煤矿变电所数显仪表字符识别研究

Study of character recognition of digital display instrument in coal mine substation

针对煤矿变电所数显式仪表存在巡检自动化程度低和识别可靠性差等问题,提出了一种基于改进遗传算法和支持向量机算法的字符识别算法。该算法采用Harr-Like特征作为字符识别特征,通过改进的遗传算法对分类器支持向量机的参数进行寻优,利用主元分析法进行降维处理,并使用支持向量机识别数显式仪表字符。实验验证了该算法的有效性和可行性。

In view of problems of low automation degree and recognition reliability of digital display instrument in coal mine substation,a kind of character recognition algorithm based on improved genetic algorithm and support vector machine algorithm was proposed.The algorithm adopts Harr-Like features as character recognition features,improved genetic algorithm was chosen to search the optimal parametersof the support vector machine classifier,and uses principal component analysis method to conduct the dimension reducing process,then applies support vector machine to identify character of the digital display instrument.The effectiveness and feasibility of the algorithm was validated by experiments.