基于直方图的红外图像细节增强算法研究  

Infrared Image Detail Enhancement Based on Histogram

作　　者：韦瑞峰[1] 赵荣普[1] 徐肖庆[1] 赵灿辉[1] 赵旋 

机构地区：[1]云南电网公司昆明供电局,云南昆明650200 [2]广州安电测控技术有限公司,广东广州510630

出　　处：《红外技术》2016年第6期472-475,共4页Infrared Technology

摘　　要：红外热成像技术广泛应用于故障检测中,但由于红外探测器各单元响应的非均匀性,使得红外图像对比度低、细节和边缘模糊,不利于人工观测和机器识别,因此对红外图像细节增强算法的研究具有深刻的意义。分析了传统直方图增强算法中的缺陷,提出了一种抑制背景的局部直方图均衡算法,并使用遗传算法加快了阈值求取的速度。实验结果表明,该方法可以在抑制背景的同时增强红外图像的细节,同时速度较快,可用于实时红外图像的预处理。Infrared thermal imaging technology is widely used in fault detection, but as a result of infrared detectors＇ heterogeneity response, the contrast of infrared image is low and is not conductive for human observation and machine recognition, so the research on the infrared image detail enhancement has profound significance. This paper first analyzes the defects of traditional histogram enhancement algorithm, and then proposes a local histogram equalization algorithm with background suppression, and uses genetic algorithm to speed up the threshold calculation. Experimental results show that the proposed method can suppress background while enhancing infrared images＇ detail and running fast, and can be used for real-time infrared image processing.

关 键 词：红外图像 直方图均衡 细节增强 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]