| Steel ball is the important basis parts,but also is a typical curved surface with high reflection characteristic.Currently,visual inspection in the steel ball surface area has a good application prospect.This paper summarizes the existing research results and designs a new type of light source after experiment verification.It puts forward a new method to detect steel ball surface defects by detecting the integrity of the reflection stripe,and the method of image processing algorithm is studied.On the basis of the circular LED monochrome dome light,combining with the actual experiments,we design a new dome-shaped light source,whose inner wall is modified by patterns with certain regular.It makes the steel ball surface defects appear in the image more clearly.We obtain various defects in the images under the new type of light through experiments.They verify the effectiveness of the light source,and we put forward the new method for surface defect detection by detecting reflection pattern on the surface of the steel ball.To processing the pattern images captured by CCD sensor,this paper puts forward two methods.One is a conventional thinking which based on curvature and the other is an algorithm based on eigenvalue which starting from corner detection.Two algorithms are introduced in detail in this paper.By comparing the characteristics of the two methods,we determine the final image algorithm used in the actual detection.On the basis of sphere surface unfolding method using two CCD cameras,using the new detection method and the corresponding image processing algorithm,we complete the development of the steel ball surface defects automatic sorter.This paper introduces the processing steps in detail,and experiment with the diameter of 1.588 mm and 0.7938 mm steel balls.The detection speed,stability,missing rate of the sorter were evaluated.Results show that the proposed method is effective and reliable for the specular curved surface defect inspection. |
PDF Full Text Request