Research On Cutter DPM Code Recognition In Complex Environment

As an emerging information technology, the2D bar code recognition has extensiveand sophisticated applications in the different fields, such as products sales,advertising, Internet, mobile communication, e-commerce, etc. In industry production,since the2D bar code has the advantages of small footprint, large information, highreliability, strong error correction and high security, it becomes the first choice in thetechnology of Direct Part Marking (DPM). As the unique and permanentidentification during the life cycle of productions, the2D bar code is called the directcode or the DPM code, which can realize real-time tracking and effectivemanagement of the production information. However, due to the inevitable negativefactors of cutting tools and many kinds of pollution factors during the process ofindustrial production and distribution, it is difficult to make the direct marking in thecutter surface. Therefor，it is urgent to study on the method of2D bar code reading forcutter.This thesis based on the cutter DPM code for production workshop, aiming at theproblems of uneven illumination, surface distortion, corrosion and wear factors,which the traditional technology could not solve, put forward a set of reading solutionfor cutter DPM code in complex environment and try to implement the accuraterecognition. The main research contents and results are presented as follows:(1) Put forward a kind of read scheme suit for the cutter DPM code with unevenillumination. For the bar code image with low contrast, using artificial colonyalgorithm to select the parameters of incomplete Beta function adaptively, enhancedthe image contrast. For the bar code image with high light, analyzed the producedcauses of high light and put forward a suppression algorithm based on the optimizedlight source device and stochastic resonance principle. The experimental results showthat the algorithm could suppress highlights from the source and have good treatmenteffect on the problem of uneven illumination.(2) Presents two kinds of reading schemes for distorted DPM code on cylindricalcutter. The first algorithm is based on the principle of the perspective projection andcamera calibration. The distortion correction model in a non-linear lens distortion condition is established and the correction of the cylindrical DPM code is completed.The second algorithm is based on the principle of image mosaic combined with theimage sequence capture device to achieve rough and accurate registration of the barcode image. Due to the DPM code image in small gap can be approximated asuniform without distortion, the spliced images can counteract the twisting modedistortion. Several groups of experiments show that both of the two methods arepracticable.(3) Proposed a cutter classification and identification method based on multipleinformation fusion. Using the features of weight, outline, texture of residual bar code,established the identification model respectively, and completed the determination ofcutter products and the recovery of the bar code information through fusion ofmultiple comprehensive decision. Experiments show that the proposed method withgood effectiveness is practicable.(4) Based on the above researches, a special system for reading DPM codes of tools inworkshop is developed. The design and implementation of the hardware devices andsoftware systems are introduced, and the specific application of the system withcommon tools is demonstrated as an example.In this paper, the research results will help to further deepen the DPM technology andits application. The preliminary experiments tested in workshop show that the relatedprinciple and method of the thesis have some practical value for improving the effectof DPM code identification in complicated environment.