基于计算机应用技术的舰船图像修复研究  

The ship image restoration based on computer application technology research

作　　者：马伟 MA Wei(Qinghai Communications Technical College,Xining 810003,China)

机构地区：[1]青海交通职业技术学院信息工程学院,青海西宁810003

出　　处：《舰船科学技术》2021年第18期190-192,共3页Ship Science and Technology

基　　金：青海省基础研究计划(2020-ZJ-709)

摘　　要：常规方法匹配待修复样本块和填充样本块时,忽略了梯度项距离这一影响因素,造成修复图像与原始图像相似度较差。为此,提出基于计算机应用技术的舰船图像修复方法。采用计算机应用技术,可视化图像待修复区域轮廓,划分多个待修复样本块,通过优先权确定样本块修复顺序,计算待修复样本块、填充样本块之间的颜色项距离、边界项距离、梯度项距离,选取距离最小的填充样本块,填充至匹配到的待修复样本块,完成图像修复。抽取舰船图像,设置对比实验,结果表明,设计方法相比2种常规方法,提高了修复图像PSNR值和SSIM值,修复图像与原始图像的纹理相似度、结构相似度更高。When the conventional method matches the sample block to be repaired and the sample block to be filled,the influence factor of the gradient distance is ignored,resulting in poor similarity between the restored image and the original image.Therefore,a ship image restoration method based on computer application technology is proposed.The computer application technology is used to visualize the contour of the image to be repaired,divide multiple sample blocks to be repaired,determine the restoration order of the sample blocks according to the priority,calculate the color item distance,boundary item distance and gradient item distance between the sample blocks to be repaired and the filling sample blocks,select the filling sample block with the smallest distance,and fill it to the matching sample block to be repaired,and complete the image repair.Extract the ship image,set the contrast experiment,the result shows that design method compared with the two conventional methods,improves the image PSNR and SSIM values,repair to repair the image with the original image texture similarity,structural similarity is higher.

关 键 词：计算机应用技术 待修复区域 填充区域 样本块 图像修复 候选块 

分 类 号：TP919[自动化与计算机技术]