摘要
为解决复杂海洋环境导致的侧扫声纳图像亮度不均、对比度低等问题,基于侧扫声纳的工作原理,对声纳图像在外业实时控制和内业数据处理过程中常用的增益补偿方法进行了深入研究,提出了一种侧扫声纳图像增益补偿判断模型,依据此模型可以实现在数据采集时对图像增益做出自动判断并进行合理补偿,提高实时生成的图像质量。模型采用灰度直方图二次拟合的方法对图像增益进行判断,考虑到侧扫声纳拖鱼两侧可能出现回波强度不一致的情况,模型也提出了相应的判断方法。最后分别利用不同型号侧扫声纳系统的扫测结果进行验证,证明了该模型的普遍适用性,在实际应用中可以减少侧扫声纳系统对使用者的依赖。
In order to solve the problems of uneven brightness and low contrast of side-scan sonar images caused by the complexity of the marine environment,based on the operating principle of side-scan sonar,this paper studies on the gain compensation methods commonly used in the field real-time control of sonar images and the inner data processing,and then proposes a gain compensation judgment model for side-scan sonar images.It can realize automatic judgment and suitable compensation of image gain during data acquisition in field work,and improve the quality of real-time generated images.The model selects the image in the appropriate region as the judgment sample,and uses the gray histogram twice fitting method to judge the image gain of the sample image.Considering that the echo intensity may be inconsistent on both sides of the towed fish of the side scan sonar,the model also puts forward the corresponding judgment principles.Finally,the data of different types of side scan sonar systems are used to verify,which proves the universal applicability of the model and can reduce the dependence of the side scan sonar system on users in practical applications.
作者
徐博阳
王方旗
丁继胜
周兴华
魏玉阔
张晓波
XU Boyang;WANG Fangqi;DING Jisheng;ZHOU Xinghua;WEI Yukuo;ZHANG Xiaobo(College of Ocean Science and Engineering,Shandong University of Science and Technology,Qingdao 266590,China;First Institute of Oceanography,MNR,Qingdao 266061,China;Beijing Hydro-Tech Marine Technology Co.,Ltd.,Beijing 100176,China)
出处
《海洋测绘》
CSCD
北大核心
2023年第2期15-19,共5页
Hydrographic Surveying and Charting
基金
国家自然科学青年基金(42106072
41606056)
山东省自然科学青年基金(ZR2020QD071)。
关键词
海洋探测
侧扫声纳
增益自动判断
增益补偿
直方图二次拟合
submarine detection
side-scan sonar
gain automatic judgment
gain compensation
histogram quadratic fitting
