激光与光电子学进展, 2020, 57 (6): 061102, 网络出版: 2020-03-06
基于离散余弦变换金字塔分解的新型偏振光学去雾技术 下载: 1285次
Novel Polarimetric Dehazing Method Using Discrete Cosine Transform Based Laplacian Pyramid
成像系统 偏振成像 图像增强 散射介质成像 拉普拉斯金字塔 imaging systems polarimetric imaging image enhancement scattering medium imaging Laplacian pyramid
摘要
偏振光学去雾技术具有细节恢复好、颜色还原度高的优点。为了进一步提高偏振光学去雾技术的去雾能力,提出了一种新型偏振光学去雾技术,该技术利用离散余弦变换构建图像金字塔,再构建图像拉普拉斯金字塔,利用传统偏振光学去雾技术对图像拉普拉斯金字塔的每一级进行去雾处理,使用去雾后的拉普拉斯金字塔重建得到去雾后图像。实验结果表明,与传统偏振光学去雾技术相比,该技术可以得到相当或更好的去雾效果,表现出良好的图像去雾能力,对于偏振光学去雾技术的进一步优化具有一定意义。
Abstract
Polarimetric dehazing methods have shown advantages of excellent detail recovery capacity and color restoration. To improve the dehazing capacity of polarimetric dehazing methods, we propose a new polarimetric dehazing method. It utilizes discrete cosine transform to construct the image pyramid and then build the Laplacian pyramid. Each level of the Laplacian pyramid is dehazed by the polarimetric dehazing method, and the dehazed image is reconstructed from the handled Laplacian pyramid. The experiment results show that compared with the traditional polarization optical dehazing method, the proposed method can achieve the comparable or better dehazing images, and shows good dehazing capacity. The proposed method is significant for the further improvement of polarimetric dehazing method.
张文飞, 任立勇, 邢飞, 张芳, 葛筱璐, 王国梅, 付圣贵. 基于离散余弦变换金字塔分解的新型偏振光学去雾技术[J]. 激光与光电子学进展, 2020, 57(6): 061102. Wenfei Zhang, Liyong Ren, Fei Xing, Fang Zhang, Xiaolu Ge, Guomei Wang, Shenggui Fu. Novel Polarimetric Dehazing Method Using Discrete Cosine Transform Based Laplacian Pyramid[J]. Laser & Optoelectronics Progress, 2020, 57(6): 061102.