Surface Roughness Extraction by Gibbs Random Fields of Laser Speckle Pattern Texture

摘要

Based on computer texture analysis methods, the relationships between laser speckle texture features of grinding surfaces and surface roughness are investigated. The laser speckle texture pictures of different surface roughness are acquired by a simple equipment which consists of a digital camera and a diode laser. The texture method based on Gibbs Random Fields model is used to analyze laser speckle patterns. Gibbs texture features with the second-order neighborhood are extracted. The experiment results display that the surface roughness information included in the laser speckle texture pictures is monotonous with β_2～β_5 Gibbs texture features. For comparing, normalized texture features has been done. This method can extract object's surface roughness information which is the same material and machined by the same method through calibrating beforehand.