互感器为电力系统中不可或缺的重要设备之一。然而由于采集到互感器红外图像受到不同的强噪声以及设备特性等因素影响，可能导致互感器无法正常识别，因此，互感器图像的清晰程度尤为重要。采用不同算法对图像进行灰度化对比试验，根据图像效果，选择加权平均法对互感器红外图像进行灰度化;针对红外图像强噪声、模糊等问题，通过采用均值、高斯、中值以及双边滤波等不同算法对互感器红外图像进行去噪对比实验，比选出峰值信噪比(peak signal to noise ratio，PSNR)值较高的双边滤波作为互感器去噪处理算法，在此基础上，为提高图像对比度，提出了一种改进的高斯-拉普拉斯金字塔图像增强算法，并通过PyCharm实验平台进行直方图、对比度(Contrast)评价指标的对照实验。实验结果表明:电流互感器Contrast值为100。688 6，比原图像增强算法Contrast值提高了19。908 7;电压互感器Contrast值为86。150 1，比原图像增强算法Contrast值提高了21。808 8，验证了该方法的有效性。
An improved gaussian laplace pyramid transformer infrared image enhancement algorithm
In order to meet the needs of national economic development,the scale and voltage level of China's power grid continue to increase,which poses more severe challenges to the safety and reliability of the power system.As a key device responsible for measurement and protection work in the power system,transformers have become one of the i ndispensable and important equipment in the power system;However,due to the influence of different strong noise and equipment characteristics on the collected infrared images of the transformer,it may be difficult to recognize the transformer properly.Therefore,the clarity of the transformer image is particularly important.Firstly,different algorithms are used to perform grayscale comparison experiments on the images.Based on the image effect,the weighted average method is selected to grayscale the infrared image of the transformer;Secondly,in response to the issues of strong noise and blurring in infrared images,different algorithms such as mean,Gaussian,median,and bilateral filtering were used for denoising comparison experiments on the infrared images of transformers.Bilateral filtering with higher PSNR values was selected as the denoising algorithm for transformers.Based on this,an improved Gaussian Laplace pyramid image enhancement algorithm was proposed to improve image contrast.And through the PyCharm experimental platform,a comparison experiment was conducted on histogram and Contrast evaluation indicators.The current transformer Contrast value was 100.6886,which was 19.9087 higher than the original image enhancement algorithm Contrast value,and the voltage transformer Contrast value was 86.1501,which was 21.8088 higher than the original image enhancement algorithm Contrast value,verifying the effectiveness of this method.
国家科技期刊平台
NSTL
万方数据
