红外图像自适应细节增强算法的研究与实现

发布时间：2018-05-03 12:54

本文选题：红外成像 + 细节增强　；参考：《华中科技大学》2016年硕士论文

【摘要】：红外成像技术是一种采集环境中的红外辐射信号并将其以二维图像形式表现出来的信号处理技术。由于红外线具有抗干扰能力强、隐蔽性好、保密性高等优势,红外成像技术在医疗检测、工业探伤以及军事目标侦察等领域有着广泛的应用。但是目前的红外图像普遍存在对比度低的缺点,不仅严重影响其本身的图像质量,还不利于后续的图像目标识别与追踪等,因此必须对其进行增强处理。然而传统的红外图像增强技术尽管可以提升图像的对比度与信噪比,但也会丢失图像中的细节,并且算法的场景适应性差,对红外成像技术的应用产生极大的限制。因此,本文从红外图像的基本特征出发,通过分析现有红外图像增强算法存在的优缺点,提出一种基于引导滤波器的红外图像自适应细节增强算法。该算法总共包含四个步骤:首先采用引导滤波器将原始图像分成包含低频背景的基本层和包含高频细节的细节层;然后采用基于图像统计特性的自适应直方图投影算法处理基本层,以实现图像对比度的改善,并提升算法的场景适应性;同时采用基于用自适应增益校正处理细节层,以增强图像细节并抑制图像噪声;最后利用经过处理的基本层和细节层构建最终的输出图像。多场景下真实红外图像数据的实验结果表明,与现有的同类算法相比,本文算法能有效处理多种场景下的红外图像,并且输出图像的对比度好、细节丰富,具有良好的图像质量。算法的主观和客观评价标准结果均表明其在红外图像动态范围压缩、对比度提升与细节保留方面有良好的处理性能以及出色的场景适应性。
[Abstract]:Infrared imaging technology is a kind of signal processing technology which collects infrared radiation signals in environment and displays them as two-dimensional images. Infrared imaging technology has been widely used in the fields of medical detection, industrial flaw detection and military target reconnaissance because of its advantages of strong anti-interference ability, good concealment, high confidentiality and so on. However, the current infrared images generally have the shortcomings of low contrast, which not only seriously affect their own image quality, but also is not conducive to the subsequent image recognition and tracking, so it must be enhanced processing. However, although the traditional infrared image enhancement technology can improve the contrast and signal-to-noise ratio of the image, it will also lose the details of the image, and the algorithm has poor adaptability to the scene, which greatly limits the application of the infrared imaging technology. Therefore, based on the basic characteristics of infrared image and the advantages and disadvantages of existing infrared image enhancement algorithms, an adaptive detail enhancement algorithm for infrared image based on guided filter is proposed in this paper. The algorithm consists of four steps: firstly, the original image is divided into basic layer with low frequency background and detail layer with high frequency details by using a bootstrap filter; Then the adaptive histogram projection algorithm based on the statistical characteristics of the image is used to process the basic layer to improve the image contrast and enhance the scene adaptability of the algorithm. Finally, the final output image is constructed by using the processed basic layer and the detail layer. The experimental results of real infrared image data in multi-scene show that the proposed algorithm can deal with infrared images in many scenarios effectively, and the output image has good contrast and rich details. Good image quality. The subjective and objective evaluation results show that the algorithm has good processing performance and excellent scene adaptability in infrared image dynamic range compression, contrast enhancement and detail retention.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【参考文献】