一种用于紫外成像系统的PCNN图像融合算法

An PCNN Image Fusion Algorithm Applied in UV Imaging System

为了解决高压电气设备局部放电故障点定位问题,把PCNN图像融合算法应用到紫外成像系统中,根据高压电气设备局部放电的同时向四周发射紫外光和紫外光波段在日盲200~400 nm的原理,在设备局放故障区域分别采集紫外光图像和可见光图像并进行图像融合,在新生成的融合图像中精确定位局部放电故障点。研究针对融合源图像自身系数特点会影响PCNN神经网络连接强度系数的特点,使用粒子群优化算法对PCNN神经网络中的连接强度系数进行全局寻优,使其可以针对不同融合源图像的各自系数特征自适应寻找最优连接强度系数。研究结果表明,优化后的PCNN算法对比于其他算法所得图像信息更加丰富、定位更加精准,融合图像可以有效地定位高压设备电晕放电故障点。

In order to solve the problem of fault location of the partial discharge for high voltage electrical equipment,the PCNN image fusion algorithm is applied to the ultraviolet imaging system,and hence the ultraviolet discharge and visible images of the discharge area can be acquired and integrated according to the fact that ultraviolet wavelength varies in sun-blind band,about 200 nm to 400 nm. Then the accurate discharge point could be found in the newly generated fusion image. In view of the influence of the coefficient of the source image fusion on the connection strength coefficient of PCNN neural network,the particle swarm optimization algorithm is used to optimize PCNN neural network connection strength coefficient in global,making the connection strength coefficient capable of self-adapting for different source images. Experimental results show that compared with other ultraviolet imager's algorithms,the image obtained using the optimized PCNN algorithm is remarkably richer in information,and more accurate in fault location,and the fusion image can effectively locate the high voltage corona discharge point of the equipment.