DocumentCode
3062428
Title
Undue influence: Mitigating range-intensity coupling in AMCW ‘flash’ lidar using scene texture
Author
Godbaz, J.P. ; Cree, M.J. ; Dorrington, A.A.
Author_Institution
Univ. of Waikato, Hamilton, New Zealand
fYear
2009
fDate
23-25 Nov. 2009
Firstpage
304
Lastpage
309
Abstract
We present a new algorithm for mitigating range-intensity coupling caused by scattered light in full-field amplitude modulated continuous wave lidar systems using scene texture. Full-field Lidar works using the time-of-flight principle to measure the range to thousands of points in a scene simultaneously. Mixed pixel are erroneous range measurements caused by pixels integrating light from more than one object at a time. Conventional optics suffer from internal reflections and light scattering which can result in every pixel being mixed with scattered light. This causes erroneous range measurements and range-intensity coupling. By measuring how range changes with intensity over local regions it is possible to determine the phase and intensity of the scattered light without the complex calibration inherent in deconvolution based restoration. The new method is shown to produce a substantial improvement in range image quality. An additional range from texture method is demonstrated which is resistant to scattered light. Variations of the algorithms are tested with and without segmentation - the variant without segmentation is faster, but causes erroneous ranges around the edges of objects which are not present in the segmented algorithm.
Keywords
CW radar; deconvolution; image restoration; image segmentation; optical radar; radar imaging; AMCW flash lidar; deconvolution; full-field amplitude modulated continuous wave lidar systems; image restoration; image segmentation; range image quality; range-intensity coupling mitigation; scene texture; Amplitude modulation; Image segmentation; Laser radar; Layout; Light scattering; Optical coupling; Optical modulation; Optical reflection; Optical scattering; Time measurement;
fLanguage
English
Publisher
ieee
Conference_Titel
Image and Vision Computing New Zealand, 2009. IVCNZ '09. 24th International Conference
Conference_Location
Wellington
ISSN
2151-2205
Print_ISBN
978-1-4244-4697-1
Electronic_ISBN
2151-2205
Type
conf
DOI
10.1109/IVCNZ.2009.5378390
Filename
5378390
Link To Document