Surface roughness measurement for castings based on wavelet energy

In this paper, we propose a novel surface roughness measurement method for castings based on wavelet transform technique. The surfaces of castings appeared as random and isotropic textures in digital image. The method of assessing surface roughness is based on two-dimensional wavelet transforms, which are performed on gray-level images of cast surfaces. Wavelet transforms and decompositions are implemented for calculating the wavelet energy. Furthermore, high-frequency wavelet energy ratio (HFWER) is proposed and calculated as a feature to assess the surface roughness in frequency domain. Experimental results on casting specimens containing several groups of casting surface images showed that the proposed method is efficient.