Extracting the features of drilling foams with image processing techniques and data analysis using artificial neural networks

ABSTRACTEXTRACTING THE FEATURES OF DRILLING FOAMS WITHIMAGE PROCESSING TECHNIQUES AND DATA ANALYSISUSING ARTIFICIAL NEURAL NETWORKSVeli Mert ALTAThe chemical foams have a significant role in industrial processes, whichuse liquid carrying such as flotation and drilling, because of their lowdensity and high viscosity compared to the liquid that they are in. Up to dateresearch showed that, the features of foam like shape and dimension have adirect effect on the carrying capacity. This study aimed to develop a modelthat perceive the features of foams with image processing techniques andmake a data analysis using artifical neural networks. The methods like localthresholding and filtering from fourier transformation are applied and theirresults are shown according to this purpose. As to this model, a developedalgorithm checks if the pixels on the image are a possible center of apotential froth or not. The techniques like tracing angular minimum, findingcenter of weight and contour tracing are applied to obtained potential froths.Radius, area, perimeter and other shape and dimension properties of thefroths, whose borders are designated, are derived and combined withchemical data to perform a data analysis using artificial neural network.Thealgorithmic benefits of techniques used in the method, over time are shown.It is observed that, the developed image processing model is moresuccessful than the techniques widely used in literature, especially withnoisy froth images. The results of data analysis performed by artifical neuralnetwork are studied. Results show that it is possible to perform froth dataanalysis by neural networks using the extracted froth data by imageprocessing techniques.Keywords: Image Processing, Analysis of Froth Images, Analysis ofBubble Images, Artificial Neural Network, Pattern Recognition