Design And Implementation Of Real-time Flaw Detecting System For Plastic Membrane

Surface defect is a key factor for the quality of plastic membrane.Traditionally,the defect detection is performed artificially,nevertheless,this method is limited by the shortcomings of low speed and short precision.A PC-based real-time visual detecting system is desperately needed in factory,though currently most defect detection systems only uses algorithms or software programs to do some matching work for textured materials,but not apply with the plastic membrane products.For products without regular pattern characteristics,the process of image acquisition and flaw detecting have to connect with the transmission system to get the real-time images,especially in cases of using linear CCD as the image capturing device. On the other hand,plastic membrane is wide-sized and the flaws are all small,so it's a challenge to solve the problem between large amount of data and required precision.In this thesis,a PC-based multi-linear CCD real-time detection system is designed for plastic flaws detecting.The main contribution of this thesis is as follows:Firstly,designs and constructs a PC-based multi-CCD real-time image acquisition system,the principle as well as the hardware and software components is described in detail.Secondly,focus on the coordinational control between the movement control module and image acquisition module.In the movement control module,the thesis proposes the method of "stepper motor + double encoders" to complement the closed-loop control,than uses the CLF node in Lab VIEW to call the dll of a program complied in VC platform to control googol motion card.The results of stepper motor test prove that the control method solves the out-of-step problem well,and achieves a high degree of accuracy pulse output.In the image acquisition module,according to the characteristics of linear CCD scanning,the thesis proposes a method based on position feedback loop to trigger multiple linear CCD synchronously.A Ring buffer solution in LabVIEW platform is also studied to solve the problem of large quantity of data transmission.Finally,introduces the defect recognizing algorithms.Since IMAQ Vision provides powerful functions,we complement the detection of the plastic flaws and give an elegant interface on the platform.A network-monitoring module is also built to share the results of detecting to the whole factory network.The real-time flaw detecting system constructed in the thesis might be used wildly in the occasions of belt,wide object.However,there are still many places that need to be improved.At the end of the thesis,the further development trends of this visual system are discussed.