针对当前三维重构技术受到光照干扰而导致重构效果不佳的问题，提出视觉传达技术的舰船图像三维重构研究方法。采用阈值分割法，提取舰船图像轮廓特征。结合模糊处理法进行纹理特征梯度分解，构建三维图像视觉传达模型。采用视觉特征提取技术重组空间信息特征，构造网络分布重组模型，匹配图像特征点。结合三维图像视觉传达动态交互机制，对三维图像的空间信息进行重构，获取纹理特征的分布集合。结合平滑滤波方法进行图像降噪处理，获取三维图像重建结果。使用视觉传达技术，实现图像增强，避免光照干扰。由实验结果可知，该方法三维点坐标与理想情况基本一致，能够得到高清晰的图像重构结果，最高信噪比为34 dB，说明使用该方法重构效果较好。
Aiming at the problem that the current 3D reconstruction technology is affected by illumination, a research method of 3D reconstruction of ship image based on visual communication technology is proposed. Threshold segmentation method is used to extract contour features of ship image. Combined with fuzzy processing method, texture feature gradient decomposition is used to construct 3D image visual communication model. Using visual feature extraction technology to obtain spatial information features, a network distribution recombination model is constructed to match image feature points. Combined with the dynamic interaction mechanism of 3D image visual communication, the spatial information of 3D image is reconstructed to obtain the distribution set of texture features. The image denoising process is combined with smooth filtering method to obtain 3D image reconstruction results. Use visual communication technology to realize image enhancement and avoid light interference. The experimental results show that the 3D point coordinates of this method are basically consistent with the ideal situation, and the high-resolution image reconstruction results can be obtained with the highest SNR of 34dB, indicating that the reconstruction effect of this method is good.
2022,44(8): 161-164 收稿日期：2021-09-17
DOI：10.3404/j.issn.1672-7649.2022.08.034
分类号：TP391.7
作者简介：邬星波(1982-),女,硕士,讲师,研究方向为视觉传达及图像三维设计等
参考文献：
[1] 宋燕飞, 罗尧治, 沈雁彬, 等. 基于双目视觉与图像识别的网架结构三维重建[J]. 空间结构, 2020, 26(4):28-35+74
[2] 尹燕运, 师影. 计算机视觉技术在交通工程测量中的应用[J]. 工程技术研究, 2020, 61(5):96-97
[3] 林平, 李琦, 申作春. 连续场景太赫兹数字全息三维重建图像的参数影响[J]. 激光与光电子学进展, 2020, 57(22):49-57
[4] 孙克强, 缪君, 江瑞祥, 等. 基于空洞卷积与多尺度特征融合的室内场景单图像分段平面三维重建[J]. 传感技术学报, 2021, 34(3):370-378
[5] 冯维, 汤少靖, 赵晓冬, 等. 基于自适应条纹的高反光表面三维面形测量方法[J]. 光学学报, 2020, 40(5):119-127
[6] 邢志勇, 肖儿良, 简献忠. 双判别生成对抗网络的红外图像超分辨重建[J]. 小型微型计算机系统, 2020, 41(03):662-667
[7] 万书亭, 张伯麟, 尹涛, 等. X射线无损检测图像三维重建软件设计[J]. 中国工程机械学报, 2020, 18(5):425-429+435
[8] 张豪, 张强, 邵思羽, 等. 深度学习在单图像三维模型重建的应用[J]. 计算机应用, 2020, 40(8):2351-2357
