- Author
-
Gilsinn, D. E.
|
Cheok, G. S.
|
O'Leary, D. P.
- Title
- Reconstructing Images of Bar Codes for Construction Site Object Recognition.
- Coporate
- National Institute of Standards and Technology, Gaithersburg, MD
- Report
-
NIST SP 989,
September 2002,
- Distribution
- AVAILABLE FROM: National Technical Information Service (NTIS), Technology Administration, U.S. Department of Commerce, Springfield, VA 22161. Telephone: 1-800-553-6847 or 703-605-6000; Fax: 703-605-6900; Rush Service (Telephone Orders Only) 800-553-6847; Website: http://www.ntis.gov AVAILABLE FROM Superintendent of Documents, U.S. Government Printing Office, Mail Stop SSOP, Washington, DC 20402-0001. Telephone: 202-512-1800. Fax: 202-512-2250. Website: http://www.bookstore.gpo.gov
- Book or Conf
- International Symposium on Automation and Robotics in Construction, 19th (ISARC). Proceedings. National Institute of Standards and Technology, Gaithersburg, Maryland. September 23-25, 2002,
363-368 p.,
2002
- Keywords
-
robotics
|
construction
|
bar codes
|
image processing
- Identifiers
- deconvolution; LADAR; object recognition; sparse matrix
- Abstract
- This paper discusses a general approach to reconstructing ground truth intensity images of bar codes that have been distorted by LADARoptics. The first part of this paper describes the experimental data collection of several bar code images along with experimental estimates of the LADAR beam size and configuration at various distances from the source. Mathematical models of the beam size and configuration were developed and were applied through a convolution process to a simulated set of bar code images similar to the original experiment. This was done in order to estimate beam spread models (beam spread models are unique to each specific LADAR) to be used in a deconvolution process to reconstruct the original bar code images from the distorted images. In the convolution process a distorted image in vector form g is associated with a ground truth image f and each element of g is computed as a weighted average of neighboring elements of f to that associated element. The deconvolution process involves a least squares procedure that approximately solves a matrix equation of the form Hf = g where H is a large sparse matrix that is made up of elements from the beam spread function.