基于DDTW的非线形建筑数字图像形貌拼接仿真  

Simulation of Digital Image Mosaic of Non-Linear Building Based on DDTW

作　　者：张华 ZHANG Hua(Jincheng College,Sichuan University,Chengdu Sichuan 611731,China)

机构地区：[1]四川大学锦城学院,四川成都611731

出　　处：《计算机仿真》2020年第6期448-451,470,共5页Computer Simulation

摘　　要：针对传统图像形貌拼接完成后,接口较为明显、消耗时间较长的问题,提出DDTW的非线形建筑数字图像形貌拼接方法。分析建筑物的轮廓曲线几何特征,根据分析结果对算法进行估计,分别计算所有子轮廓的线特征串以及反向特征串,方便进行匹配检测。以计算结果为基础采用相位的相关算法,获取图像间的相对位移距离,同时提取图像特征点,并采用PERANSANC对特征点进行过滤,完成图像匹配。对匹配后的图像进行拼接预处理、图像缝合以及图像画出,从而完成非线性建筑图像形貌拼接。分析实验结果得出,提出的方法与传统方法相比,图像拼接完成后没有明显的拼接痕迹,细节也较为完整,同时拼接速度较快,说明其具有显著的有效性。Due to obvious interface and high time consumption in traditional image mosaic technology, a non-linear digital image mosaic method based on DDTW was proposed. The geometric features of building contour curve were analyzed. The algorithm was estimated by analysis results. The line feature strings and reverse feature strings of all sub-contours were calculated respectively. This was convenient for the matching detection. Based on the calculation results, the phase correlation algorithm was used to calculate the relative displacement distance between images and extract image feature points. PERANSANC was used to filter the feature points and thus to complete the image matching. Finally, the matching images were stitched and quilted and drawn to complete the non-linear building image mosaics. Compared with the traditional method, the proposed method has no obvious splice marks after the image stitching, and the details are relatively complete. Meanwhile, the image mosaic speed is faster, so the proposed method has significant effectiveness.

关 键 词：非线性建筑 数字图像 特征串 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]