Font Size: a A A

Research On Film’s Defect Detection And Cutting

Posted on:2015-09-28Degree:MasterType:Thesis
Country:ChinaCandidate:W K LvFull Text:PDF
GTID:2298330431989269Subject:Physical Electronics
Abstract/Summary:PDF Full Text Request
There are two problem existed in film’s cutting stock procedure.The illogical cutting scheme lead to large trim loss and the detects existed inthe film make the film’s quality unqualified. In order to solve the aboveproblems, a film’s defect detection and cutting system design wasdesigned.The film’s cutting stock algorithm and detect’s detecting algorithmwere researched in this paper. Following are the main contents in this paper:An improved random search algorithm was proposed. A hierarchicallyrandom search algorithm was proposed on the basis of the traditional randomsearch algorithm. The cutting pattern was divided into two parts, the first dealingwith iterms random combination according to the order number, and the second withthe length constraint. In the solution, each pattern was selected from the collection ofa number of alternative patterns through performing heuristic evaluation, and theselected patterns would be used as many as possible. The order number was updatedbefore generating pattern for each item until the cutting was completed.Experimental results show that the material utilization is better and the pattern countis less, which verify the effectiveness of the proposed algorithm.A new algorithm was designed to solve the one-dimensional cutting stockproblem that the film is piecewise and discontinuous. A solution was proposed tosolve the piecewise and discontinuous model base on the hierarchically randomsearch algorithm.The Experimental results show the effective of the algorithm withtwo piecewise model.Detection of the film defects. The defects in the film was processed by thedigital image processing technology.Firstly,several image preprocessingoperations such as graying, denoising, corrosion and inflation were processedin the image.Then the OTSU algorithm was adopted to segment the image intoa binary image.Finally,a boundary contour detection was adopted to processthe segmented image.And the information of the detects’size and locationwould be detected and imported to the database,constituting an integratedsystem with the cutting stock system.
Keywords/Search Tags:One-dimensional cutting, Random search, Heuristic algorithm, defectdetection, image processing
PDF Full Text Request
Related items