Abstract: To detect and quantitative evaluate wrinkles in the film of soft cigarette packets, a film defect inspection system based on spectral confocal technology was designed. The designed system used a spectral confocal sensor to construct the three-dimensional morphology of the packet film, and adopts the random sample consensus(RANSAC)algorithm for accurate point cloud segmentation and wrinkle region extraction. A mathematical model was then developed to evaluate the wrinkle status, including wrinkle rate and wrinkle intensity. The results showed that: 1)The relative error of measurements of standard components of transparent glass and metal sheets were all below 1%, with the relative standard deviation (RSD) of 10 repeated measurements also below 1%. 2) The inspection results of 15 commercially available cigarette brands in soft packets showed that their wrinkle rates varied between 4% and 24%, while the wrinkle intensity ranged from 0.02 to 0.48. There was a large variation in the quality of the film packaging of different brands of soft cigarette packets. This method provides a new digital nondestructive tool for evaluating the film packaging quality of soft cigarette packets.