| Abstract [eng] |
So far, laminated chipboard determining the quality challenges of quality controller, using a variety of measurement tools, relying on their own vision and other senses and skills. Because the man has become the slowest link in the production process, the aim is to create a laminated chipboard quality inspection system prototype for the inspection hole after drilling, using a computerized image analysis system. In order to investigate what is used for image analysis methods to identify objects on the first section of this paper carried litertūros review. Finding out non-standard methods, their principles of operation and application areas. The second chapter analyzes the development of prototypes to choose methods such as Sobel operator, Canny edge detector, Laplace operator, Eigen decomposition, Moravec corner detection, Harris corner detector. They found that all of these algorithms has different parameters, allowing image processing volumes and opportunities. Also selected image comparison method - linear correlation. The third part deals with assembling the prototype hardware and software. The listed hardware components, their mutual connection and influence the functioning of the more one looks at the key parts such as camera settings. The influence of light intensity of the importance of the results obtained. The following software and overall system performance flow diagram. Analyzed and illustrated the principles of operation. Further work is carried out the verification and performance test. Trying to camera settings, lighting strength, pixel how much influence the reliability of the system. Selects the most appropriate algorithm parameters to different factors, wanted the best option, for precise and rapid remediation of defects. |
