is shown that the weights associated with the expectation-maximization iterations provide a reliable indicator for terminating the coarse-to-fine resolution progression. At the weight-determined stopping point, estimation performance approaches the ultimate limit set by the complete-data bound

Keywords

image resolution; iterative methods; laser ranging; maximum likelihood estimation; optical radar; radar imaging; transforms; wavelet transforms; Haar-wavelet basis; coarse-to-fine resolution progression; estimation algorithm; expectation-maximization algorithm; iterations; maximum-likelihood fitting; multiresolution basis; multiresolution laser radar range imaging; performance; pulsed-imager; weight-determined stopping point; Algorithm design and analysis; Analytical models; Computational modeling; Image resolution; Laser radar; Maximum likelihood estimation; Optical pulses; Performance analysis; Radar imaging; Robustness;

fLanguage

English

Journal_Title

Image Processing, IEEE Transactions on

Publisher

ieee

ISSN

1057-7149

Type

jour

DOI

10.1109/83.552095

Filename

552095