| Machine vision is one of the main means that industrial robot gets surrounding information. It can improve the ability and flexibility of robot. According to such application occasion as assembly line, this dissertation develops identification software on the basis of knowledge of working situation and parts. It offers high efficiency and stability. During the whole developing process of this system, the main work includes: 1. Chooses the hole camera with one rank lens radial aberration as the CCD camera calibration model, and chooses Radial Alignment Constraint method to work out extrinsic and intrinsic parameter of CCD camera.2. In the image preprocessing, it analyzes results of several kinds of smoothing filter; median filter stands out for its good edge maintenance and adaptability. This article takes median filter as the way to process the collected image. Also, it uses optimal thresholding arithmetic to segment image and use sequential arithmetic to mark connected component.3. Analyses the classical edge detectors, and put forward the edge detection method based on binary connected component. It not only makes use of segmentation image, but also avoids inspecting plentiful non-object area, and enhances the detection speed. In the given application situation, this method completely satisfies demand and has superiority.4. According to the difference of dimension and shape of parts, the dissertation identifies and orients them by position, area, circularity, and rectangularity. In addition, it analyzes the characteristics of different parts and ensures their direction in different ways which is simple and turns out good results.5. Analyzing the Serial communication protocol between PC and PLC controller, Using communication test program, which is based on MSComm controller, monitor hypogyny computer.
|
PDF Full Text Request