Feature Abstraction of Time-Sequence Curve and Its Similarity Calculation

Time-sequence data are visually shown in curves. A great deal of information can be derived from them, which lays the foundation for the recognition of curves. Along the x-axis direction and by means of averagely dividing 0deg ~180deg into small intervals, standard codes are adopted to represent the general tendency of the curve. Based on standard codes, corresponding revisal codes are adopted to represent the deviation from them. The combination of the two-level codes describes the curve shape exactly. A similarity function is defined to compute the similarity between two curves, which in most cases are not of the same length. The method described above is proved effective in the recognition of sediment microfacies by well-logging curves from the Luliang Area of Xinjiang Kelamayi Oil Field.