Research on Visual Detection Method of Sound Film Broken Glue Defect

Broken glue defect is a common defect in sound film dispensing. Aiming at the problem of fuzzy glue region boundary frequently occurring in the detection process, an improved iterative maximum interclass variance method was proposed to detect broken glue defects. Due to the skewness distribution of the gray curve of the sound film, the median value of the gray value of the image was replaced by the average value of the improved iterative method to eliminate the sound film. The initial threshold of the sound film dispensing region was quickly converged. The maximum inter-class variance method was combined to improve the extraction accuracy of the glue region boundary. The image features of single gap, multi gap, and dislocation defect types in qualified and broken products were compared, and the number and offset of glue areas were calculated to identify defects accurately. Fifteen thousand products were randomly selected on the production line for testing, and the results showed that the detection accuracy rate of qualified products was 100%, and the detection accuracy rate of broken glue was 99.1%, which met the needs of enterprises.