The Research Of Plate Defect Detection Based On Machine Vision Technology

With the development of industrial automation, machine vision systems have been applied to various industries'product defect detection. Taking pulp environmen--tal protection plate surface defects as the research object, this paper uses the object-oriented language C# and image processing software HALCON to program jointly and has carried on some beneficial researches on the tests of different kinds of defects in the plate. By building a set of bright field illumination detection device of automatic water feeding, fliping, sorting, it can detect faults in positive and negative plates respectively and offset the limit due to single road lighting--cannot detect the shallow depth of defects and the defects existing at the bottom of the plate. According to this kind of plate defect feature differentiation, it can be divided into different types of defects, using a variety of optimal image algorithms for joint detection, and in this way improving the defect recognition rate and the stability of the system, finally realizing the real-time automatic test plate breakages, non-standard size and circular degree, edge defects or bumps, dirts, scratches and surface residual wires and other defects.And by applying this system to the actual plate production line, it can resolve a series of problems and limitations about the artificial detection plate production line.Firstly, this article reviews the technology background of the machine vision and image processing used for product appearance detection, and then introduces the system structure of the machine vision engineering and development train of thought in this project, and the product features of the pulp environmental protection plate, discusses the reason of plate surface defects from the point of production process, and analyses the image characteristics of all kinds of defects, then puts forward the basic purpose of plate surface defect detection.Next, it introduces the self-check machine device's modules composition and defect detection process. This system is divided into four modules:image processing module, software module, the mechanical structure and electric transmission module, and elaborates the operation process, control parameters and location of each module in detail.In addition, according to the different characteristics of the plate defect, designing corresponding detection scheme, using threshold segmentation, neighbor--hood area of growth and filtering to realize the extraction of the plate area and regional division, the plate is divided into internal round, small dark ring, internal ring, ring and edge ring area. And each area is studied by using Hough transform circle of coarse detection.Finally, it introduces the shape, defect detection, edge detection smudgy wire detection, scratches and other defect detection scheme applicable to a variety of types of plate. Through data analysis and comparison, it shows the effect of various detection scheme.At last, a summary of the shortcomings in the work is made, and subsequent improvement scheme is put forward.