光学 精密工程, 2016, 24 (5): 1159, 网络出版: 2016-06-15   

基于光纹特征的激光主动照明图像去模糊

Image deblurring for laser active illumination based on light vein features
作者单位
1 中国科学院 长春光学精密机械与物理研究所 激光与物质相互作用国家重点实验室, 吉林 长春 130033
2 北京航天自动控制研究所 宇航智能控制国家重点实验室, 北京 100039
摘要
受光学系统离焦、大气扰动、平台振动的影响, 激光主动照明系统捕获的图像容易被模糊, 而传统的去模糊方法难以取得良好的复原效果, 故本文提出基于光纹特征的盲解卷积复原方法来实现图像去模糊。首先将模糊图像降采样, 建立尺度金字塔, 在尺度空间查找光纹特征图像块。随后基于激光主动照明图像饱和像素较多的特点, 提出新的图像退化模型。最后针对模糊核估计、光纹参数更新、清晰图像复原3个步骤, 提出适用的能量函数, 迭代复原出无噪清晰图像。搭建了主动照明系统, 在捕获的激光主动照明图像上进行了实验, 并与现有方法进行了对比。结果表明: 本文方法不仅能够复原出清晰图像, 而且能有效抑制振铃效应, 其客观评价指标峰值信噪比(PSNR)优于已有的其他算法。
Abstract
The images captured by a laser active illumination system are easily blurred by optical system defocus, atmospheric disturbance and platform vibration, and traditional deblurring methods can not achieve good image restoration. Therefore, this paper proposes a blind deconvolution restoration method based on light vein features to implement the image deblurring. Firstly, the blurred images were down-sampled, a scale pyramid was established, and the light vein patches were searched along a scale space. Then, according to the laser illumination imaging characterized by more saturated pixels, a new nonlinear image degradation model was proposed. Finally, to implement the three steps, blurred kernel estimation, light vein patch renewal and the image restoration, the proper energy function was proposed and the latent image without noise was iterated and recovered. A laser active illumination system was established and the experiments were performed on the images captured by the system and compared with state-of-art deblurred methods, The experiment results indicate that the proposed method not only obtains more clear images, but also availably suppresses ringing artifacts, and its Peak Signal to Noise Ratio (PSNR) is superior to that of other algorithms.

王灿进, 石宁宁, 孙涛, 王锐. 基于光纹特征的激光主动照明图像去模糊[J]. 光学 精密工程, 2016, 24(5): 1159. WANG Can-Jin, SHI Ning-ning, SUN Tao, WANG Rui. Image deblurring for laser active illumination based on light vein features[J]. Optics and Precision Engineering, 2016, 24(5): 1159.

本文已被 3 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!