不同波长光源照明的水下成像及光学图像实时处理  

Underwater Imaging and Real-Time Optical Image Processing under Illumination by Light Sources with Different Wavelengths

作　　者：聂瑛[1] 何志毅[1] 

机构地区：[1]桂林电子科技大学信息与通信学院,广西桂林541004

出　　处：《光学学报》2014年第7期59-65,共7页Acta Optica Sinica

基　　金：国家自然科学基金(61077036);无线宽带通信与信号处理重点实验室开放基金(12108)

摘　　要：水下成像环境复杂,为改进水下成像质量,在清水中掺入一定量的蓝绿墨水和牛奶来模拟不同的水质环境,采用不同波长的照明光源进行水下图像采集实验。考虑到水下的吸收和散射的影响,将不同光源采集到的图像特点进行对比分析,设计了一种新的光学方法来进行图像处理。结合长波长的红光抗散射和较短波长的蓝绿光抗吸收两方面的光传输特性,通过红光和绿光照明采集图像的灰度线性差值去除散射噪声信息,获取目标图像信息,在Windows和Android系统中进行图像采集和清晰化。实验结果表明该方法能有效增强图像清晰度,与虚光蒙板(USM)锐化算法相比计算量小,实时性高。To improve the different water environments underwater imaging quality under the complex underwater imaging environment, are simulated by adding the blue-green ink and milk for the absorption and scattering effects respectively, and light sources with different wavelengths are employed for illumination. To eliminate or alleviate the effect of absorption and scattering, an optical image processing technique is proposed after analyzing the collected characteristics of images with different light sources. The experiments show that the blue-green light has better transmission to overcome the water absorption and the red light with longer wavelength exhibits clearer illumination to alleviate the scattering influence and obtain target image information. The scattering noise can be removed with the collected differential image grey-scale value of the red light and blue-green light. The image processing technique in both Windows and Android systems is demonstrated, achieving a high performance. Compared with the unsharp mask （USM） sharpening algorithms, it has the advantages of simpler computation and good realtime performance.

关 键 词：图像处理 水下成像 照明光源 散射 吸收 

分 类 号：TN29[电子电信—物理电子学]