航空影像雾霾滤除技术研究

发布时间：2019-01-03 06:47

【摘要】：航空带雾影像具有视觉模糊、对比度低、图像信息缺失等缺点,大大降低了图像的应用价值及内业成图精度。随着计算机视觉技术的不断提高,雾天图像清晰化处理可以实现从信息量低的有雾图像到信息丰富无雾图像的恢复与还原。因此,如何改善和提高由雾霾造成的退化图像质量已经成为航空摄影测量领域研究的新热点,本文以探讨实用化去除航空影像雾霾的软硬件系统为目的进行了以下研究:1)暗通道先验的改进算法在软件算法方面,主要介绍了以直方图均衡化、Retinex理论为代表的图像增强去雾方法和基于图像复原的暗通道先验去雾算法,从理论基础、处理效率、色彩保真性等方面对比分析了这几种方法的主要特点及不足,从而选择了更为严谨有效的基于图像退化模型的暗通道先验算法进行本文航空影像的雾霾滤除技术研究。针对航空带雾影像特点,总结分析了暗通道先验算法在图像去雾应用中的不足,通过引入自动白平衡算法,使用交替顺序滤波器替代原算法软抠图操作以细化初始透射率,提出了一种适合航空影像的暗通道先验改进去雾算法,实验结果表明:本文优化方法具有更优去雾效果,图像色彩保真度高,且处理效率高。2)基于红外影像和可见光影像融合的去雾方法在硬件方面,通过红外光滤波镜、加固技术对哈苏单机进行改造处理,研发集成了一套能同时获取可见光与红外光影像的双头相机,基于红外线比可见光具有更强穿透性,提出了一种双影像融合去雾方法,实验验证了该去雾方法具有高效的可行性。3)两种方法的程序实现通过Matlab开发软件平台,使用图像用户界面(Graphical User Interfaces,GUI)工具设计了暗通道先验改进算法与融合去雾方法的简单程序系统。
[Abstract]:Aeronautical fogged image has many disadvantages such as blurred vision, low contrast and missing image information, which greatly reduces the application value of image and the precision of internal mapping. With the continuous improvement of computer vision technology, the clear processing of fog image can realize the restoration and restoration from the fog image with low information content to the rich fog image. Therefore, how to improve the quality of degraded images caused by haze has become a new research hotspot in the field of aerial photogrammetry. The purpose of this paper is to study the software and hardware system for the practical removal of aerial image haze. The following research is done: 1) the improved dark channel priori algorithm in software algorithm mainly introduces histogram equalization. The image enhancement de-fogging method represented by Retinex theory and the dark channel priori de-fogging algorithm based on image restoration are compared and analyzed from the aspects of theoretical basis, processing efficiency, color fidelity and so on. Therefore, a more rigorous and effective dark channel prior algorithm based on image degradation model is selected to study the haze filtering technology of aerial images in this paper. According to the characteristics of aeronautical fogged image, the deficiency of dark channel priori algorithm in image de-fogging is summarized and analyzed. By introducing automatic white balance algorithm, alternating sequence filter is used to replace the original soft matting operation to refine the initial transmittance. An improved dark channel prior de-fogging algorithm for aeronautical images is proposed. The experimental results show that the proposed method has better effect of de-fogging, and the color fidelity of the image is high. And the processing efficiency is high. 2) based on the fusion of infrared image and visible light image, the method of de-fogging is used in the hardware aspect. In the aspect of hardware, the harzu single machine is modified by the infrared filter mirror and the strengthening technology. A dual-head camera which can obtain both visible and infrared images at the same time is developed and integrated. Based on the fact that infrared is more penetrating than visible light, a dual-image fusion de-fogging method is proposed. Experimental results show that the de-fogging method is effective and feasible. 3) the program of the two methods is implemented using the image user interface (Graphical User Interfaces,) through the Matlab development software platform. GUI) tool has designed a simple program system of dark channel priori improved algorithm and fogging method.
【学位授予单位】：河南理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】