论文部分内容阅读
为了早期发现变电站事故隐患,建立了一个红外温度巡检与预警系统,该系统能够自动循环采集变电站电气设备的红外图像,通过提取目标设备的温度信息进行故障判别。为了解决机械传动造成的图像位置偏差导致系统难以对感兴趣设备进行定位监测的问题,提出了一种基于对象分割和图像配准的校正方法,首先对图像进行对象分割和组态定义,并将待测目标提取出来,然后采用基于相位相关和Harris角点匹配的图像配准方法对序列偏差图像进行配准。使用上述方法对某变电站中获取的15组电气设备红外遥视图像进行了实验,平均正确匹配率达到93.3%,并且配准精度较高,达到了亚像素级。实验表明该方法能够对偏差图像进行有效校正,使系统能够及时准确地获得设备部件的温度,保障监测系统的可靠运行。
In order to detect potential accidents in substations early, an infrared temperature monitoring and early warning system was set up. The system can automatically collect the infrared images of substation electrical equipment and identify the fault by extracting the temperature information of the target equipment. In order to solve the problem that the system is difficult to locate and monitor the interested equipment in order to solve the problem of the image position deviation caused by the mechanical transmission, a method based on object segmentation and image registration is proposed. Firstly, the image is segmented and the configuration is defined, The target to be detected is extracted, and then the image of sequence deviation is registered with image registration based on phase correlation and Harris corner matching. The above method was used to test the infrared telemetry images of 15 sets of electrical equipment obtained in a substation. The average correct matching rate was 93.3% and the registration accuracy was high, reaching sub-pixel level. Experiments show that this method can effectively correct the deviation image, make the system obtain the temperature of the equipment components timely and accurately, and guarantee the reliable operation of the monitoring system.