Design And Implementation Of The Bar Recognition And Counting System Based On Image Processing

It is particularly important to calculate the number of bars quickly and accurately during the process of transaction and storage.Currently these are three popular counting methods on the market include manual counting,radio-frequency tag,photoelectric sensors.the manual counting method not only needs intensive labor,and the accuracy of the count also can't achieved hundred percent;though the accuracy rate of RFID tag count method can reach hundred percent,but the spending of this way is too high for the public;however,the photoelectric sensor is not suitable to calculate the number of bars.In recent years,researchers in the field of image have developed a kind of image counting method.Compared with the others,the cost of this method is least and the accuracy rate is highest,At present,it is a popular research method in the field of bar recognition.In order to solve the problem of fast and effective counting in the acceptance of the bar and the storage,this thesis designs a kind of automatic recognition and counting system,the industrial control board as the main atom,cooperate with USB camera as the hardware platform of the system,the visual 2013 studio as a software platform under Windows 7 operating system.In order to solve the problem that the bars are often blocked in the bar and the adhesion between the bars,a series of pretreatment methods are adopted to improve the quality of the bar image.By contrast,using the traditional connectedcomponent labeling counting algorithm for counting,and adding the algorithm optimization of connected domain area,contrast identification mark as the core algorithm of this system,the algorithm can improve the traditional communication domain counting method of omission,misjudgment phenomenon.Install the USB driver in the industrial control board on the atom,the cap function called directly in the OpenCV library to open the camera,Through the ESC key on the keyboard image acquisition and preservation of the control side,algorithm processing on COME1054-IOBoard core board,realize the counting system of bar images called computer vision library OpenCV3.0.This thesis achieved one of the bars image counting system with the using of a improved communication domains counting method and a COME1054-IOBoard as the hardware platform.After dealing with the article's algorithm,both the Identify bars and the unidentified bars will be clearly marked In the bar images,so we can adjust the identification number by adding artificial intervention,to ensure that the identify counting accuracy up to one hundred percent.