摘要
侧扫声呐在水下搜寻作业的过程中会受到环境、噪声以及机器本身等诸多不确定的因素干扰,声呐生成的图像会受到较大的影响,从而干扰探查人员对水下目标的判断。本文利用Opencv计算机视觉算法工具,使用Python建立Clahe、平滑中值卷积、形态学算法以及拉普拉斯变换等的图形处理模型,利用实际数据为样本进行实验。实验结果表明,Clahe算法最适用于去噪降噪以及消除横向条纹的场景,平滑中值卷积适用于降噪且保留原特征,形态学运算适合单一特征提取,拉普拉斯变换在底部环境声波反射较强情况下的特征物提取能力最优。
Side scan sonar is often disturbed by many uncertain factors such as environment,noise and equipment instability factors in the process of underwater operation.The image generated by sonar will be greatly affected,thus interfering with the exploration personnel′s judgment of underwater targets.In this paper,the Opencv visual algorithm tool is used to establish the graphics processing models of Clahe,smooth median convolution,morphological algorithm and Laplace transform by using Python.The data from actual operations are used as samples for experiments.The experimental results show that the Clahe algorithm is most suitable for denoising and eliminating transverse stripe.The smooth median convolution can denoise and retain the original features.The morphological operation can be used in the extraction of single feature.The Laplacian transform has the best feature extraction ability in the case of strong reflection in the bottom section.
作者
陈淇玮
CHEN Qiwei(Nanhai Rescue Bureau of the Ministry of Transport,Guangzhou 511477,China)
出处
《舰船科学技术》
北大核心
2024年第24期104-108,共5页
Ship Science and Technology
关键词
声呐图像
图形优化
算法模型
sonar image
image optimization
algorithmic model
