Research And Application On The Non-contact Size Inspection Method Of Parts Based On CCD

In recent years, with the development of industrialization and information, parts inspection technology based on CCD has replaced the traditional manual measurement and formed a certain scale of industry. Compared with the traditional detection methods, this method has the ad-vantages of non-contact, fast response, automation and convenient operation. But the current popular detection methods based on CCD can only meet the detection of specific models of parts and need for manual interactive operation. So there are some limitations. Therefore, research on a more general method, will have important significance for the development of intelligent detec-tion technology for parts detection.Aiming at the problem of current limitations, a non-contact parts size detection method based on CCD is proposed. The method uses CCD camera to collect contour images of parts and a computer for processing; the proposed algorithm is for the general inspection which is adaptive; the method uses the outline as the observation object, which reduces the amount of data, and im-proves the efficiency. This method needs no prior knowledge, and has superior intelligence, uni-versality.In the aspect of image processing, this paper presents a technique of extracting contours. And the contour can be extracted and ensured be smoothed. By using simple edge detection algo-rithm for contour extraction, there will be a lot of pseudo edge and gap phenomenon which is not rational. Contour extraction criterions that are proposed in this paper can guarantee the single pixel extraction. And they can also be applied to other image task.Then this paper proposes an adaptive corner detection technology. Corner detection tech-nology can locate corner position, segment and form curve parameter estimation. The technology uses KD curvature to calculate contour curvature, realizes non maxima suppression, and finally obtains the position coordinates of corners. The algorithm firstly proposes the concept of corner strength. Without prior request such as circle, line, or other type specifically, the method has a certain degree of adaptability.Finally, three experiments are used to test our method. The first experiment is to test the ro-bustness and accuracy; the second experiment by the non-noise parts pictures, mainly to test the performance of this algorithm; third experiment is to test actually ability in practical application with noise parts pictures. And the complexity of the algorithm is analyzed. From results of three groups experiment, we can conclude that our method is designed reasonably and has high accu-racy and short response time, which can meet the test requirements of general parts.