Design and implementation of insulators material hydrophobicity measure system by support vector machine decision tree learning

Hydrophobicity is an important parameter to measure electronic properties of insulated material. How to decide the hydrophobic level of insulated material surface conveniently, quickly and accurately, is a problem needing to be solved urgently. IMHMS (insulator material hydrophobicity measure system) is a system designed to solve it using misjudging-cost based support vector machine decision tree learning and predicting. In IMHMS, support vector machine decision tree (SVMDT) is learned from training samples dataset including plenty of spraying images of insulated material´s surface with different hydrophobic levels by a novel learning algorithm, and is used to predict hydrophobic level of new sample. Information of samples includes attributions of spray image of insulated material´s surface which are extracted by digital image processing methods, and hydrophobic levels are given by field experts. The result of testing shows hydrophobic level of insulated material´s surface outputted by IMHMS can satisfy the precision requirement of practicality application.