Font Size: a A A

Research And Development Of On-line Inspection System For The Detection Of Painting Product's Surface Defects

Posted on:2016-12-29Degree:MasterType:Thesis
Country:ChinaCandidate:M ChenFull Text:PDF
GTID:2348330479954458Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
With the development of manufacturing technology, people become more demanding on the quality of painting products. The manual inspection method can't meet the detecting requirements any more. This paper studied how to apply the machine vision technology on the detection of painting product's surface defects, and realized the system upon it.In the aspect of general technical scheme, according to the characteristics of products, the general technical scheme was put forward.In the aspect of mechanical structure, a delivery platform structure was designed, and the stiffness and modal analysis of structure were made.In the aspect of imaging acquisition, main technical parameters of camera, lens and frame grabber were analyzed, and the selection methods were discussed. For the design of light, we studied the condition to achieve maximally flat illumination of linear LED array, this condition was verified by simulation. A line dome light and a direct spot light were designed, the effect of distance between LED array and the arc wall of line dome light for the light intensity, and the relationship between the direct spot light's intensity and working distance were studied. Two light were tested, to verify the performance in system applications.In the aspect of detection algorithm, the defect types were divided into two classes, namely flaw defects and shape defects. Efficient detection algorithm for each type of detects was designed. For detect of flaw defects, a detection algorithm based on the self-subtraction after bottom hat operation was designed. For detect of shape defects, we designed a registration method using subspace for the search of registration points, and then using image subtraction operation to achieve defects image. A search algorithm based on line-scanned for fast searching and locating defects was designed.Finally, this article discussed the realization and operation of the system in the industrial field. And by comparing with manual inspection and machine inspection, it has been proved that the inspection system is able to achieve detection accuracy by manual method, and the inspect speed and reliability of inspection system are much better than the manual method. After the long time running of the inspection system in the industrial field, the inspection system was proved to have the advantages of high precision, high speed and flexibility for different kinds of products.
Keywords/Search Tags:machine vision, surface defects detection, mechanical design, light design, image registration, defect extraction
PDF Full Text Request
Related items