智能配电房危险情况的图像特征识别方法

摘要

对智能配电房无人化管理是智能电网的重要发展方向.提出BRISK算法和增加平移因子的方法相结合的智能配电房危险情况图像识别方法,对智能配电房的实时图像进行白化处理,去除白斑对配电房内图像特征的影响,采用投影分析法,求出智能配电房实时情况图像的Zernike矩.以此为基础,采用小波多尺度分析法,获取智能配电房图像的频域信息,采用BRISK算法和增加平移因子的方法相结合,对图像中的危险情况进行识别.仿真实验结果表明,该方法能智能化的对相关特征进行识别,识别时间、识别效率及识别精度均优于一般的监控系统.%The unmanned management for the intelligent substations is an important direction of smart grid.The image feature recognition method combined BRISK algorithm and increasing the translation factor method was proposed for dangerous situation in the intelligent substation,the real-time image in the intelligent substation was bleached,the effect of the white spot on the image characteristics in the substation was removed,and the Zernike moment of the real-time image in the intelligent substation was gotten by using the projection method.On this basis,the image frequency domain information in the intelligent substation was obtained by using the wavelet multi-scale analysis method;the dangerous situation in image was recognized by using BRISK algorithm and method of increasing the translation factor.The simulation experimental results show that the method can recognize the related characteristics,and its recognition time,efficiency and precision are better than general monitoring system.