一种用于超景深融合的快速多曝光融合

A Fast Multi-Exposure Fusion Algorithm for Ultra Depth of Field Fusion

摘要: 超景深融合在刀具，电路板，器件外壳和部分小型零件等测量领域的应用越来越广泛，但上述测量目标大多是由多种材料组成，对光线有着不同程度和类型的反射，使得成像点的强度超过了相机的感光范围，导致过曝和过暗的低质量成像点出现。为了解决现有的高动态成像融合算法(HDR)在超景深融合中融合速度缓慢的问题，采用了聚焦评价和灰度评价的方式，获取高细节信息和剔除低细节信息。在限制图像动态范围不失真的同时，且利用聚焦算子的评价速度要高于HDR的优势，达到了高速融合多曝光图片的目的。实验结果表明，在融合速度上，本文算法远快于现有的HDR算法，达到了0.1463 s，且融合质量与主流HDR算法的图像质量相当。

Abstract: Ultra-depth of field fusion is increasingly widely used in measurement fields such as cutting tools, circuit boards, device shells and some small parts. However, most of the above measurement targets are composed of various materials, which have different degrees and types of light reflection, making the intensity of the imaging points exceed the photographic range of the camera, resulting in the appearance of low quality imaging points with over exposure and too dark. In order to solve the problem of the slow fusion speed of the existing high dynamic imaging fusion algorithm (HDR) in the ultra-depth of field fusion, the focus evaluation and gray evaluation methods are used to obtain the high detail information and eliminate the low detail information. In addition to limiting the dynamic range of the image without losing the truth, and using the focusing operator’s advantage of higher evaluation speed than HDR, the purpose of high-speed fusion of multi-exposure images is achieved. Experimental results show that the fusion speed of the proposed algorithm is much faster than that of the existing HDR algorithm, reaching 0.1463 s, and the fusion quality is comparable to that of the mainstream HDR algorithm.