船舶立体图像的特征提取能够获取船舶的几何特征、视觉特征、纹理特征等，对于船舶识别、目标搜索等有重要的意义。获取过程受图像、视频设备、数据传输链路等影响，立体图像可能存在局部缺失等问题。本文首先介绍3种立体图像的质量评估方法，并针对无参考质量评价方法进行研究，详细介绍基于颜色、自然度的无参考质量感知特征提取和基于局部分析的无参考船舶立体图像质量特征提取，对于改善船舶立体图像的识别效果有显著的作用。
Feature extraction of ship stereoscopic images can obtain ship geometric features, visual features, texture features, etc., which is of great significance for ship identification and target search. Due to the influence of image, video equipment and data transmission link, the stereo image may have some problems such as local missing. First of all, this paper introduces three kinds of stereoscopic image quality evaluation methods, and conducts in-depth research on the non-reference quality evaluation methods, and introduces in detail the non-reference quality perception feature extraction based on color and naturalness and the non-reference ship stereoscopic image quality feature extraction based on local analysis, which has a significant effect on improving the recognition effect of ship stereoscopic image.
2023,45(16): 145-148 收稿日期：2023-2-27
DOI：10.3404/j.issn.1672-7649.2023.16.030
分类号：U665.2
作者简介：方银龙(1966-),男,实验师,研究方向为航海技术
参考文献：
[1] 翟方乾, 高尚, 赵昱博, 等. 立体图像数字水印技术研究综述[J]. 信息与电脑(理论版), 2023, 35(5): 41–43
[2] 卢佳佳, 蔡坚勇. 基于增强视觉质量的图像感兴趣区域检测研究[J]. 计算机仿真, 2023, 40(1): 234-238.
[3] 罗传未, 张子慧, 贺子婷, 等. 一种基于立体注意力机制的立体图像超分辨算法[J]. 电视技术, 2023, 47(1): 30-35.
[4] 冯丽露, 康耀龙. 立体图像任意剖面轮廓线提取方法仿真研究[J]. 计算机仿真, 2022, 39(8): 239-242+285.
[5] 张晓媛, 于洋, 王新蕊. 三维图像虚拟视点生成优化研究仿真[J]. 计算机仿真, 2022, 39(3): 205-209.
