Research On Key Technologies Of The On-line Surface Detection System For Marine Steel-Plate

In recent years, China’s shipbuilding industry has entered a rapid development track, and with the expansion of the industry scale demand for ship has sharply rised. In ship manufacture a large number of different color steel plates are needed. However in the manufacturing process of steel continuous casting billets, rolling equipment and processing techniques, multifaceted reasons lead to a variety of different types of defects on steel plate’s surface. The plate quality affects the appearance of the hull surface as well as the performance of corrosion resistance, abrasion resistance and fatigue strength of the ship ultimately. At present, the domestic shipbuilding enterprises have been using artificial visual sampling to detect surface defects without effective and intelligent techniques.According to the demand of the steel plate detection system, this system adopts portal frame structure, the whole measuring system contains 16 separate sets of imaging measuring device and 5 sets industry computer control. when the steel plate with a certain speed motion, image measuring device collects real-time image on the upper and lower steel plate surfaces. Image acquisition system for CCD image processing in real time, on-line identification plate surface defects.This system according to the characteristics of the steel plate operation, adopts the steel plate parallel detection method to process parallel steel plate. The image gradient of processing and gray stretch etc image preprocessing method, the image preprocessing, then use Fisher standards on image threshold defect extraction, and the rapid clustering filter method to threshold after extraction of the image noise filter, and ultimately get the defect treatment results. The defect information is transmitted to the master control machine simulation display.This project developed a software system by using Visual C++and made installation and commissioning on the scene. This article gives the results of experimental treatments. It fits well with the design requirement of the system, effectively detection and location of the steel plate flaw and also meets real-time’s demands.