Research Of Paper Defects Detection System Based On Distributed Technology

The paper defects detection is one of the most important elements of qualitymanagement and control of the paper sheets. The modern paper making machine hasincreasing web width and machine speed. In order to improve the quality of papersheets and production efficiency, the online paper defects detection system based onmachine vision is required and imperative to replace the traditional manual inspectionwhich is subjected to low efficiency, high labor intensity and low calibration accuracydue to human factors. In this thesis, a distributed paper defects detection system basedon multi-camera is designed and implemented to meet the requirements of modernpaper making production line with characters of wide format, fast speed and highexpectation of accuracy. This system aims to provide the high precise online real-timedetection of the common paper defects. The main contents of this thesis is showingbelow:Firstly, on the basis of the requirements of paper defects detection technology andanalysis of typical machine vision inspection system, a overall design of the distributedpaper defects detection system is presented. We introduces the operation principle andtechnological process, builds a distributed paper defects detection system, analyzes theperformance and model of the module of the hardware system.Secondly, the synchronization of real-time image acquisition and multi-camera isanalyzed in this thesis, which introduces the acquisition principle of line-scan cameraand the content of GigE Vision protocol. Queue buffer structure is used to ensure thereal-time acquisition of the large amount of images, which avoids the frame loss duringthe process of acquisition. By external trigger mode, synchronous acquisition of themulti-camera becomes possible.This paper does a study of the network communication of distributed paper defectsdetection system. This network communication is effectuated by the Socketprogramming and describes the communication process of client and server. The"heartbeat detection" allows the online self-diagnosis of this inspection system, whichmeans the client could be able to reconnect the server automatically when theirdisconnection is detected by the system.Finally, the software design of distributed paper defects detection system isintroduced. It is designed to provide detection algorithm for paper defects. Based on the Visual C++6.0software platform, the applications of this software can include paperdefects images acquisition and transmission, paper defects extraction and classification,online paper defects monitoring and historical data inquiry. A easily controlledman-machine interface is also presented. This system has been analyzed and tested tomeet the design requirements.