应用改进最小距离分类法识别泵功图工况

Applying the Improved Minimum Distance Classification to Identification the Operation Conditions of Pump Diagrams

最小距离分类法广泛应用于文字，图像识别领域。泵功图是封闭的二值曲线图像，可用于判别抽油井各种工况。对最小距离分类法进行改进，并把获得的泵功图进行形态学处理，使之优化为标准功图，然后应用最小距离分类法对油井现场获得的61幅泵功图进行工况分类。实验结果表明，这种方法对泵功图分类是可行的，正确率较高。

The minimum distance classification has been broadly applied to the areas of character recognition and image recognition. The pump diagram is a binary image of closed curve, and it can be used to identify all kinds of operation conditions of an oil well. In this paper, the minimmn distance classification has been improved, and a!l pump diagram have been dealt with morphology method to an optimized standard pump diagram, then it has been applied to identify the operating conditions of 61 pump diagrams which are collected on-site in the oil field. Test results show that the method is feasible and it has high correctness in condition of pump diagrams classification.