An improved method for high-curvature detection with applications to automated inspection

Abstract : Digital curve acquisition with a vision system introduces a large amount of degradation. At the same time ‘corners’ convey important visual information. Unfortunately they are found in the same spectral domain as the quantization noise. Low-pass filtering of the curvature function is compared with more sophisticated methods that keep the angles ‘sharp’. An automated inspection is successfully performed, demonstrating the efficiency of a new method.
Document type :
Journal articles
Complete list of metadatas

Cited literature [6 references]  Display  Hide  Download

https://hal.inria.fr/hal-01119236
Contributor : Radu Horaud <>
Submitted on : Wednesday, February 25, 2015 - 10:58:14 AM
Last modification on : Thursday, January 11, 2018 - 6:15:57 AM
Long-term archiving on : Tuesday, May 26, 2015 - 1:30:17 PM

File

PinedaHoraud1983.pdf
Files produced by the author(s)

Identifiers

Collections

UGA | LAG

Citation

José-Cruz Pineda, Radu Horaud. An improved method for high-curvature detection with applications to automated inspection. Signal Processing, Elsevier, 1983, 5 (2), pp.117-125. ⟨10.1016/0165-1684(83)90018-X⟩. ⟨hal-01119236⟩

Share

Metrics

Record views

219

Files downloads

162