The High-temperature Billet Crack Detection Research Based On Digital Image Processing

Digital Image Processing Technology has been researching by the academics sincethe50s of the last century. Through dozens of the accumulation, it not only becamefrom the research for the theory to practical application, but also succeeded inapplication on medical, optical, communication and so on. Digital Image ProcessingTechnology has most of advantages especially high precision of recognition, theidentification of automation, high efficiency of detection, which makes the academicsbe fascinated to research it along and it does not change. For this age, with rapiddevelopment of the industry, the criterion of Steel manufacturing is more and moreserious, but actually end product often appears of all kinds of defects, such as surfacecracks. These defects less than the machine can not work, at worst, it causes theindustrial accidents, and results in loss of economic. Because the method of detection ofthe billet still drop behind. For example, end product of billet is still detected with eyes,or with the Simple and crude instrument, it seriously constraints the development of thebillet production. If the billet crack could be detected before the billet formed, andRapidly let the defected billet recast, which could mostly improve the efficiency ofproduction. Because the specificity of billet production, the high temperatureenvironment, which results in common Contact Detection Method not coming intoeffect. This paper, contecting the specificity of the high temperature of billet production,using Digital Image Processing Technology and instrument knowledge, seted up asystem of crack detection of the high temperature’s billet, which make the later researchof the detection of the billet crack of automation supplying the strong basis.This paper introduced the design principle of the system of crack detection of thehigh temperature’s billet, and some related elementary knowledge about Digital ImageProcessing Technology, based on the target and environment of the detection, designeda system of crack detection of the high temperature’s billet. Still that, it described theframework of the system, operating principle and the use of the hardware facility,beside those, it still described their performance index. Especially, it analysed appliedrange, effect of filtering algorithm and edge detection, contecting the thinking of C++ Programming, and designed the software of the detect system. It designed the procedureof the detection of crack. Specific steps are as follows. First, it took photos with CCDcamera, and then using Gradation Correction improved contrast of the photo. And usedmean filter to clear noise, then detected the crack edge by the Gauss-Laplace operator,besides these, many handling method could be used. For example, Image Sharpening,Histogram Equalization and Threshold Segmentation. At last, feature information ofcrack could be recognized, such as area, perimeter, length and so on. Basing onidentification rule, it recognized the crack. On the exploitation of the software of system,it used VC6.0to exploit the detect instrument, which has simple interface and strongfunction.This paper not only summaried the research results in the research of crackdetection, but also prospected the future work of billet crack detection. It provided theguidance for the future research of crack detection.