This paper deals with the problem of defect detection on highly reflective surfaces making use of vision systems. A new mechatronic system has been developed, based on a nonflat mirror. According to the method described in this paper, the light rays emitted from a source hit a suitably designed nonflat mirror, and are reflected so as to illuminate the curved surface under investigation. The path of the light rays from the source of light to the mirror and then to the object surface is mathematically traced making use of the optical geometry laws. After the reflection on the object surface, the light rays are collected by a charge-coupled device (CCD) camera and elaborated by a vision system, which manages to detect the surface defects as shadows of various shape and size within the picture. Simulations have been carried out in order to provide the optimal mirror shape. Moreover, a prototype of the mechatronic system, including the synthesized mirror, has been built to perform some experimental tests to validate the method. The results, reported in the paper, definitively show the effectiveness of the proposed method.
[1]
Giovanni Tantussi,et al.
The quality control of natural materials: defect detection on Carrara marble with an artificial vision system
,
1997
.
[2]
Richard W. Conners,et al.
A machine vision system forautomatically grading hardwood lumber
,
1992
.
[3]
J. B. Gomm,et al.
Inspection of ceramic tableware for quality control using a neural network vision system
,
1994,
Electronic Imaging.
[4]
Nasser Sherkat,et al.
Real-time vision for automatic lace cutting
,
1994,
Electronic Imaging.
[5]
Andrew Blake,et al.
Detecting Specular Reflections Using Lambertian Constraints
,
1988,
[1988 Proceedings] Second International Conference on Computer Vision.
[6]
Yinhong Li,et al.
Detecting and characterizing small voids in mostly diffuse materials
,
1994,
Other Conferences.