Singular Value Decomposition Filter: an Effective Method to Enhance the SwissRanger SR-3000's Range Images
Guruprasad M. Hegde, Cang Ye · 2008
This paper presents the research results on the enhancement of the SwissRanger SR-3000's range images by a Singular Value Decomposition (SVD) filtering method. The image enhancement is performed by converting a conventional range image into an enhanced range image where each pixel's intensity embodies the surface normal and depth information of the corresponding pixel in the original image. This range image representation makes an object's edges distinctive. But it corrupts the object's surfaces due to the noise in the range data. We propose a filtering method based on the SVD to alleviate the surface corruption and preserve the edges. The efficacy of the proposed method is validated by numerous experiments in various environments. Index Terms—Singular Value Decomposition filter, normal-enhanced range image, flash LADAR.