Study on evaluating seam pucker of clothing based on binocular CCD camera

Three-dimensional information of seam pucker on sewed sample can be obtained with binocular CCD camera, which avoids the disturbance of color and pattern to seam pucker information. Then the data along Z axis are analyzed in order to distract characteristic parameters that can reflect different seam pucker. Based on these characteristic parameters, Probabilistic neural network model is constructed to objectively and quantificationaUy evaluate the seam pucker of clothing. By tested, this model is available, which can avoid the fuzziness of subjective evaluation.