The Study Of Digital Imaging Detection System On The Analysis Of Metallurgical Melts’ Surface Properties

The surface properties and the interfacial properties play a leading role in the reactionand separation of melts and are also the foundation of melts’ interface dynamics. The surfacetension is one of the important parameters and plays the great significance in themeasurement of surface tension. Metallurgical melts are in molten state at high temperature,and the state can’t be seen by the naked eyes. It is necessary to use X ray to obtain the imagesof this state. We can not only observe the state of the melts through the images, but alsoachieve the surface properties by analyzing the images, such as surface tension.Firstly, the paper introduces the source and the meaning of the project, as well as theresearch status of surface properties and digital imaging detection technology. Thencombined with the study of the tutor and the sensors in my laboratory, according to theconditions and the requirements of metallurgical melt test, the paper designed the CCDcamera and the data acquisition system. And according to the requirements of the surfacetension measurement method-the drop shape method, the image preprocessing, imagesegmentation and contour extraction are studied to get the parameters of the surfaceproperties of melts.In the image preprocessing module, firstly, the paper analyses the noise source anddetermines the model. Then the traditional noise reduction methods are adopted to reduce thenoise, such as the median filter, adaptive median filter and multi-frame superposition averagefilter. After that, the enhancement methods including airspace enhancement act and frequencydomain enhancement act are adopted to enhance the images, in order to improve the clarity and contrast of the images. In the image segmentation module, the threshold segmentationmethods and the edge detection methods are studied. The threshold segmentation methodsmainly introduce the OTSU method and iterative method. The edge detection methodintroduces several traditional algorithms and mainly introduces the SUSAN edge detectionalgorithm and the algorithm based on the mathematical morphology. After processing theimages, the results are compared with each other. In the contour extraction module, in orderto obtain a complete and smooth edge curve, the curve fitting methods based on least squareare adopted. Finally we can get the surface properties of the melts, such as surface tensionand the contact angle and so on.All the algorithm functions mentioned above are realized in the platform of VC++, andthe digital imaging detection system for dealing with the analysis of metallurgical melts’surface properties is developed.At last part, the summary and outlook are made, and the deficiencies and the placesneeded to be improved are pointed out. The experiment shows that the paper has alreadyachieved the requirement of the subject.