Image Classification And Defect Recognition In The Process Of Polymer Ultrasonic Sealing

Polymer ultrasonic sealing is a green and effective assembly process and an advanced assembly technology that is widely used.With the diversification of material sealing and various polymer sealing technologies,ultrasonic polymer sealing has been widely used with its significant advantages of low cost and high efficiency.The thermoplastic polymer ultrasonic sealing process and the quality of the sealing are very random,which is necessary for the sealing process and the identification of defects on the sealing surface.This article is based on the image in the ultrasonic sealing process,and based on MATLAB to identify the defects in the image and the sealing process.Firstly,the thermoplastic polymer ultrasonic sealing image is preprocessed,including image denoising and enhancement,which is conducive to image feature extraction and image segmentation.Perform regional positioning and segmentation of the thermoplastic polymer ultrasonic sealing image,effectively mark the segmented area,and calculate the defect area,roundness,etc.Secondly,analyze the spectral energy spectrum of the thermoplastic polymer ultrasonic sealing image,and observe the energy distribution of the entire sealing process and the distribution of energy at different frequencies.The feature extraction of the thermoplastic polymer ultrasonic sealing image is carried out by statistical method,mainly using gray difference statistics and gray co-occurrence matrix method to extract the image texture feature,in which the gray co-occurrence matrix method scans the sealed image horizontally and vertically to extract The obtained feature matrix is labeled,and the training set and the test set are made.Combined with the neural network,the image of the ultrasonic sealing process of the thermoplastic polymer is classified.Finally,the dimensionality of the feature matrix of the thermoplastic polymer ultrasonic sealing image is reduced,and the neural network is optimized to improve the recognition speed and recognition rate.The image defect recognition rate of thermoplastic polymer ultrasonic sealing is over 85%,and the recognition rate of polymer sealing process is over 85%.Image recognition technology is feasible in the classification and recognition of thermoplastic polymer ultrasonic sealing images.It is also effective in the identification of defects in the ultrasonic sealing surface of thermoplastic polymers,and it plays an effective role in monitoring the ultrasonic sealing process and sealing quality of the polymer.