Research Of Talcum Mineral Feature Extraction And Classification Based On Image Processing

Talcum is one of important and significant minerals with high economy value and applied in many national economy areas, such as papermaking, medical, dope and many other industrial fields. Pure Talcum is rare in nature and generally contains various impurities in different grades and levels, like chlorite, magnesite, tremolite, dolomite, quartz, pyrite and so on. In order to achieve the goal of different grades and levels Talcums' applications, the important key is to classify Talcum minerals, so that it can fully make economic value. For research purpose, this dissertation uses OpenCV, wavelet analysis theory, image processing theory and methods for classifying and analyzing Talcum minerals, the major research work has been done as follow:To begin with, Talcum mineral's appearance characteristics were research, including color, gray and texture, and classifying Talcum mineral by ordering Talcum purity. As Classification is required, Talcum mineral is sorted in three different kinds which are including gray, color and texture.Secondly, Talcum mineral images are preprocessed with using the methods include Gaussian filter, image intensification, Gaussian pyramid decomposition sampling, and further Canny operator and threshold segmentation integrate morphology theory are applied in to the project for removing the background of Talcum mineral.Thirdly, through analysis Talcum mineral images, containing color, gray and texture, RGB color space to HSV color space is proposed of the process. According to calculate the image gray histogram feature values of Talcum mineral, as the result it is the first classification' gist which could isolate mass-tone white Talcum mineral. To calculate H figure in HSV color space as the second time classification. Finally, in order to figure out and make the textural features of Talcum mineral clearly in image and digital platform, this dissertation is offering a way to use wavelet analysis and Mallat fast decomposition algorithm to decompose longitude direction and latitude direction extract image texture features. And completed the related Talcum mineral classification experiment.Finally, on the basis of theoretical analysis and the simulation results has been discussed and proven, Talcum mineral automatic classification software was designed and developed with processing Via Visual Studio 2008 development platform and OpenCV library. Via Visual Studio 2008 applies image preprocessing, function of color and gray classifications, wavelet decomposition, the eigenvalue extraction and setting threshold classification module, etc. Theoretical analysis and Talcum mineral image classification experiment's outcome and result prove the correctness and effectiveness of this mentioned method.