雾气环境下基于引导滤波的中压开关图像检测  

Image Detection of Medium-voltage Switches Based on Guided Filtering in Fog Environment

作　　者：康璐瑶 邵振华 吴保露 KANG Luyao;SHAO Zhenhua;WU Baolu

机构地区：[1]福建农林大学,福州350000 [2]闽江学院,福州350000

出　　处：《科技创新与应用》2025年第4期51-55,共5页Technology Innovation and Application

基　　金：国家自然科学基金项目(51277091);福建省自然科学基金工业引导性项目(2021H0040,2021H00055)。

摘　　要：针对复杂条件下获取的图像质量不佳、中压开关标识图识别率低下等问题,提出一种在雾气环境下对中压开关图像进行智能检测与识别的算法。在雾气环境下拍摄中压开关图像,用快速引导滤波优化透射率的颜色衰减先验法去雾。将结果与一系列经典算法的客观评价指标对比,表明此算法可以在雾气环境下更好地恢复图像特征,同时提升运行速率。将去雾后的图像在MATLAB中进行基于HSV颜色空间的分割与识别仿真,即可实现中压开关标识的准确定位。Aiming at the problems of poor image quality obtained under complex conditions and low recognition rate of MV switch identification maps,an algorithm for intelligent detection and recognition of MV switch images in fog environment is proposed.Images of medium-voltage switches are taken in a foggy environment,and the color attenuation prior method that optimizes transmittance is used to defog.The results are compared with objective evaluation indicators of a series of classic algorithms,which shows that this algorithm can better restore image characteristics in a fog environment while improving the running speed.The defogged image is subjected to segmentation and recognition simulation based on HSV color space in MATLAB,and accurate positioning of medium-voltage switch logos can be achieved.

关 键 词：图像增强 雾 引导滤波 颜色分割 中压开关 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]