基于机器学习的电力调度机房静态健康度超分辨率图像识别方法  

Super-resolution Image Recognition Method for Static Health of Power Dispatching Room Based on Machine Learning

作　　者：安天瑜 王铎钦 王海宽 AN Tianyu;WANG Duoqin;WANG Haikuan(Northeast Branch of State Grid Corporation of China,Shenyang 110180,China)

机构地区：[1]国家电网有限公司东北分部,辽宁沈阳110180

出　　处：《微型电脑应用》2024年第4期157-161,共5页Microcomputer Applications

摘　　要：针对电力调度机房运行态势不一,终端信号灯色彩难以识别的问题,提出了基于机器学习的电力调度机房静态健康度超分辨率图像识别方法。设定残差阈值,采用模糊最大熵方法,计算电力调度机房静态图像目标类和背景类的最佳分离点,引入模糊隶属度函数,运用RGB极大比值法,提取与增强超分辨率图像信号灯的色彩特征,构造最佳分类面,设置分类约束条件,依据二次分类器函数,识别电力调度机房静态健康度。实验结果表明,该方法能够提高电力调度机房静态图像质量,色彩特征识别效果较佳,确保电力调度机房静态健康度识别准确性。Aiming at the problem that the operation situation of the power dispatching equipment room is different and the color of the terminal signal light is difficult to identify,a super-resolution image recognition method of the static health degree of the power dispatching equipment room based on machine learning is proposed.The residual threshold is set,and the fuzzy maximum entropy method is used to calculate the optimal separation point between the target class and the background class of the static image of the power dispatching room.The fuzzy membership function is introduced,and the RGB maximum ratio method is used to extract and enhance the super-resolution image signal light.The optimal classification surface is constructed,the classification constraints are set,and the static health degree of the power dispatching room is identified according to the secondary classifier function.The experimental results show that the method can improve the static image quality of the power dispatching room,and the color feature recognition effect is better,which ensures the accuracy of the static health recognition of the power dispatching room.

关 键 词：机器学习 支持向量机 电力调度机房 超分辨率图像 

分 类 号：TP725[自动化与计算机技术—检测技术与自动化装置]