基于优化去雾算法的配网开关状态视频识别技术研究  

Research on Video Recognition Technology of Distribution Switch Based on Optimized Defogging Algorithm

作　　者：王丹[1] 李建岐 廖斌[1] 

机构地区：[1]华北电力大学电气与电子工程学院,北京102206 [2]全球能源互联网研究院,北京102209

出　　处：《电力信息与通信技术》2017年第10期31-37,共7页Electric Power Information and Communication Technology

基　　金：国家重点研发计划资助项目(2016YFB0101902)

摘　　要：为解决雾霾天气下室外电力设备人工巡检的不便,以经典暗通道先验去雾算法为基础,针对传统算法复杂度高、天空区域处理效果不理想等问题,提出了一种优化的去雾算法。该算法通过灰度开运算来自适应获取大气光强,采用下采样和插值算法降低算法复杂度,并引入形态学算法和容差机制优化透射率。最后,以柱上跌落式开关为例,采用尺度不变特征变换匹配算法(Scale-Invariant Feature Transform,SIFT)进行开关定位,利用Hough变换进行开关状态识别。仿真和实验结果表明,该方法可有效降低传统算法的复杂度,处理速度快且效果理想,可有效实现雾霾天气下视频图像中开关的状态识别。In order to solve the inconvenience of manual inspection of outdoor power equipment under fog and haze, based on the classical dark channel prior defogging algorithm, this paper proposed an improved algorithm to solve the problem of the traditional algorithm＇s complexity and the difficulty for much sky area processing, The algorithm used gray-opening to acquire the intensity of the atmosphere adaptively, and used lower sampling and interpolation algorithm to reduce the complexity of the algorithm, then introduced morphological algorithm and tolerance mechanism to optimize the transmittance. Finally, taking the dropping switch as an example, the SIFT algorithm is used for switch location, and the Hough transform is used to identify the switching states. Simulations and experiments showed that this method effectively reduced the complexity of the traditional algorithm, processing speed and effect were ideal, which realized the effective recognition of switching state in video images under fog and haze.

关 键 词：配网开关 暗通道先验 形态学 状态识别 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]