基于计算机图形学的声呐图像生成方法  

Sonar Image Generation Method Based on Computer Graphics

作　　者：胡宇翔 张武[1] 李宝奇 HU Yuxiang;ZHANG Wu;LI Baoqi(Key Laboratory of Science and Technology on Advanced Underwater Acoustic Signal Processing,Institute of Acoustics,Chinese Academy of Sciences,Beijing,100190,China;University of Chinese Academy of Sciences,Beijing,100049,China)

机构地区：[1]中国科学院声学研究所中国科学院先进水下信息技术重点实验室,北京100190 [2]中国科学院大学,北京100049

出　　处：《网络新媒体技术》2023年第3期34-42,共9页Network New Media Technology

摘　　要：声呐图像由于采集困难,耗费的人力物力成本高,导致目前几乎没有公开的声呐图像数据集,其严重限制了声呐图像相关研究的进展。为了解决此问题,本文提出了一种基于计算机图形学的声呐图像生成方法,首先,通过场景建模、渲染获取海底地物场景的散射信息和深度信息,其次,通过回波生成计算模拟的原始回波数据,最后利用合成孔径成像算法对其进行处理得到生成的声呐图像。此外,考虑了海洋中常见的声压衰减和散斑噪声现象,并介绍了相关水声信息嵌入方法,可以进一步提高生成声呐图像的真实性。同时,可以通过控制拖体高度、目标高度、目标形状等参数来实现声呐图像的可控生成,能够有效缓解声呐图像数据集短缺的问题。Due to the difficulty of collecting as well as expensive cost of manpower and time,there is nearly no public sonar image datasets which critically limits the development of relevant research.To alleviate this problem,a sonar image generation method based on computer graphics is proposed.Firstly,acquiring the backscatter information and depth information by scene modeling and rendering.Secondly,calculating the raw echo data through echo generation.Finally,processing it by utilizing synthetic aperture imaging algorithm to get generated sonar images.Moreover,some classic underwater phenomenon have been considered,such as sound pressure attenuation and speckle noise.We introduce the acoustic information embedding methods which are aimed to further enhance the realism of results.Meanwhile,realistic sonar images can be controllably generated by changing the height of towfish,the height of object,the shape of object and other parameters.In conclusion,the proposed method can generate realistic sonar images rapidly and alleviate the lack of sonar image datasets.

关 键 词：声呐图像生成 计算机图形学 声压衰减 散斑噪声 可控生成 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] U666.7[自动化与计算机技术—计算机科学与技术]