Experimental Research On Shrinkage Direction Of Injection-molded Parts Based On Digital Image Processing

Injection molding is a very complex process coupled by many dynamic factors. Shrinkage of the plastic product is a key factor which affects the quality of the molding product, not only degrading precision and apparent quality of the plastic product, but also causing structural deformation and intrinsic quality defects. Therefore, shrinkage has become a major problem in the field of injection molding. In order to explore the relationship between the shape of the plastic product and direction of shrinkage, seven groups of experiments were executed and a image processing system including the image stitching function was developed based on the image processing module of the Matlab software, realizing the test of shrinkage direction of plastic parts.Firstly, shooting equipments including the visual mold and a screw-driven moving mechanism were designed and fabricated independently. The screw-driven mechanism for moving of the camera provided foundation for precision shooting and image stitching. After 7 different cavities were designed and fabricated, seven groups of visual experiments were executed. After the plastic part had been kept in cavity for 24 hours, a digital single lens reflection camera was employed to take the shrinkage picture.Secondly, a new denoising method based on human-computer interaction in the image processing was presented, effectively solving the interference points during the curve fitting process. An improved Harris corner algorithm was proposed, which not only retains the corresponding pixel values of the higher interest, but also makes capture of matching points more conveniently. Matching accuracy of the improved Harris operator and of curve-fitting method was coMPared and analyzed. Analyzing results showed that the improved Harris operator is less accurate than the method based on curve-fitting. Then curve-fitting method was used in this paper for acquiring the shringkage direction of the plastic parts.Thirdly, the developed image processing system was used for No.1 to No.4 groups of experiments to distinguish the boundary of the mold cavity and of the plastic part in the pictures. The data of shrinkage direction were got through locating the feature points of the mold cavity and of the plastic part by using line fitting and intersectant point computing. For No.5 to No.7 groups of experiments, the mold cavity containing the plastic part was shot block by block; after the block images were stitched, the method used for No.1 to No.4 groups of experiments were adopted to get the feature points of the mold cavity and of the plastic part.Finally, the conclusions were summarized through comprehensive analyzing seven groups of injection molding shrinkage tests via digital image processing. The shrinkage of injection-molded plastic parts is affected by combination of the gate and of the part's Centroid. If contraction of the plastic part is not hindered by the mold cavity, the shrinkage direction of the plastic part points to a location between the gate and the Centroid of the plastic part. If there is obstruct coming from the mold cavity to contraction of the plastic part, shrinkage of the plastic parts shows a certain degree of partial shrinkage.