基于卷积神经网络的海底地形识别  

Seabed Terrain Recognition Based on Convolutional Neural Network

作　　者：王永东 陈文锋 刘果 WANG Yongdong;CHEN Wenfeng;LIU Guo(CRCC Harbour&Channel Engineering Bureau Group Co.,Ltd.,Zhuhai 519000,China)

机构地区：[1]中国铁建港航局集团有限公司,广东珠海519000

出　　处：《电声技术》2025年第2期78-82,共5页Audio Engineering

摘　　要：针对海底地形识别任务中识别精度低的问题,提出一种运用卷积神经网络(Convolutional Neural Network,CNN)进行海底地形识别的算法。该算法通过提取收集到的海底点云数据并进行预处理使其与CNN卷积层输入相适应,随后使用CNN进行识别与检测。通过混淆矩阵对算法性能进行验证与评估,实验结果表明,基于CNN的海底地形识别算法在识别与分类中的准备率都超过90%,错误预测率均低于10%,表明该算法具备应用于识别海底地形任务的潜力。Aiming at the problem of low recognition accuracy in the task of seabed terrain recognition,an algorithm for seabed terrain recognition using Convolutional Neural Network(CNN) is proposed.The algorithm extracts and preprocesses the collected submarine point cloud data to make it adapt to the input of CNN convolution layer,and then uses CNN to identify and detect it.The performance of the algorithm is verified and evaluated by the confusion matrix.The results show that the preparation rate of the underwater terrain recognition algorithm based on CNN is over 90% in recognition and classification,and the false prediction rate is less than 10%,which shows that the algorithm has the potential to be applied to the task of recognizing underwater terrain.

关 键 词：卷积神经网络(CNN) 地形特征 深度学习 地形识别 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论] TN98[自动化与计算机技术—计算机科学与技术]