基于改进支持向量机的无人值守变电站遥视门禁人脸识别方法

Modified-SVM-based Face Recognition for Remote Access Control of Unattended Substations

当前的无人值守变电站遥视门禁人脸识别节点设定一般为分布式形式,识别范围较小,导致门禁识别状态延时增加,为此提出对基于改进支持向量机的无人值守变电站遥视门禁人脸识别方法的设计与验证分析。根据实际识别的需求及标准,对识别图像进行预处理,采用多层级的方式,扩大实际的识别范围,设定多层级识别节点,构建改进支持向量机门禁人脸识别模型,采用级联Adaboost处理实现人脸识别。最终的测试结果表明:对比于传统Face++嵌入式遥视门禁人脸识别测试组、传统大数据遥视门禁人脸识别测试组,本次所设计的改进支持向量机遥视门禁人脸识别测试组最终得出的门禁识别状态延时被较好地控制在15 ms以下,说明该种识别方法的针对性较强,识别的速度及效率较高,具有实际的应用价值。

Current remote access control of unattended substations generally adopts distributed arrangement for face recognition nodes,resulting in small recognition range and increased recognition status delay.Therefore this paper proposes the design and validation analysis of a modified-SVM-based face recognition method.According to actual recognition requirements and standards,the recognition image is preprocessed using a multi-level approach to expand the actual recognition range,set multi-level recognition nodes,construct an modified support vector machine access control face recognition model,and use cascaded Adaboost processing to achieve face recognition.The final test results show that compared to the traditional Face++embedded face recognition group and the traditional big data face recognition group,the modified-SVM-based face recognition group can achieve an access recognition status delay below 15 ms,indicating strong pertinence,high recognition speed and efficiency of the proposed method.